Osgeo Journal the Journal of the Open Source Geospatial Foundation Volume 1 / May 2007

OSGeo Journal The Journal of the Open Source Geospatial Foundation Volume 1 / May 2007

In This Volume

FAO case study

Introducing MapWindow, GeoNetwork ...

Topology Basics

GRASS & distributed computing

Interview Why we use OpenSource software

Software Updates

News, and more... OSGeo Journal Editorial Vol. 2, August 2007

Editorial

Community Flexibility

I always enjoy working on this Journal. It is always varied so much between volumes. In this volume ..... interesting to see what kinds of projects and ideas are going through the minds of those in our communities. It is also exciting to see the various developments happening both inside and outside of OSGeo. Tyler Mitchell I am one of those people that likes continual change, Editor in Chief so this Journal ﬁts into that interest of mine because http: // osgeo. org both the contributors and the types of content have tmitchell AT osgeo.org Contents of this volume: Case Studies 21 FAO GeoNetwork in a Multinational Develop- ment Programme ...... 21 Editorial 1 The West Alexander Neighbourhood in Win- Community Flexibility ...... 1 nipeg ...... 25 Navigating with Innovation ...... 26

News 4 Topical Studies 29 Brief News from the OSGeo Community .... 4 Spatial Relationships in GIS – Geospatial Topology Basics ...... 29 Event Reports 7 Events of the OSGeo India Chapter ...... 7 Programming Tutorial 33 OSGeo at Géoévénement 2007 ...... 9 Porting a GRASS raster module to distributed computing ...... 33

Project Spotlights 13 Interview 43 Introducing the MapWindow GIS Project ... 13 LizardTech: Why We Use and Support Free GeoNetwork Open Source ...... 16 and Open Source Software ...... 43

ISSN 1994-1897 1 OSGeo Journal Editorial Vol. 2, August 2007

Sponsor Spotlight 47 Nascent UK Local Chapter Update ...... 53 Spatial Data Quality and the Open Source Spanish Local Chapter Update ...... 54 Community ...... 47 Ottawa Local Chapter ...... 56

Local Reports 51 ...... 51 Developer Announcements 57 Activity Report of OSGeo Taiwan ...... 52 So simple to install ...... 57

ISSN 1994-1897 2 OSGeo Journal Editorial Vol. 2, August 2007

ISSN 1994-1897 3 OSGeo Journal Brief News from the OSGeo Community Vol. 2, August 2007

News

Brief News from the OSGeo Community

Compiled by Jason Fournier With the 2007 conference approaching1 and the ex- citement building in the FOSS4G community, the 2008 conference will be an important date on the cal- General News endar not to miss. The 2008 organizing committee has already begun preparations, and will be follow- FOSS4G 2008 Call for Proposals Closes ing up with a larger announcement. The detailed 2 The RFP solicitation closed July 20th, 2007. Four or- proposal is available online. ganizations submitted proposals to host next year’s Again, thank you for all of the excellent submis- event: sions. OSGeo Conference Committee • Bangkok, Thailand • Cape Town, South Africa • Florianópolis, Santa Catarina, Brazil Google Summer of Code Projects • Seville, Spain There are 15 OSGeo-sponsored projects underway in the Google Summer of Code program: FOSS4G 2008 Awarded to Cape Town, South Africa • GRASS Modules for line generalization and smoothing The OSGeo Conference Committee is pleased to an- • GDAL Raster Driver for PNG/JPG Tile Struc- nounce that the 2008 Free and Open Source Software ture + gdal2tiles utility for Geospatial (FOSS4G) conference will be held in • Materialized Views with Geometric Support Cape Town, South Africa, with the tentative dates • Plugins for multidimensional raster data of September 29th to October 3th 2008. The OS- sources. Geo Conference Committee would like to thank all of • Caching data in uDig those who submitted to host the 2008 conference, we • JTileCache were all very impressed with the number and quality • Shortest path in free (vector) space avoiding of proposals. obstacles module in GRASS 1FOSS4G 2007 web site: http://www.foss4g2007.org 2South Africa proposal to host FOSS4G: http://www.osgeo.org/files/webfiles/conference/proposals/South_Africa_ proposal_to_host_FOSS4G_2008.pdf

ISSN 1994-1897 4 OSGeo Journal Brief News from the OSGeo Community Vol. 2, August 2007

• GDAL: KML read support for the existing MapServer driver • New Transformation Algorithms for GeoTools MapServer 5.0 is slated to be released just in time for and uDig FOSS4G 2007 on Wed Sept. 12, 2007. With this re- • Coverage model and operations for PostGIS lease comes a number of feature enhancement pro- • GeoServer Style Editor posals including: • Implementation of An Interactive GeoRSS tool in uDig • Enhanced image output quality via the AGG li- • A Synthetic Aperture RADAR Processor using brary GDAL • Dynamic Charting Capability • 3D Rendering Pipeline for GeoTools • Labelling prioritization • OGC WMS GDAL driver • and many more!

For more information on each project, including the For more details please visit the MapServer Release participant and mentor, please visit the Google SoC Plan web page.5 OSGeo page.3

QGIS Project News It is our great pleasure to announce the immediate availability of Quantum GIS (QGIS) Version 0.8.1. The release includes numerous bug ﬁxes and stabil- MapGuide ity improvements to the 0.8 code base. QGIS is available in source form, and will be available as binary MapGuide Open Source announces that the Fusion executables for Microsoft Windows, Mac OS X, and Technical Preview is available for download. The an- GNU/Linux. nouncement was made by Paul Spencer of DM Solu- 6 tions Group Inc.: For more information see the QGIS web site.

"Hello MapGuide Open Source users and developers, MapBender I am very happy to announce that DM Mapbender 2.4.2 was released on August 2, 2007. A Solutions Group will be contributing our number of features have been added including: Fusion technology to the MapGuide Open Source project. Fusion is the project name for a Software Developers Kit that pro- • Bulgarian translation of the introduction to vides tools and templates to speed devel- Mapbender opment of web mapping applications. • improved module Add WMS from filtered list: Having worked with MapGuide Open loading a configured WMS from a GUI is now Source since its first release, we saw an possible opportunity to better meet our clients ap- • JavaScript compression; Mapbender will load plication needs by creating a new toolkit faster based on our experience with MapServer • first draft of an API documentation for - and we made quite a few improvements JavaScript classes on the way!" • and many more!

For more information please visit the MapGuide OS- For a complete listing of changes and bug ﬁxes please Geo home page.4 visit the MapBender changelog page.7 3Google SoC OSGeo page: http://code.google.com/soc/2007/osgeo/about.html 4MapGuide home page: http://mapguide.osgeo.org/ 5MapServer Release Plan page: http://mapserver.gis.umn.edu/development/release_plans/mapserver_5_0/release_plan/ 6QGIS web site announcement: http://qgis.org/content/view/143/65/ 7MapBender changelog page: http://www.mapbender.org/index.php/Template:Changelog_2.4.2_rc1

ISSN 1994-1897 5 OSGeo Journal News Vol. 2, August 2007

GDAL/OGR ber of new message translations, and updates for the help pages. The GDAL/OGR team is pleased to announce the re- For more information please visit the GRASS GIS lease of GDAL/OGR 1.4.2. This is intended to be news item.9 a bug fix release extending the GDAL/OGR 1.4 stable branch and does not include the many new features in the development trunk (which will appear as GeoTools GDAL/OGR 1.5.0). For more information please visit the GDAL/OGR GeoTools releases version 2.3.3 on August 2 2007. news item.8 This is a bugfix only release. Most notable changes are Oracle datastore improvements (bug fixes and speed ups), as well as improvement in coverage serv- GRASS ing and rendering speed. 10 GRASS 6.2.2 is a new stable release which fixes sev- For a full change log, see the online release notes. eral bugs discovered in the 6.2.1 source code. This release is solely for stability purposes and adds no new OpenLayers features. The default Graphical User Interface has been further stabilised, and the LIDAR processing OpenLayers releases version 2.4. OpenLayers has tools and Linear Reference System support signifi- addressed over 100 outstanding tickets with the 2.4 cantly improved. GRASS 6.2.2 also includes a num- release. Release notes are available online.11

8GDAL/OGR release announcement: http://www.osgeo.org/node/367 9GRASS GIS release announcements: http://www.osgeo.org/node/369 10GeoTools release notes: http://jira.codehaus.org/secure/ReleaseNote.jspa?projectId=10270&styleName=Html&version= 13532 11OpenLayers release notes: http://trac.openlayers.org/wiki/Release/2.4/Notes

ISSN 1994-1897 6 OSGeo Journal Events of the OSGeo India Chapter Vol. 2, August 2007

Event Reports

Community Reports from Meetings and Conferences

Events of the OSGeo India Chapter Free and Open Source Software In National Conference on Biodiver- Science sity Conservation

4-6 January, 2007, Union Christian College, Alwaye, 8-10 February, 2007, Osmania University Kochi The National Conference on Biodiversity Conser- In this programme by an educational institution, vation and Human Well-Being was organized by De- the participants were introduced to Free and Open partment of Zoology, Osmania University and co- Source GIS software, and the organisations that pro- sponsored by SACON Deccan Regional Station, Hy- mote them. A dataset prepared for the city of Hyder- derabad. The conference was held at Osmania Uni- abad, using freely available satelite imagery along versity during February 8-10, 2007. with free downloads of SRTM data was demon- SACON conducted a workshop on Wetland In- strated by V.Ravi Kumar, emphasising FOSS GIS. formatics. It was aimed to address some of the issues For more information see the online report of the related to spatial information generation, processing course.12 and data base design, dissemination and software development. In this workshop SACON presented the Wetlands of India website and it’s future with the extensive use of FOSS tools. The 1st OSGeo Workshop V.Ravi Kumar, Joint Secretary OSGeo India Chap- ter presented about the various FOSS4G’s GIS soft- The very ﬁrst OSGeo Workshop was held at the Map ware and their use taking the City of Hyderabad as World Forum, Hyderabad 22-25th January 2007. This an example. Prof K.S.Rajan, treasurer of OSGeo India well attended event was also participated by OSGeo chapter also actively participated in the proceedings. delegates from China, Vietnam and Japan. This event P.K.Sinha and A.V.Satya Kumar demonstrated FOSS has been covered in several other reports.13 GIS by showing OpenJUMP and it’s user-friendly 12Free and Open Source Software In Science site: http://www.space-kerala.org/fsc/index.php/Fsc/FreeAndOpenSourceGIS 13OSGeo India Workshop event coverage in Geospatial blog: http://geospatial.blogs.com/geospatial/2007/01/map_world_ forum.html, GIS Development Magazine: http://www.gisdevelopment.net/magazine/years/2007/march/46.htm, OSGeo Journal: http://www.osgeo.org/files/journal/final_pdfs/OSGeo_vol1_Events.pdf

ISSN 1994-1897 7 OSGeo Journal Events of the OSGeo India Chapter Vol. 2, August 2007 capabilities in performing vector GIS tasks like on- sultancy Servicesv (TCS), MapInfo, Adroitec, GIS screen digitization of raster images to shape file with Consortium, Eicher, 5Map, Infosys, Autodesk, Na- attribution. tional Institute of Health and Family Welfare, and Bharti Vidyapeeth’s Institute of Computer Applica- tions and Management (BVICAM). National Convention On Free Soft- The workshop was opened by OSGeo-India rep- ware resentative Dr. P.S.Roy, with an opening remark that,"GIS has emerged as an important technology tool. 3-4 March, 2007, Hyderabad, Andrhra Pradesh GIS initiatives need to be upscaled in terms of outreach National Convention On Free Software, 3rd - and this is where OSGeo comes into play." One of the 4th of March 2007, Hyderabad, Andhra Pradesh in- most important objectives, from an Indian context, cluded a session on Geospatial data. Dr. K.S.Rajan, is to convert the free GIS tools like GRASS and and V.Ravi Kumar of OSGeo-India chapter, pre- MapServer in vernacular languages to form a large sented OSGeo Philosophy, and Open Source GIS al- interest group. Dr. Roy also elaborated upon some ternatives respectively, which was attended by Free of the noteworthy initiatives taken by organizations Software Foundation (FSF) enthusisasts from around in India in the open source domain. 14 the country. V.Ravi kumar, who is a founder member of OSGeo India, spoke about various FOSS GIS like GRASS, Quantum GIS, UDIG, Open Jump available Geohorizon 2007 Symposium today. The demonstration on ’Locating a Nuclear Power Plant’, using Open Jump, was not only in- 16-17 March, 2007, Anna University, Chennai teresting but also testified how FOSS GIS can be Society of Geoinformatics Engineering at Anna equally useful as commercial GIS software. Ravi Ku- 15 University , Chennai, conducted a two-day Geo- mar also talked in length about utility of FOSS GIS horizon 2007 symposium (16-17 March 2007). in India. The most pertinent point raised by him was There were four participants from the OSGeo In- that FOSS GIS is cost effective and can, therefore, be- dia chapter, two on each day. Sri Neeraj Deekshith come popular in a developing country like India. It of Autodesk India and Santosh Gaekwad of SACON can also provide employment through promotion of Hyderabad, presented use of Open Source Geospa- Village Cadastral GIS where all that one has to pay tial software in their organisations. for is a branded computer (available for less than Mr V.Ravi Kumar, and Mrs Mahalakshmi Rs.20,000) as no cost is incurred on the software be- Narayanan presented Open Source GIS Galore, and cause it is, free and open. Open Source GIS for web-services respectively. Santosh Gaikwad presented the use of GeoSever Sridhar Saraswati of Autodesk India, also gave for internet mapping of Wetlands of india. He a presentation, which answered a vital question – delievered the SACON’s experience while studying why is Autodesk supporting free and open source GeoServer – a open source server for Web-GIS. Use of software? The reason is that firstly, Autodesk wants Udig and GeoServer for publishing vector or raster a broader adoption of Internet mapping technology maps onto web by storing the spatial data into Post- and secondly, FOSS would create higher demand GIS was demonstrated. for related products. It would also be beneficial for the customers as the frequency of software releases will increase, cost of ownership will come down and OSGeo Workshop there will be more innovation and support for standards. 11 May, 2007, GIS Development, Noida Satyajit Rath from CDAC, Noida presented On 11th May 2007 one day OSGeo-India work- CDAC’s Initiatives on Open Source GIS. A talk enti- shop was held at, GIS Development, Noida (Delhi) tled Geospatial Convergence with Open Source was pre- with participants from the Indian geospatial com- sented by Amit Jain, of Infosys. The workshop came munity. They represented organizations like Na- to an end with the closing remarks by Dr. P.S Roy. He tional Remote Sensing Agency, Centre for Develop- expressed hope that "commercial software should invest ment of Advanced Computing (CDAC), Tata Con- in open source to make it a win-win situation for all." 14National Convention On Free Software: http://www.swecha.org/node/29 15Society of Geoinformatics Engineering: http://societyofgeoinformaticsengineers.blogspot.com/index.html

ISSN 1994-1897 8 OSGeo Journal OSGeo at Géoévénement 2007 Vol. 2, August 2007

OSGeo Capacity Building Work- ber and help spread the awareness. The resource per- shop sons were from Geological Survey of India. 26-28 June, 2007, Salim Ali Centre For Ornithology & Open Source GIS Lecture Natural history, Hyderabad The OSGeo-India chapter has conducted it’s 1st A two hour lecture come demonstration of Open Capacity building 3 day workshop from 26th to 28th Source GIS GALORE, was presented by V.RaviKumar June 2007 at Salim Ali Centre For Ornithology & Nat- of OSGeo India, for the students of Geoinformatics ural History (SACON) Hyderabad. The workshop post graduation course Nannaya University, Rajah- was attended by 18 participants from NIC, NRSA, mundry on 3rd July 2007. GRASS GIS and Open- IIIT, Spec Systems, Army and Survey of India. All the JUMP were demonstrated using sample data. The participants were conversant with commercial GIS students who could interact well in the question and software and were eager to acquaint with FOSS al- answer session were presented GRASS6.2 Live CDs. ternatives. The 1st day included demonstration and hands- on practice with OpenJUMP GIS and map composi- Photographs tion through Inkscape. The 2nd day was devoted to ILWIS GIS through raster analysis and generation of Several event photographs are available online.16 stereo pair and 3D visualization. The 3rd day was for using GRASS GIS for network analysis and cre- Ravi Kumar ation of 3D visualisation and ﬂy-bi. The participants Co-Chairman Events, OSGeo-India Chapter were shown how one can join the OSGeo as a mem- ravivundavalli AT yahoo.com

OSGeo at Géoévénement 2007 by Yves Jacolin

Presentation of the Géoévénement

Géoévénement is the annual french geomatics tradeshow. In 2007, it took place at Porte de Ver- sailles in Paris. It offers booth exhibits, thematic presentations and Birds of a Feather sessions. This year, more than 35 companies had a booth from which 13 were from the open source community. These 13 companies created an open source village which proposed more than 30 free presentations, free as in free beer and in free mind. Some companies proposed Figure 1: Tux welcomed people in the Open Source workshops about Mapserver, Cartoweb and Open- Village layers installation. OSGeo participated in this village with a booth and two presentations. There was also a Birds of a Feather about the project of a OSGeo Francophone Local Chapter discussing topics such as: how to or- This report aims at presenting the results of the ganize it, how to manage the election of ofﬁcers, booth, presentations and BoF. First of all, here is a which organisation need to be contacted, etc. presentation of the Francophone Local Chapter. 16Event photographs: http://www.kodakgallery.com/I.jsp?c=cjns10h3.batchqkn&x=0&y=o6f4mq

ISSN 1994-1897 9 OSGeo Journal OSGeo at Géoévénement 2007 Vol. 2, August 2007

The Francophone Local Chapter Local chapter ran a sprint translation project to translate the OSGeo Flyer,20 created a panel and showed The idea of creating a Francophone Local Chapter a presentation on two laptops. first emerged at FOSS4G 2006, Lausanne. In Octo- The booth welcomed a continuous stream of peo- ber 2006 the Francophone mailing list and the Fran- ple who showed a lot of interest in the OSGeo and its cophone wiki page were created. People subscribed projects, here are some frequently asked questions: to the mailing list and the number of members grew • What is OSGeo? to more than 75 people (still growing). Following • Is OSGeo different than AFIGEO? (AFIGEO is the Local Chapter Guidelines, the next step, after the a French geomatics association) auto-organisation, is to define the groups missions. • What is the difference between OSGeo and Here they are: OGC? Free Organisation • Do I have to pay something or subscribe to become an OSGeo member? • create a portal for the Francophone users of • How can I contribute? open source geomatics (data, software, jobs, ...) – How can I find what help is needed? • set up meetings and tradeshows – Where can I read the process to con- • promote OSGeo and its missions tribute? Free Data – Who should I contact? • Do you have some project demonstration on • promote open source licence Live-CD or anywhere? • host, share and create free data • Do you have any examples of projects? Free Software – Example of contribution? – Example of funding projects? • promote and demonstrate open source soft- • What is the equivalent of MapInfo/Cartes et ware Bases in Open Source? (Cartes et Base is a pro- • localise, translate software prietary GIS analysis tool) • translate and share documentation • Do you think you will create a summary of • support in french (mailing-list and forum) projects with a comparison of each? Some projects are still in development: • Does a simple tool exist to create a WMS web service? • Translation of the OSGeo Journal vol.1 (now • Does OSGeo plan to provide training? published17) • Does OSGeo fund any external projects? • Poll: "What are the interests of the geomatics Some questions focused on the migration from pro- communities about the OSGeo Francophone prietary software to FOSS and to learn if this soft- Chapter"? ware is good enough for their use, if they can find some training, etc. Promotion of OSGeo BoF Exhibition Booth This meeting took place at the GéoCafé (an open central place in the tradeshow) on the Wednesday 4th of The Francophone Local chapter put up an exhibition April at 14h00. About 20 people participated. The booth to promote the OSGeo mission and its hosted content included an OSGeo Présentation, Presenta- projects. We focused on FOSS and missions promotion of Local Chapters and the development of the tion rather than demonstrate some projects which Francophone one. Discussion also included: will be possible next time using some LiveCD (i.e. Omniverdi LiveCD18, Geolive19). In order to pro- • How can we organize ourself? Organisation, mote OSGeo missions and projects the Francophone election, ... 17OSGeo Journal Vol. 1 in French: http://www.osgeo.org/journal/volume_1 18http://livecd.ominiverdi.org/index.php?page=LiveCD&toc=livecd 19http://geolive.cartography.st/geolive/wiki/frenchDesc 20OSGeo Fliers: http://www.osgeo.org/visibility

ISSN 1994-1897 10 OSGeo Journal OSGeo at Géoévénement 2007 Vol. 2, August 2007

• Who can help support the Francophone Chap- Presentations ter? Contact some associations that share common missions (join, organize, structure, ...) Géoévénement organised some thematic presentations, one of them focused on "How to get your hands on Open Source technologies": Topics of the meeting were free and was larger than expected. • Rising of a FOSS industrial ecosystem - David Jonglez, Camptocamp France SAS and François Badet, WaynaSoft • OSGeo: why a francophone local chapter is important - Yves Jacolin, Géorezo and Gwenael Bach- elot, Autodesk • Free GIS: freeness for reachness? - Nicolas Klein, SI17 • Some elements to compare 3 GFOSS - Olivier Ertz, Erwan Bocher, Nicolas Ribot and Yves Jacolin • Development of an Information System for Water (SGBD - SIG) in the Charente county - Harold Rethoret, EPTB Charente

Figure 2: A presentation at the Open Source Village The Open Source Village also organised more than 35 presentations. Themes of this conference Some ideas which were discuss on the Franco- covered testimonials, presentation of OSGeo or Open phone mailing list were accepted: companies have Source projects (OpenLayers, cartoweb, Geonet- the same place as inside OSGeo (sponsor role), mem- work, etc.), free data, Web Services, open source li- bers are only physical individual people. Others cences, and OSGeo and project presentations. Al- ideas were proposed such as, the creation of a we- most all presentations are publicly available in the bring which can promote each other, running a poll OSGeo Library.21 to know more about user and contributor interests (in progress). Organisation of the proposed Francophone Local Chapter followed closely to its missions: a lead manager to connect the local chapter to the OSGeo (providing monthly reports, etc.) and three other committee managers to lead one of the three main missions: organisation, free data and open source software. Each of them have active projects (translating OSGeo Journal for example). If one of these projects is especially important a project manager will be chosen by the committee manager with agreement of the list. Election of this structure ran at this end of June but result are not known at the moment I am writing this article. Figure 3: A presentation at the Open Source Village Free data is a part of the mission that seems to get the main interest from Francophone communi- Conclusion ties: How to host data? Where? Do we need to set up a data catalog? It seems that some datasets could The existence of this Francophone OSGeo Local be free in the current year. Chapter project allowed us to inform others about Another meeting aims to do a synthesis about the OSGeo Foundation, and about the future Fran- the Open Source Village. Everybody seems to agree cophone local chapter. This is a success with 3 pre- about its success and will renew next year. We hope sentations about the OSGeo, a lot of people meeting that OSGeo-fr may manage this to get stronger. at the booth, and a great richness in the discussion. 21OSGeo Library: http://www.osgeo.org/library

ISSN 1994-1897 11 Event Reports OSGeo Journal Community Reports from Meetings and Conferences Vol. 2, August 2007

Furthermore, a first contact between some contrib- geomatics community. utors of the future francophone local chapter gave All this effort has been made possible by the gen- some great ideas and future project (e.g. about free erous financial support of AutoDesk inc, Campto- data topics, among others). camp SAS and Autodesk France. I would like to And now? We need to formalize the Francophone thank them for that. local chapter by electing the local chapter and com- Some links: mittees leader (end of june), and sending to OSGeo a proposal about its structure and organisation (plan • Conference papers are available online in for july). For the next tradeshow we aim to translate French22 and English.23. many more brochures and fliers into french, share • For more information about the Francophone Geomatics LiveCD and present some workshops in Local Chapter, see the OSGeo wiki page.24. addition to presentations. • For more information about the Géoévéne- Some media reported the success of the Open ment, see the conference web site.25 Source Village "An Open Source Village in the Géo- événement 2007 tradeshow" (Le Monde Informatique), Yves Jacolin "Rendez-vous at Géo Événement 2007" (SIG la lettre), Local Chapter Contact, OSGeo Francophone Chapter portailsig which relays OSGeo-fr informations to the yjacolin AT free.fr

22Conference papers (French): http://wiki.osgeo.org/index.php/Fr_geoevenement2007 23Conference papers (English):http://wiki.osgeo.org/index.php/GeoEvenement2007 24Francophone Local Chapter: http://wiki.osgeo.org/index.php/Francophone 25Géoévénement, see the conference web site: http://www.ortech.fr/geo-evenement/index.php

ISSN 1994-1897 12 OSGeo Journal Introducing the MapWindow GIS Project Vol. 2, August 2007

Project Spotlights

Introducing the MapWindow GIS Project by Daniel P. Ames, Christopher Michaelis and Ted regular source code committers. Dunsford The MapWindow components and end user application support manipulation, analysis and view- Introduction ing of geospatial data in many standard GIS data formats. Hence, MapWindow is a mapping tool, a GIS MapWindow is an open source geographic informa- modeling system, and a GIS API, in a redistributable tion system (GIS) and an application programming open source form that is appealing for its simplicity interface (API) distributed under the Mozilla Pub- of use and for running on Microsoft Windows. This lic License (MPL), built upon the Microsoft Dot Net paper presents the project history, current status, key Framework 2.0. The project is maintained by an elements and future plans. active group of international developers who regularly release updates and bug fixes through the MapWindow.org web site.26 Project Background In 2005, the United States Environmental Pro- tection Agency adopted MapWindow as the core In 1998, researchers at Utah State University in Lo- GIS platform for its BASINS watershed analysis and gan, Utah, USA were using MapObjects LT 1.0 from modeling software which is used by environmental Environmental Systems Research Institute (ESRI) as professionals at all levels of U.S. government and in- a GIS component in several research projects. How- ternationally. This was followed by adoption at the ever, restrictive redistribution licenses and the inabil- United Nations University and many other govern- ity to edit vector or attribute data files or display and mental, non-profit, and commercial entities through- interact with raster data became significant barriers out the world. Indeed, from November 2006 through to use of this component. July 2007 the software was downloaded over 60,000 Indeed, dynamic and interactive functions that times and has over 7,000 members on the opt-in mail- require low level access to spatial data were not well- ing list. There are also currently 630 subscribers to supported in freely distributable Windows-based source code commit notification mailing lists, and 43 GIS programming components at that time. An al- 26MapWindow web site: http://www.mapwindow.org

ISSN 1994-1897 13 OSGeo Journal Introducing the MapWindow GIS Project Vol. 2, August 2007 ternative approach for these projects was to move to these are available for download from the MapWin- a proprietary desktop GIS system (e.g. ArcView 3.1), dow.org web site. writing the needed tools as extensions. However, In addition to building plug-ins, one can cus- many applications in environmental science and en- tomize the look and feel of the MapWindow appli- gineering are more suited to a standalone environ- cation by changing the included default.mwcfg XML ment due to the need to distribute the tool to indi- configuration file. Changes to this file can affect ti- viduals who do not own a desktop GIS or are not tle bar, splash screen, about box, plug-in loading and GIS experts. window layout settings, among others. An XML The core MapWinGIS component was developed project file allows users to save a list of loaded data to meet this need for an alternative to existing pro- layers and session settings. The Newton Reservoir prietary GIS components. MapWinGIS is an Ac- Information System project in Figure 1 is an exam- tiveX control that gives full ability to display, query, ple project included in the MapWindow installation manipulate, and use spatial data. Because of the package. All of the data loaded into MapWindow engineering-oriented nature of the projects that ne- and all settings (e.g. symbology, loaded plug-ins, cessitated its development, MapWinGIS was partic- and project projection), are saved to the project file. ularly optimized for rapid display of data and is intended to be a fully functional model interface, not simply a map viewer. Development focuses included high-speed image and raster display, limiting the MapWindow Plug-ins amount of re-drawing, and including APIs for low- level access to spatial data. MapWindow has an extensible architecture that al- The MapWindow GIS desktop application wraps lows developers to write plug-ins to add function- all of the MapWinGIS ActiveX functionality, provid- ality using Dot Net compatible languages such as ing end users with a common legend, toolbar and a VB.Net or C#. This is done by implementing a spe- consistent interface that can be extended by adding cific plug-in interface in a custom DLL file which is plug-ins or a customized configuration file. In short, placed in the application directory and automatically MapWinGIS ActiveX was built to add GIS function- detected at runtime. Plug-ins can also be written and ality to custom end user applications and MapWin- compiled directly in the MapWindow plug-in editor dow GIS desktop was built to help users deploy GIS- eliminating the need for an external development en- enabled applications assembled from pre-built plug- vironment. The MapWindow plug-in interface pro- ins and datasets. vides extensive capabilities allowing the developer to use as little or as much as needed to accomplish Project Description a particular goal. Default plug-ins packaged with MapWindow include tools for editing shapefiles and MapWindow GIS Application attributes, identifying features and performing common geoprocessing tasks. Plug-in developers retain The MapWindow application is a lightweight spatial all copyright and licensing rights for their work and data viewer with an intentionally slim default inter- hence can redistribute their products as proprietary, face intended to simplify its use and improve adapt- open source, or otherwise as needed. ability to specific uses. The main GUI and functionality can be extended through plug-ins and scripts which add capabilities as needed. Also the default application itself can be customized to change the MapWindow ActiveX Control look, feel and even its title-bar name at runtime. By default, the MapWindow layout includes a map, a The core MapWindow component is an ActiveX con- legend, and a preview-map. Built-in toolbar buttons trol, MapWinGIS.ocx that can be used in Visual Ba- allow one to manage project files (collections of data sic or any language that supports ActiveX (e.g. C#, layers), to print, and to navigate the map. Figure Microsoft Access, Microsoft Excel). MapWinGIS has 1 shows the MapWindow interface with the Shape- been optimized for fast image and raster display, lim- file Editor plug-in enabled and with the title bar cus- iting the amount of redrawing, and including APIs tomized for a specific project. Both open source and for low-level access to spatial data. A simple Visual commercial plug-ins have been developed by users Basic program using the control and just a few lines worldwide for a variety of applications; many of of code is shown in Figure 2.

ISSN 1994-1897 14 OSGeo Journal Introducing the MapWindow GIS Project Vol. 2, August 2007

Figure 1: Basic MapWindow interface with Shapeﬁle Editor enabled.

Version 5.0 also introduces an interface layer around every component. By doing this, any given component (e.g., MapWindow.Data) may be replaced with a third-party component if needed. In- deed any code that implements the speciﬁed interfaces will be compatible with other MapWindow 5.0 components. Finally, MapWindow 5.0 components are designed and written entirely in Microsoft Dot Net code (Managed C ++, C#, and VB.Net) except for linked-in components (GDAL and OGR) which are maintained in their native form. This eliminates many long-standing problems with COM registra- Figure 2: A Simple Mapping Application tion on the Windows platform. An overview diagram of the planned MapWindow 5.0 components and linkages is shown in Figure 3. Ongoing Development and Future Plans

Development plans for MapWindow include a component structure upgrade in version 5.0. This version focuses on incorporating the Geospatial Data Access Library (GDAL) and OGR Simple Feature Library for additional data format support, and Net Topol- ogy Suite (NTS) for topological support. Version 5.0 also targets OGC compliance and improved modu- larization. An ASP.Net-based web component for performing online mapping and geoprocessing has also been introduced. Figure 3: MapWindow 5 Component Diagram

ISSN 1994-1897 15 OSGeo Journal GeoNetwork Open Source Vol. 2, August 2007

Conclusion Idaho State University http: // www. hydromap. com/ The MapWindow GIS project is a dynamic and active dpames AT gmail.com effort to build and distribute open source GIS tools that integrate tightly with the Microsoft Windows Chris Michaelis operating system. Supported by a large number Idaho State University of commercial and non-commercial funding sources, http: // www. happysquirrel. com/ volunteer developers, and student research activi- cmichaelis AT happysquirrel.com ties, this project is expected to grow in the foresee- able future and continue to be a model of a successful open source for geoinformatics software develop- Ted Dunsford ment and use. Idaho State University http: // www. MapWindow. org/ Daniel P. Ames hadunsford AT mapwindow.org

GeoNetwork Open Source

Internationally Standardized Distributed Spatial as the Consultative Group on International Agricul- Information Management tural Research (CGIAR), the European Space Agency (ESA), the US Federal Geographic Data Committee by Jeroen Ticheler and Jelle U. Hielkema (FGDC), in countries implementing National Spatial Data Infrastructures (NSDI), in Europe in the con- GeoNetwork opensource is a standardized and decen- text of The INfrastructure for SPatial InfoRmation tralized Geospatial Information Management Sys- in Europe (INSPIRE), as contribution to the Group tem, based on the concept of distributed data and on Earth Observations (GEO) and others, GeoNetwork information ownership and is designed to enable ac- opensource has also seen a remarkable uptake. cess to geo-referenced data and cartographic products through descriptive metadata. Collections maintained at different organizations are accessible The GeoNetwork opensource through a single entry point. This approach of ge- project ographic information management enhances easy and timely access to existing spatial data for a wide GeoNetwork opensource is a decentralized geospatial community of users, supporting informed decision- information management system based on interna- making in a variety of application ﬁelds. Web2 tech- tional standards. It is designed to enable easy niques have been adopted to allow more interactive and timely access to geo-referenced data and carto- and intuitive use of the system and to offer building graphic products through descriptive metadata from blocks for future web service offerings. a variety of data providers. The system tries to The main goal of the GeoNetwork opensource support a wide community of spatial information project is to provide free and open source soft- users to make better-informed decisions. It im- ware that improves accessibility to a wide variety of proves geospatial information sharing between or- geospatial data and information products at differ- ganizations and their users, using the capacities of ent scales and from distributed sources. The collec- the Internet. It often triggers improved collaboration tions are organized, documented and published in within and between organizations, reducing duplica- a standardized, consistent way both at the metadata tion and enhancing information consistency and data and data level following international standards. As preparedness. such the tool has become a popular and widely ac- Maps, including those derived from satellite im- cepted geospatial Information Management System agery, are effective communication tools and play an within the UN system, forming a central and basic important role in the work of various user communi- element in the development of a United Nations Spa- ties: tial Data Infrastructure (UNSDI). In other national and international organizations and initiatives such • Decision Makers, e.g. sustainable development

ISSN 1994-1897 16 OSGeo Journal GeoNetwork Open Source Vol. 2, August 2007

planners and humanitarian and disaster emer- were possible using the Z39.50 catalog protocol. gency managers in need of quick, reliable and With support from the UN Environment Pro- up to date cartographic products as a basis for gramme (UNEP) a second GeoNetwork version was better planning, action and monitoring of their developed in 2004. The release allowed users to activities. work with multiple metadata standards in a trans- • GIS Experts in need of exchanging consistent parent manner. It also allowed metadata to be shared and updated geographic data. between catalogues through a caching mechanism, • Spatial analysts in need of multidisciplinary improving reliability when searching in multiple cat- geospatial data to perform geographic analysis, alogs. A further major improvement was made with resulting in more reliable forecasts that help to the release of version 2.0.2 in April 2006 which was better organize interventions in vulnerable ar- accompanied by CD version of the software, includ- eas. ing a variety of other open sources packages, as well as a detailed user installation manual and an attrac- In 2001 the Food and Agriculture Organization tive flyer, followed up by a DVD version in October (FAO) of the United Nations FAO started the de- 2006. velopment of a prototype catalog system to system- The latest release of GeoNetwork opensource, ver- atically archive and publish geospatial data devel- sion 2.1 (July 2007), is the result of another round oped and used within the organization. The proof critical improvements, supported by among oth- totype build on experiences within and outside the ers FAO, the UN Office for the Coordination of Hu- organization. It used metadata content available manitarian Affairs (OCHA), the Spatial Consortium from legacy systems that was transformed into what of the Consultative Group on International Agri- was than a preliminary draft metadata standard, ISO cultural Research (CSI-CGIAR) and UNEP. Support 19115. for the final metadata standard ISO19115:2003 has been added by using the ISO19139:2007 implementation specification released in May 2007. The release also serves as the open source reference implementation of the OGC Catalog Service for the Web (CSW 2.0.1) specification. Improvements to give users a more responsive and interactive experience have been made on parts of the system, including the web map viewer and the search interface. At present, GeoNetwork opensource is a central and basic building block in the UN Spatial Data Infras- tructure (UNSDI) currently under development. In other national and international organizations and initiatives such as the (CGIAR), the European Space Agency (ESA), the US Federal Geographic Data Figure 1: GeoNetwork opensource version 2.0. Shown Committee (FGDC), in countries implementing Na- is the Map Viewer that is embedded in the applica- tional Spatial Data Infrastructures (NSDI), in Europe tion. in the context of The INfrastructure for SPatial InfoR- mation in Europe (INSPIRE), as contribution to the Joined in 2003 by the World Food Programme Group on Earth Observations (GEO) etc., GeoNetwork (WFP), the prototype was taken as input for a first opensource has also seen a remarkable uptake. version of GeoNetwork opensource. It was decided to develop the software as a Free and Open Source Soft- ware (FOSS) to allow other parties to benefit from the development efforts and to allow for their par- Free and Open Source Software ticipation. The result was a release 1.0 in 2003 of and Community Development the software as well as the establishment of operational catalogues in FAO and WFP. The system was Started as a small open source project in a very spe- based on the ISO19115:DIS metadata standard and cialized field, the community of users and devel- provided an embedded Web Map Client, InterMap opers of the GeoNetwork opensource software has in- that supported Open Geospatial Consortium (OGC) creased dramatically since the release of version 2.0 compliant Web Map Services. Distributed searches in December 2005 and the subsequent pre-releases

ISSN 1994-1897 17 OSGeo Journal GeoNetwork Open Source Vol. 2, August 2007 of version 2.1. At present the project sees over 800 incubation process is currently ongoing and is close downloads of the software per month from the web- to conclusion. The project websites have been moved site. The user and developer mailing lists have close to servers accessible under the umbrella of the OS- to 200 subscriptions each. Members provide feed- Geo foundation. Web pages have been updated to re- back within the community and provide transla- flect the OSGeo umbrella function and a source code tions, new functionality, bug reports, fixes and useful review with related changes to the code has been instructions to the project as a whole. done. The GeoNetwork team has dedicated efforts Building a self sustaining community of users to develop a DVD containing the best free and open and developers is one of the biggest challenges for source software in the field of Geoinformatics. This the project. This community building process relies DVD has been printed and distributed in hard copy on active participation and interaction of its mem- to over three thousand people and is also available bers. It also relies on building trust and operating in through the Internet. a transparent manner, thereby agreeing on the overall objectives, prioritization and long term direction of the project. A number of actions have been taken by the project to facilitate this process. A work plan is developed and presented at the yearly GeoNetwork workshop. This plan is maintained and updated throughout the year where needed. The foundation for the establishment of a GeoNetwork Advisory Board was laid at the 2006 workshop in Rome and membership was defined. The project management team reports back to the advisory board during the GeoNetwork workshop. Two public websites have been established. One focuses on the users of the software, the other focuses on the developers. Both can be updated and maintained online by trusted members of the com- Figure 2: GeoNetwork opensource version 2.0. Shown munity. They provide documentation, bug reporting is a screenshot of a metadata record. and tracking, WIKI pages et cetera. Source code is maintained in a publicly accessible code repository, hosted at an independent ser- Technology vice provider, SourceForge.net that hosts thousands of FOSS projects. Developers and users have full ac- As discussed, the GeoNetwork opensource has been de- cess to all source code, while trusted developers can veloped to connect spatial information communities make changes in the repository. Standardizing doc- and their data using a modern architecture, which is umentation development is the next challenge taken at the same time powerful and low cost, based on the up by the project to ensure versioning and multiple principles of Free and Open Source Software (FOSS) output formats (e.g. HTML and pdf) are supported. and International and Open Standards for services A small part of the community connects through and protocols, among others from ISO/TC211 and Internet Relay Chat (IRC) on a public #geonetwork the OGC. The GeoNetwork opensource architecture, in channel. But most interaction takes place on the user fact, is largely compatible with the Geospatial Portal and the developer mailing lists. Subscription to these Reference Architecture, which is the Open Geospa- lists is open to anyone interested. The archive of the tial Consortium (OGC) Guide to implement a stan- mailing lists provide an important resource for users dardized geospatial portal. and can be browsed online. A special mailing list To deploy a full Spatial Data Infrastructure, a has been established to monitor changes in the code range of related software packages can be imple- repository. This so-called commit mailing list deliv- mented in addition to GeoNetwork opensource and the ers change reports by email to its subscribers. embedded InterMap opensource. Within the range During the 2006 workshop, the project advisory of FOSS solutions, these include Web Map Server board decided to propose the GeoNetwork opensource software like MapServer, GeoServer and MapGuide project as an incubator project to the newly founded Open Source; GIS desktop software like GRASS, Open Source Geospatial Foundation (OSGeo). This OSSIM, Quantum GIS, uDIG and gvSIG; Web Map

ISSN 1994-1897 18 OSGeo Journal GeoNetwork Open Source Vol. 2, August 2007

Viewers like MapBender, MapBuilder and OpenLay- active functionality has been implemented that pro- ers. vides easy access to the system conﬁguration pages dealing with site settings, catalog harvesting and scheduling and maintenance.

Conclusion

This article provided an insight in the history, project structure and technologies behind the GeoNetwork opensource Geospatial Information Management Sys- tem. It only touched the related FOSS components that are required to implement a full Spatial Data In- frastructure and that are part of the OSGeo Founda- tion and the wider Geoinformatics FOSS community. The future of the project looks bright, receiving a wider attention every day. The biggest challenges will be in further community building and in keep- ing focus on the original goals of the project. At the same time, more demanding and challenging re- Figure 3: GeoNetwork 2.1 preview. quirements have to be met and new technologies have to be used where they provide significant ben- The software uses an approach that separates efits to the project. content and format as much as possible from the in- The system will have to be split into well defined ternal business logic. It therefore uses XML encod- server and client components in order to allow third ing for much of its internal information and data parties to integrate components into their existing handling. Using Stylesheet transformations (XSLT), evironments. A plug-in mechanism has to be devel- XML content is modified to provide the required out- oped for key functions in the system, like language put at different levels in the application. Metadata support, custom user interfaces as well as metadata is handled in this way and is stored in its original application profiles. It is likely that the use of AJAX format, giving full flexibility when manipulating its will be taken further in the future releases, includ- content as well as format. ing most importantly in the online metadata editor. A generic framework has been developed in Java. This poses exciting technological challenges result- Web services are exposed to the network through a ing in more intuitive and responsive user-system in- servlet engine like Apache Tomcat or Jetty and pro- teraction. duce among others HTML and XML output. The A large user and developer community will al- framework provides all the required functionality to low the project to flourish and advance the devel- deal with XML, XSLT, database connections through opment of interoperable global spatial data infras- JDBC, Internet protocols, user sessions and file sys- tructures. Taking the Free and Open Source Soft- tems. Custom web services can be developed using ware development approach results in comprehen- existing functionality or by writing small extensions. sive Geospatial Information Management Software Recently, the use of AJAX techniques in the web that can be used by any interested party as a straight- interface has started. By splitting existing services forward and cost-effective means of publishing geo- into smaller services, the user can now interact with graphical metadata and data on the web. the system in a more interactive manner. A comprehensive search interface has been build without Jeroen Ticheler and Jelle U. Hielkema (rtd) compromising on speed. This has allowed to in- Environment, Climate Change and Bioenergy Division tegrate the existing web map viewer in the home (NRC) page. In earlier versions this viewer was available Food and Agriculture Organization of the United Nations as a pop-up window, causing confusion to users. In (FAO)27 the administrative part of the system, similar inter- UN Geospatial Information Working Group 27Food and Agriculture Organization of the United Nations (FAO): http://www.fao.org/geonetwork 28UN Geospatial Information Working Group (UNGIWG): http://www.ungiwg.org

ISSN 1994-1897 19 Project Spotlights OSGeo Journal Vol. 2, August 2007

(UNGIWG)28 Jeroen.Ticheler AT fao.org GeoNetwork opensource29 jellefroukje AT gmail.com

29GeoNetwork opensource: http://geonetwork-opensource.org

ISSN 1994-1897 20 OSGeo Journal FAO GeoNetwork in a Multinational Development Programme Vol. 2, August 2007

Case Studies

FAO GeoNetwork in a Multinational Development Programme

The Case of the Programme Against African Try- able on the FAO geospatial portal, thus complement- panosomiasis ing other existing dissemination channels and media (i.e. PAAT web site, CD-ROMs, FAO/PAAT statu- by Giuliano Cecchi and Raffaele C. Mattioli tory meetings). More recently, GeoNetwork was included in the agenda of PAAT training activities. In this context, selected PAAT partners from affected Abstract countries were trained on the basics of GeoNetwork to foster a more effective sharing and documentation The Programme Against African Trypanosomiasis30 of data in the context of ongoing multinational inter- - Information System (PAAT-IS) is hosted by FAO vention projects. The potential of such a tool in the and it produces, collates, stores, analyses, and dis- management of trypanosomiasis and other vector- seminates information on tsetse, trypanosomiasis borne and transboundary animal diseases is briefly and related sustainable agriculture and rural devel- discussed and progress and setbacks in its dissemi- opment for the benefit of 37 trypanosomiasis af- nation to PAAT partners in developing countries are fected countries of sub-Saharan Africa and of in- outlined. ternational FAO/PAAT partners. In 2006 PAAT-IS began to utilize the infrastructure and functionalities of FAO GeoNetwork to improve the dissemina- The Programme Against African tion of PAAT-IS spatial datasets, mainly small scale Trypanosomiasis (PAAT) predictive distribution maps of different tsetse fly species, the vector of the disease. Standardized meta- Tsetse-transmitted trypanosomiasis (also called data were created and key GIS datasets made avail- sleeping sickness in humans) is an infectious dis- 30PAAT is an international, inter-agency alliance aiming at harmonizing and coordinating the activities of the four mandated organizations (AU/IBAR, FAO, IAEA, WHO) in relation to tsetse-trypanosomiasis and related sustainable agriculture and rural development in affected areas of sub-Saharan Africa. (http://www.fao.org/ag/paat.html)

ISSN 1994-1897 21 OSGeo Journal FAO GeoNetwork in a Multinational Development Programme Vol. 2, August 2007 ease unique to Africa, caused by various species of FAO GeoNetwork for PAAT-IS trypanosomes. The disease affects both people and animals and occurs in 37 sub-Saharan countries cov- FAO, hosting the PAAT-IS, utilized and promoted ering more than 9 million km2, an area which corre- GeoNetwork opensource31 among the partners and sponds approximately to one-third of Africa’s total beneficiaries of the Programme. In a first phase, stan- land area. The infection threatens an estimated 60 dardized metadata (ISO-19115) for key PAAT spatial million people and about 50 million head of cattle. datasets were created and posted on FAO GeoNet- 32 PAAT is an international forum which guides work . The first datasets to be uploaded were the and assists the trypanosomiasis affected countries predicted distributions of tsetse fly, at 5 km resolu- in the development of policies and strategies aimed tion for the whole sub-Saharan Africa and at 1 km at the implementation of integrated intervention in resolution for some priority areas of intervention. identified priority areas. PAAT brings together in- Figure 1 shows two of the PAAT-IS datasets now ternational institutions [African Union (AU), Food available through FAO GeoNetwork. and Agriculture Organization of the United Nations From an operational standpoint, the choice of an (FAO), International Atomic Energy Agency (IAEA) open source solution and of FAO GeoNetwork of- and World Health Organization (WHO)], other UN fered several advantages. By using FAO GeoNet- Agencies, international, regional and national re- work, PAAT-IS did not need set up a new hardware search institutes, extension services and scientific and software infrastructure to target a large com- and technical staff from affected countries. munity of GIS data users. Access to the existing FAO platform was granted to PAAT-IS, which also benefited from technical assistance of FAO GeoNet- work staff on all matters related to the user interface and metadata creation. GeoNetwork administration PAAT - Information System (PAAT- functionalities allowed to create a new group of users labelled “PAAT-Information System”, administered IS) and the use of GIS datasets by the PAAT-IS manager. The latter was registered in the system with the privileges of “User adminis- trator”. This profile allows the PAAT-IS manager to PAAT-IS aims at generating, harmonizing and shar- create new users within their group; PAAT-IS users ing technical and scientific information among PAAT will be able to access dataset that may not be visible partners and stakeholders. GIS datasets and tools or downloadable for occasional (i.e. non-registered) constitute important components of PAAT-IS and users. This hierarchical policy of users’ management more emphasis has been placed on them during the allows to restrict access to sensitive information to recent system upgrade. a predefined group of authorized people. At the Geospatial datasets are used in several phases of present stage, no visualization or downloading re- the tsetse and trypanosomiasis decision making pro- striction is applied to PAAT-IS datasets in GeoNet- cess. Global datasets (i.e. continental level datasets) work, with the exception of test datasets or datasets are used by international agencies and donors to “under development”. identify priority areas of intervention in relation In a second phase, PAAT-IS introduced the us- to feasibility and potential benefits of tsetse/try- age of FAO GeoNetwork to key partners in trypanosomiasis control or elimination projects. Key panosomiasis affected countries, which have em- datasets at the continental scale are the predicted barked on a tsetse-trypanosomiasis elimination ini- distributions of tsetse species, maps of the agro- tiative at continental level, the Pan African Tsetse and ecological potential and poverty maps. Higher res- Trypanosomiasis Eradication Campaign (PATTEC). olution datasets, normally at the national or subna- PATTEC is structured on several national projects tional level, are needed to plan, implement, moni- which are expected to address the transboundary na- tor and evaluate interventions in the field. Typical ture of the problem in a coordinated manner. All na- spatial datasets used in this context are field-based tional projects established GIS units to support en- entomological and parasitological data, land cover tomological mapping, land use planning and strate- maps, digital elevation models and medium resolu- gic decision making. GeoNetwork opensource could tion satellite imagery (e.g. Landsat, Spot, etc.) play a role in storing, documenting and sharing the 31GeoNetwork opensource: http:/geonetwork-opensource.org 32FAO GeoNetwork: http://www.fao.org/geonetwork

ISSN 1994-1897 22 OSGeo Journal FAO GeoNetwork in a Multinational Development Programme Vol. 2, August 2007

Figure 1: Two sample datasets produced by PAAT and also available through FAO GeoNetwork. geospatial information in this type of multinational Interactive mapping in FAO project. During an interactive training workshop held in Rome, three selected GIS specialists from GeoNetwork Burkina Faso, Ghana and Uganda received train- In addition to the on-line publication of metadata ing on several aspects of information management and the download of GIS datasets, FAO GeoNetwork and GIS-based decision making, including GeoNet- offers to authorized users the possibility to create work. They were provided with FAO GeoNetwork Web Map Services (WMS), a standard of the Open accounts as “Editors” within the PAAT-Information Geospatial Consortium (OGC) which allows to visu- System group, thus allowing them to create and up- alize, overlay and query maps interactively, either load their own test metadata. Feedback was positive through a GIS application or directly from the web and the potential of the tool was widely recognized. browser (through InterMap33, an interactive map Still, in the aftermath of the workshop, the usage of viewer). The software used by FAO to create the the system failed to gain momentum and follow-up WMS is MapServer34. actions such as completion of test metadata and further dissemination lag behind. This can be ascribed PAAT-IS set up WMS for the land cover maps of to shortcomings in the management of the national eight trypanosomiasis-affected countries, all derived projects in general, and of the information systems from the FAO Africover databases. The maps are tai- in particular. Further training and pro-active actions lored to tsetse habitat mapping and they were cre- by FAO will probably be needed in this field for last- ated through the process of thematic aggregation of ing assimilation of data sharing approaches to be the original full-resolution, multipurpose databases. achieved. Such a customization took advantage of the inherent flexibility of the Land Cover Classification System35 (LCCS), upon which Africover maps are based. 33InterMap viewer: http://sourceforge.net/projects/intermap 34MapServer: http://mapserver.gis.umn.edu/ 35Land Cover Classification System: http://www.glcn-lccs.org/

ISSN 1994-1897 23 OSGeo Journal FAO GeoNetwork in a Multinational Development Programme Vol. 2, August 2007

Figure 2: Web Map Service of the land cover of Kenya customized for tsetse habitat mapping, visualized through the FAO GeoNetwork interactive map viewer “InterMap”.

ISSN 1994-1897 24 OSGeo Journal The West Alexander Neighbourhood in Winnipeg Vol. 2, August 2007

The dissemination channel of Africover products PAAT - Information System: http://www.fao.org/ is a dedicated web-site36; in particular, access to the ag/paat-is.html full resolution datasets requires authorization from FAO GeoNetwork: http://www.fao.org/ the National Focal Point Institutions. In addition geonetwork to this preferential dissemination channel, an aggre- gated multipurpose version of Africover is available Giuliano Cecchi as WMS through FAO GeoNetwork. PAAT-IS fol- Food and Agriculture Organization of the United Nations lowed this example to ﬁrst introduce the technology (FAO), Animal Production and Health Division of web map service to the tsetse and trypanosomi- Viale delle Terme di Caracalla, 00100 asis community. WMS is one of the most funda- Rome, Italy mental OGC’s interoperability standards that PAAT Environmental engineer can use to allow its partners to access datasets re- He is currently managing the Information System of motely through a variety of applications. PAAT the Programme Against African Trypanosomiasis (PAAT) partners could also use WMS to display datasets and supporting FAO in the ﬁeld of spatial analysis of ani- for which direct download might not be advisable. mal diseases. FAO GeoNetwork already provides a robust hard- giuliano.cecchi AT fao.org ware and software infrastructure that, through the support of PAAT, could be used to these purposes. Raffaele Mattioli Doctor of veterinary medicine, PhD More info He has more than 20 year experience in veterinary epi- demiology in developing countries; since 2000 he is in Programme Against African Trypanosomiasis: charge of the PAAT. http://www.fao.org/ag/paat.html raffaele.mattioli AT fao.org

The West Alexander Neighbourhood in Winnipeg

A Mapserver Case Study In the end, the public was able to discover that an area that is often described as “run down” is rich by David A. Enns in history both with people and buildings. Our goal was achieved by creating a public website offering interactive mapping tools. This web mapping site Abstract allows some basic interactive tools such as zoom in / out, pan, etc, and one of my favourites, the infor- West Alexander, a neighbourhood located in Win- mation tool which brings back detailed information nipeg has been facing ongoing destruction of afford- about the parcel clicked on, along with a photograph able housing making way for the ever expanding of the area. nearby hospital and community college. The issue at hand was how to increase public awareness of this situation before complete sections of a neighbourhood go missing and to determine how people liv- Technology: ing in the community view their neighbourhood as opposed to those outside the neighbourhood. In a The ﬁrst step in this project was obtaining the photo- sense we deﬁned our goal as a social GIS project. It graphic record of each parcel within this neighbour- required the gathering of several data sources, some hood, and then tying these photos to base data that purchased and some supplied by the City. The soft- was obtained from the City of Winnipeg. This base ware used to display this data and make it accessible data included parcels, roads and hydrography. The to the public is MapServer/Chameleon. 20cm imagery (1 tile) was purchased just for the West 36Africover: http://www.africover.org

ISSN 1994-1897 25 OSGeo Journal Navigating with Innovation Vol. 2, August 2007

Alexander area. increasing redraw times on the map. We chose to Once we had all the data and photos in place, it use some large imagery which had to be resampled was time to create an interactive mapping website. several times until I got a good cross between qual- The technology chosen was based on Open Source ity and performance. In terms of development time, mapping applications. For the web map portion, the most time consuming part of the project was the I used MapServer, an internet mapping application cleaning up of layers along with the photos. The link- and for developing the maps with widgets, etc. I ing of the photos took some time, but I had ﬁgured used Chameleon. that out in a previous project. The beauty of this tech- Initially we had looked at some commercial prod- nology is that when you create your template/wid- ucts, but software cost was an issue and our client gets, they can then be re-used. had no existing GIS software so we were free to choose. In this case, I chose open source tools such as MapServer and Chameleon because of the power In Summary: and stability available in these products. It was actually the advantages of using these two tools that For anyone wanting to start a similar project, I would was the deciding factor in the choice to go with recommend these tools, as there are several books open source. I especially found the online forum available on this topic and plenty of good people and the community to be very useful with quick willing to help on the forums. It was really the en- turnaround times on questions I had. I consulted couragement and goodwill help from others in the both the MapServer IRC channel and the MapServer open source community that made this project suc- forum frequently and found that people were very cessful. friendly and more than willing to help me get on the The site37 is open to the public so hopefully cit- right track, even when I had basic Linux questions izens of Winnipeg and other cities across the world that related to MapServer. can look at our example and implement the same In the beginning it can be quite difﬁcult, not be- idea in their neighbourhoods. cause of the tools, but because it is a new way of thinking. For someone used to commercial products, David A. Enns moving into the open source arena had its learning Map It Out, Inc. curve. Perhaps the most important, but not hardest, http: // wa. mapitout. com/ issue was the getting the look and feel (GUI) plus info AT mapitout.com

Navigating with Innovation

Laboratòrio de Computação Aplicada sels Tracking Program) in Brazil. Given the impor- Grupo de Estudos Pesqueiros tance of this measure, it is relevant to consider its Universidade do Vale do Itajaì meaning and potential consequences on the development of the ﬁshing activity. by Rafael Medeiros Sperb, Roberto Wahrlich, and Marcelo Rodrigues Ribeiro Background

Introduction The first experiences of tracked fishing vessels via satellite on Brazilian territorial waters occurred in The recent Conjunct Normative Instruction SEAP- 2000, when the Government demanded the tracking PR/MB/MMA n. 02, from September 4th, 2006, is of foreign vessels under lease. By that time, tracking a significative step towards Responsible Fishery in technology was already widely used in the freight Brazilian waters, as it establishes the Programa Na- transport on highways, as well as in some sectors cional de Rastreamento de Embarcações Pesqueiras related to security in urban and rural areas. Being por Satèlite - PREPS38 (National Satellite Fishing Ves- an entirely new application in Brazil, performance 37West Alexander web mapping site: http://wa.mapitout.com 38http://www.preps.gov.br/

ISSN 1994-1897 26 OSGeo Journal Navigating with Innovation Vol. 2, August 2007

Figure 1: West Alexander Housing Study was disappointing due to diversity of the systems for geospatial support who‘s goal was the integration of sending, delivering, and visualizing data: each track- tracking data from different companies that provide ing company introduced variability according to the this sort of service. The Normative Instruction men- system employed. There was no technological infras- tioned above was elaborated based on this version tructure to allow the simultaneous receipt of data on and on the experience obtained from five years of the position of the vessels, unless they all employed monitoring the leased fleet. service by the same tracking company. There was no legal basis to assure the reliability and integration of a tracking system for the fishing sector.

The first point worth mention in the newly- created PREPS is the combination of interests and ju- risdictions among SEAP-PR, the Brazilian Navy, and the Ministry of the Environment. This combination was designed in order to avoid conflicts of interests and waste of financial and human resources. At the same time, it intends to increase the efficiency of the management and control of commercial fishery. With this perspective, there is the first innovative feature was promotion of shared management of the fishing activity.

Figure 1: The operational model of the tracking system‘s ﬁrst version

This was the picture in which the Grupo de Estu- In addition Federal Government regulations re- dos Pesqueiros - GEP39 and the Laboratòrio de Com- quire that vessels with a gross tonnage above 50 tons, putação Aplicada - G1040, both from the Universi- or greater than 15 meters must adhere to PREPS. dade do Vale do Itajaì, developed the ﬁrst version Their part is deﬁned in this process, and it is the sec- of Rastro ( 1), a Web-based information system with ond innovating point of the NI. 39http://www.preps.gov.br/ 40http://www.univali.br/g10

ISSN 1994-1897 27 Case Studies OSGeo Journal Vol. 2, August 2007

petition adjusts quality and cost of service, with positive impacts for PREPS and the productive sector. In order to maximize the benefit from the established uniform data interface, the Government invested in the development of a tool for receiving and integrating telemetry data. The system employs state-of-the-art technologies and architecture, allowing the three organizations involved in PREPS to act, within their responsibilities, from one system only. This may be considered the fourth innovating aspect: establishment of an information system to support plans and decisions operating in a shared environment. The same information may be used, for example, by the Ministry of the Environment to inspect specific areas; by SEAP/PR to control a certain fleet‘s operations; and by the Brazilian Navy, to find vessels at any moment of threat to the security on the sea. Finally, the fifth innovative element is the track- Figure 2: The first interface ing system‘s interoperability with other governmen- tal systems. The concept of the system developed for The government‘s guiding principle is that this PREPS permits exchange of data with a system for segment‘s biggest interest is the maintenance, in managing the subvention for diesel, and with a com- long terms, dos potenciais de captura dos estoques puterized system for managing fishing permissions, pesqueiros existentes. Not by coincidence, this is also among other applications. the Government‘s interest. Therefore, PREPS for- mally establishes the sector‘s responsibility related to its performance, for example, when recognizing fishing areas and providing data on catch associated to Conclusion these areas. These elements are essential when dealing with fishing resources. In implementing PREPS, Both the Grupo de Estudos Pesqueiros - GEP41 and the Government makes the private enterprise co- the Laboratòrio de Computação Aplicada - G1042 are responsible for the future of fishery on national wa- satisfied with the results from this process of techno- ters. logical development that began 6 years ago and that The third point addresses pioneering technolog- will not be limited to the University‘s walls from now ical innovation. By choosing the web-based service on. The results present a clear example of academic to implement tracking, Tracking Central of PREPS advances that are successfully employed to produce made it possible for any company willing to provide and to transmit knowledge that transforms the soci- tracking to offer and integrate its data. Each service ety. provider must only adhere to the format established by the NI SEAP/PR n. 20, from September th, 2006. Rafael Medeiros Sperb, PhD As a consequence of this standardization, the Laboratòrio de Computaçõo Aplicada - G10 Government benefits by having conditions to inte- Environmental Engineering graduation course grate tracking data sent by different tracking compa- Universidade do Vale do Itajaì - Univali nies. And the tracking companies, when answering Centro de Ciências Tecnològicas da Terra e do Mar the technical requirements set by the Government, Rua Uruguai, 458, Centro - Cx.Postal 360 establish themselves in a free market, unlike the CEP 88302-202 - Itajaì/SC monopoly that other countries have experimented Telephone number: +55 (47) 3341-7960 with this type of service. Accordingly, market com- rsperb AT univali.br

41http://siaiacad04.univali.br/ 42http://www.univali.br/g10

ISSN 1994-1897 28 OSGeo Journal Spatial Relationships in GIS – Geospatial Topology Basics Vol. 2, August 2007

Topical Studies

Spatial Relationships in GIS – Geospatial Topology Basics

Landon Blake that remain invariant under certain transformations, as bending or stretching.” If you wanted to translate this definition to a simpler, more common lan- Introduction guage you might say that topology is “the study of the relationships between shapes or geometries that This article provides the reader with an introduction do not change when the shapes or geometries are to geospatial topology. It is a continuation of the arti- subjected to a common transformation or manipula- cle in the last issue of the OSGeo Journal, which pro- tion.” In GIS, topology can be considered a special vided the reader with an introduction to spatial rela- type of spatial relationship. It is a spatial relation- tionships in GIS. This article is intended for a reader ship that does not depend on the coordinate geom- with a basic knowledge of GIS, and does not cover etry of the shapes or geometries participating in the advanced concepts. To get the most benefit out of relationship. this article the reader should [1] understand what a “feature” is in the context of a GIS, [2] understand A short example will help us to understand the how features are commonly represented by vector definition of geospatial topology. Consider two line geometry types like points, lines, and polygons, and segments that intersect one another. You could apply [3] have a basic knowledge of geometry and coordi- several transformations to both lines, such as a trans- nate systems. lation, rotation, scaling, or warping, and the lines will still intersect. The fact that the two line segments intersect does not change if the transformations are What is topology? applied to both line segments in the same manner. We also do not need to know anything about the co- The word “topology” has several different defini- ordinate geometry of the two lines to understand this tions. The Random House Webster's College Dictionary particular spatial relationship. Therefore the inter- provides one definition of topology as “the mathe- section of the two lines is a topological relationship. matical study of those properties of geometric forms This is different from the typical type of spatial

ISSN 1994-1897 29 OSGeo Journal Spatial Relationships in GIS – Geospatial Topology Basics Vol. 2, August 2007 relationship that we discussed in the ﬁrst article on this topic. As an example, the interior angle formed by two line segments will change if the two line segments are warped or stretched. This is an example of a spatial relationship that is not topological.

Article Scope

In this article we will deal specifically with the topology that can exist between features modeled with “vector geometries”. For the purposes of this discus- From this example we can see that containment sion we de[FB01?]ne “vector geometries” as shapes and adjacency relationships can usually be expressed that can be expressed using distances, angles, or co- in a bi-directional manner. Thus, if Parcel 1 is con- ordinates. The shape and location of features in a GIS tained by Parcel 3, Parcel 3 must by necessity contain are commonly described in two dimensions, but may Parcel 1. be described in only one dimension, or in three or more dimensions. In this article we will concentrate on the topology that can exist between real world ob- Network Topology jects represented in only two dimensions. We will only discuss two types of geospatial topology in this The second basic type of geospatial topology that we article. The first type will be containment and adja- will discuss is known as network topology. Network cency. The second type will be network topology. topology represents the relationships between segments in a linear network, or a collection of line segments. (This could be straight line segments, curved lines segments, or both.) There are 3 basic relation- Containment & Adjacency Topol- ships between line segments that are important in ogy network topology. 1. The end point of one line segment can connect The first basic type of geospatial topology that we to the end point of another line segment. This will discuss is containment and adjacency. This type connection is typically called a node. of topology considers two similar types of spatial re- 2. A line segment can intersect a node, but not lationships. The first is that one feature geometry connect to that node or the line segments joined may contain, or be contained by another feature ge- by that node. (This often takes place when ometry. The second is that one feature geometry my a line segment passes “above” or “below” a be next to, alongside, or adjacent to another feature node. For example, a railroad track might cross geometry. over the junction of two streams.) An example will help us to understand contain- 3. A line segment can intersect another line segment and adjacency topology. Diagram 001 shows ment. a group of polygons that could be used to represent 4. A line segment can terminate at some point parcels in a GIS. Consider the containment and adja- along another line segment. Typically this ter- cency topology evident in this diagram. mination point is also represented with a node.

An example will help us understand network • Parcel 1 is contained by Parcel 3. topology. Diagram 002 shows a group of line seg- • Parcel 3 contains Parcel 1. ments that could be used to represent a road network • Parcel 3 is adjacent to Parcel 2. in a GIS. The end point of each line segment is rep- • Parcel 2 is adjacent to Parcel 3. resented by a node. Each node has been numbered • Parcel 4 is not contained by, does not contain, in the diagram for the purposes of our discussion. and is not adjacent to Parcel 1, Parcel 2, or Par- Let's consider some of the network topology that is cel 3. evident in this diagram.

ISSN 1994-1897 30 OSGeo Journal Spatial Relationships in GIS – Geospatial Topology Basics Vol. 2, August 2007

the spatial relationship between feature geometries. Geospatial topology allows us to do this in some special ways. It can be used to verify that the geometries of two or more features participate in a topological relationship in a way that does not violate the rules or restrictions set up for that particular relationship.

Let's consider a couple of examples.

Consider the containment and adjacency topology that was evident in Diagram 001. In this diagram we see parcels of land or real property represented by two-dimensional polygons. We discussed how these 1. Node #2 and Node #4 connect to line segments polygons can contain other polygons, and how they to one another. Node #3 connects 3 line seg- can also be adjacent or next to other polygons. When ments, while Node #1, Node #5, Node #6, and modeling parcels of land, or a “parcel fabric” as it is Node #9 are present at the end point of line seg- sometimes called, it is helpful to set up some basic ments, and represent the termination points of rules that govern how the polygons used to repre- the network. sent the parcels can be related when it comes to con- 2. The line segment between Node #8 and Node tainment and adjacency. We typically do not want #9 crosses over, but does not connect to the line adjacent parcels to overlap, nor do we want to have segment between Node #1 and Node #2. a gap between adjacent parcels. This is a rule that we 3. The line segment between Node #10 and Node can set up and verify using geospatial topology. We #11 terminates at a point on the line segment might also set up a rule about the number of land between Node #8 and Node #9. This point is parcels that can be contained in any single assess- represented by Node #11. ment district. (In this case the tax assessment district would also be represented by two-dimensional poly- Relationships in network topology can be ex- gons.) Because we are talking about containment, we pressed in a bi-directional manner, as the relation- can set up a rule and verify it for this scenario using ships in containment and adjacency topology can. containment and adjacency topology. Thus, if the line segment between Node #1 and Node #2 connects to the line segment between Node #2 and Node #3, then the line segment between Node #3 and What about network topology? How can it be Node #3 must also connect to the line segment be- used to verify data quality in a GIS? Consider for tween Node #1 and Node #2. a moment a network of sanitary sewage pipes and the manholes that connect them. What type of rules could we set up for this type of utility system mod- Practical Application of Geospatial eled by a series of straight line segments and nodes? (A straight line segment would represent a sewer Topology pipe, while a node or a point would represent a manhole.) We could define a direction of flow for each What is the practical application of geospatial topol- manhole and then require the manhole at one end of ogy in a GIS? Like the more basic type of spatial rela- a line segment to be represented by a node with an el- tionships, topology can be used to aid in spatial anal- evation that is higher or greater than the elevation of ysis. For example, topology can be used to greatly the node representing the manhole at the other end increase the speed and efficiency of the operations or of the line segment. We might also have a rule that tasks that must be performed as part of spatial anal- indicates only line segments belonging to a sanitary ysis. But in this article we will consider another prac- sewer segment can connect to a sewer manhole, and tical application of topology that allows it to be used that line segments representing other types of pipes in situations where typical spatial relationships just would not be allowed. We might also dictate that won't do what we need them to. a pipe leaving a manhole be at least the same size Spatial relationships can also be used to a tool or bigger than the pipe entering the manhole. All of for data quality assurance. This is done by plac- these things could be verified by setting up rules that ing g a constraint or a check one a unique aspect of take advantage of network topology.

ISSN 1994-1897 31 Topical Studies OSGeo Journal Vol. 2, August 2007

Conclusion "Towards Usable Topological Operators at GIS User Interfaces" Catharina Reidmann http://www. In this article we deﬁned geospatial topology. We agile-secretariat.org/Conference/greece2004/ took a brief look at the two main types of topol- papers/8-1-3_Riedemann.pdf ogy, containment and adjacency topology and net- "Simple Features For OLE/COM" Open Geospatial work topology. We also brieﬂy examined a practical Consortium Inc. http://www.agile-secretariat. application of geospatial topology, a tool to ensure org/Conference/greece2004/papers/8-1-3_ data quality in a GIS. In a future issue of the OS- Riedemann.pdf Geo Journal we will examine how we can use the open source Java Topology Suite to construct two- "Point-Set Topological Spatial Relations" M. Egen- dimensional geometries used to represent features, hofer and R. Franzosa International Journal of Geo- and then set up constraints on those geometries to graphic Information Systems ensure data quality. Landon Blake sunburned.surveyor AT gmail.com References

The following references were suggested for inclu- sion in this article:

ISSN 1994-1897 32 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007

Programming Tutorial

Porting a GRASS raster module to distributed computing Examples for MPI and Ninf-G Introduction

Satellite image processing plays a vital role for re- Shamim Akhter, Yann Chemin, Kento Aida search developments in Remote Sensing, GIS, Agri- culture Monitoring, Disaster Management and many other fields of study. However, processing those Abstract even increasing spatial resolution satellite images require a large amount of computation time due to its complex and large processing criteria. This seems Satellite imagery provides a large amount of use- a barrier for real time decision making. Distributed ful information. To extract this information and un- computing can be a suitable solution to complete the derstand them may require huge computing power job timely. Cluster and Grid are two well-known dis- and processing time. Distributed computing can re- tributed systems have been associated with high per- duce the processing time by providing more compu- formance computing of massive CPU bound appli- tational power. GRASS, an open source software, has cations. GRASS GIS (Neteler et al. , 2004) is an open been used for processing the satellite images. To let source software/tool, which has been used to pro- the GRASS modules benefit from distributed com- cess satellite images. Inside GRASS, different mod- puting, an example module r.vi is ported in both ules have been developed for processing satellite im- MPI (r.vi.mpi) and Ninf-G (r.vi.grid) program- ages. GRASS module r.vi is developed by (Kam- ming modules. Their implementation methodolo- ble et al. , 2006), and is used as a test example for gies are the main discussion issue of this paper which this study. Developing the methodology, which en- will guide the basic way of representing any GRASS ables to run GRASS GIS environment for satellite im- raster module in distributed platform. Additionally, age processing on distributed computing systems, is a comparative study on modified r.vi, r.vi.mpi the main concerning issue of this paper. Addition- and r.vi.grid is presented here. ally, two different implementation methodologies for

ISSN 1994-1897 33 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007 distributed r.vi are discussed for two different pro- structure of serial running r.vi module (for simplic- gramming platforms MPI (Message Passing Interface ity only two band images have been presented). , MPI) and Ninf-G (Ninf-G Homepage. , 2007). Vegetation Index (VI) is the major set of indicators for vegetation. The GRASS module r.vi, is used to process 13 different vegetation indexes for the satellite images. NDVI (Normalized Difference Vegeta- tion Index (Weier et al. , 2007) is one of them. The NDVI is calculated from these individual measure- ments as NDVI=(NIR-Red)/(NIR+Red), where RED and NIR stand for the spectral reﬂectance measure- ments acquired in the red and near-infrared regions, respectively. Other vegetation indices are as follows: Figure 3: Serially Running r.vi Module Structure

Objectives

Under the assumption that distributed style of computing will remove computational time constraints, new image processing distributed algorithms for re- motely sensed images can be considered. Different GRASS modules have been developed for solving different Remote Sensing Image analysis problems. To evaluate the performance of the GRASS modules in distributed computing environment is the major objective of this paper. Additionally, it is a new dimension for the remote sensing users to port their jobs in distributed computing environment. Another issue addressed in this paper is to ﬁnd out the sufﬁ- cient amount of workload that is needed to dispatch among the worker nodes for getting better performance than the serial module.

Figure 1: Vegetation Indexes Methodology They are derived using various methods of dif- ferentiations and contrast. Fig. 2 shows a snapshot To fulfill the above objectives or requirements, the of an output of GRASS Software. GRASS module (r.vi) has been parallelized by using GRASS module r.vi works with raster images the master-worker model. The master process runs (rows x columns). Different band raster images are in the GRASS environment, and decomposes the tar- required for different indices. The generic indices get images in rows and dispatches the computation (NDVI, RVI, etc) use red and NIR (Near Infra-Red) of rows to multiple worker processes. Worker pro- band images. However, ARVI uses red, NIR and blue cesses are free from GRASS, they just run the com- band images, GVI uses red, NIR, blue,green, chan5 putation and send back the row wise result to the and chan 7 of landsat images and GARI uses red, master process. The module r.vi is implemented NIR, blue and green bands. GRASS functions are using MPI on a PC cluster system (r.vi.mpi) and used to extract row-wise data from the specific band Ninf-G on the same PC cluster system (r.vi.grid) images and store them in buffers. Then, each column to hold the similarity between the experimental en- value is extracted sequentially from the buffers and vironments. However, r.vi.grid module has been sent for generating the specific VI values. Thus, after structured as it is capable to run in distributed GRID completing the VI from row buffers, the row wise VI system. Additionally, experiments were done to values are put back into output image and this pro- analyse the results with distributed r.vi modules by cess will continue for each row. Fig. 3 presents the increasing the number of operations, to find out the

ISSN 1994-1897 34 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007

Figure 2: NDVI calculations with GRASS

ISSN 1994-1897 35 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007 amount of workload that is necessary to get a perfor- Speciﬁcation of the Davinchi Cluster mance beneﬁt from the distributing environments. The Cluster nodes for the experiments are as follows: In Fig. 4, the implementing structure of distributed r.vi is presented (for simplicity only two band images have been presented). Here, S1, S2, .....,Sn are different worker processes.

Figure 5: Speciﬁcations of Davinchi Cluster

Distributed r.vi Module in MPI (r.vi.mpi) For r.vi.mpi, GRASS source code is installed in master node and GRASS libraries are copied to the slaves local memory location exactly the same as master (/usr/local/grass-6.0.2/) and the following configuration steps are completed to setup the MPI- Figure 4: Distributed r.vi Module Structure (r.vi.mpi GRASS environment. and r.vi.grid) • grass.con f file is created inside /etc/ld.so.con f .d/ directory and the required grass library names are written inside this file. Implementation • Then the /sbin/ld config command is exe- cuted. MPI and Ninf-G Framework MPISend and MPIRecv both functions have been MPI (Message Passing Interface) is a library of func- used for data communication. tions (in C) or subroutines (in FORTRAN) that one The following Makefile is created for compilation can insert into the source code to perform data MODULE_TOPDIR = ../.. communication between processors (Sain et al. , CC=mpicc 1996). MPI was designed for high performance on PGM = r.vi.mpi both massively parallel machines and on worksta- LIBES = \$(GISLIB) \$(GMATHLIB) tion clusters and developed by a broadly based com- DEPENDENCIES = \$(GISDEP) \$(GMATHDEP) include \$(MODULE_TOPDIR)/include/Make/Module.make mittee of vendors, implementors, and users. default: cmd The Ninf-G has been developed by AIST (Na- tional Institute of Advanced Industrial Science and To run the code the following shell command has Technology, Japan) and TITECH (Tokyo Institute of been used: Technology, Japan). Ninf-G is a reimplementation of mpirun -np 3 location_of_exec_file parameters.. the Ninf system on top of the Globus Toolkit (Foster In this experiment the location of the executable et al. , 1997). Globus serves as a robust and com- file is: GRASSCOMPILED IR/dist.i686-pc-linux- mon platform for implementing higher level middle- gnu/bin/r.vi.mpi The pseudo-code of r.vi.mpi is ap- ware and programming tools, etc., ensuring interop- pended here: erability amongst such high level components, one of #include "gis.h" which being NinfG. NinfG system is based on client #include "glocale.h" server computing. The computational resources are #include "mpi.h" available as remote libraries at a remote computation host which can be called through the global network from client program written in existing languages /* main.c: Declare the following MPI code */ such as FORTRAN, C, or C++. /* NUM_HOSTS is total host number */

ISSN 1994-1897 36 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007

/* me is defined processor rank/number */ /* k Loop Start:

MPI_Status status; MPI_Recv(R,ncols+1,MPI_DOUBLE,k,1,\ MPI_Init(&argc,&argv); MPI_COMM_WORLD,&status); MPI_Comm_size(MPI_COMM_WORLD,&NUM_HOSTS); row_n=R[ncols]; MPI_Comm_rank(MPI_COMM_WORLD,&me); for (cn=0;cn

/* cn Loop Finished. /* Extract total row and column number */ /* and sends them to the slaves */ /* k Loop Finished.

/* Processes row put back to the result images for(i=1;i

MPI_Send(&nrows,1,MPI_INT,i,1,MPI_COMM_WORLD); for(j=0;j

/* i Loop Finished. ((DCELL *) outrast)[j] = outputImage[k][j]; G_put_raster_row(outfd,outrast,data_type_output); /* Row data are extracted from images and */ /* dispatched them among slaves with specific*/ /* j Loop Finished. /* row number by Round Robin Fashion */ /* k Loop Finished. for (r = 1; r*(NUM_HOSTS-1) <= nrows;r++ ) /* r Loop Started: /* r Loop Finished.

for(k=1;k=me) MPI_COMM_WORLD); n_rows++;

/* k Loop Finished. /* Data Receiving from master and process for(i=0;i

for(k=1;k

ISSN 1994-1897 37 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007

&status); Worker code is as follow: MPI_Recv(db,6*ncols,MPI_DOUBLE,0,1,MPI_COMM_WORLD,\ #include &status); #include #include for (col=0; col /* col Loop Started: void VI_CALC(int n, int *I, double *a,\ ndvi(db[0][col],db[1][col]);// Process ndvi() double *b,double *c, double *d, double *e,\ double *f,double *r){ /* col Loop Finished. int col; r[ncols]=I[ncols]; for (col=0; col

MODULE_TOPDIR = ../.. CC=ng_cc if((ret=grpc_initialize("/.../raster/r.vi.grid \ PGM = r.vi.grid /vi.conf")!=GRPC_NO_ERROR)){ LIBES = \$(GISLIB) \$(GMATHLIB) fprintf(stderr, "Error in grpc_initialize, \ DEPENDENCIES = \$(GISDEP) \$(GMATHDEP) %d\n",ret); include \$(MODULE_TOPDIR)/include/Make/Module.make exit(2); default: cmd } Like any other GRASS modules r.vi.grid can run by the command //creating one handle for each host node for(i = 0; i < NUM_HOSTS; i++) r.vi.grid parameters grpc_function_handle_init(&handles[i], hosts[i],\ Worker IDL ﬁle is as follow: "VI_Server/VI_CALC"); Module VI_Server; Define VI_CALC (IN int n,IN double I[n], \ for(row = 0; row < nrows; row++) IN double a[n], IN double b[n], IN double c[n],\ /* row Loop Started: IN double d[n], IN double e[n],\ IN double f[n], OUT double r[n]) Required "VI_ServerC.o" host_n=host_n%NUM_HOSTS; Calls "C" VI_CALC(n,I,a,b,c,d,e,f,r); if(G_get_raster_row(infd_redchan,\

ISSN 1994-1897 38 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007

inrast_redchan,row,data_type_redchan)<0) G_fatal_error(_("Could not read from <%s>"),\ /*Destruct the handles redchan); for(i = 0;i < NUM_HOSTS; i++) if(G_get_raster_row(infd_nirchan,\ grpc_function_handle_destruct(&handles[i]); inrast_nirchan,row,data_type_nirchan)<0) grpc_finalize(); G_fatal_error(_("Could not read from <%s>"),\ G_free(inrast_redchan); nirchan); G_close_cell(infd_redchan); ...etc...(Free memory)

for(col=0; col < ncols; col++) /* col Loop Started: Experimental Results

Fig. 6 is generated by running r.vi and r.vi.mpi /* each column cell values for all GRASS module with increasing the operation num- /* band images are extracted together bers. To calculate different vegetation indices the /* and put them in 2D arrays workload in the slaves is too small to get benefit db[0][col]= d_redchan; db[1][col]= d_nirchan; from r.vi.mpi module. Thus, the purpose of this db[2][col]= d_greenchan; experiment is to find out the sufficient amount of db[3][col]= d_bluechan; workload that will make the parallel version faster. db[4][col]= d_chan5chan; In Fig. 6, to calculate the NDVI takes only 3 op- db[5][col]= d_chan7chan; erations (subtraction, addition and then division) /* vegetation indexing array valuess are and this amount of workload takes just few seconds /* filled in I[ncols] array to execute, where the serial version takes less processing time then parallel version. However, increasing the number of operations provide better /* col Loop Finished. performance in parallel version. It is clear from Fig. 6 that to get the benefit from the parallel ver- if(grpc_call(&handles[host_n],ncols,I,db0,db1,\ sion the total workload needs to be around 300 oper- db2, db3,db4,db5, R) != GRPC_NO_ERROR){ ations or more so that the communication overhead fprintf(stderr,"grpc_call ERROR\n"); will be overpassed by the computational workload. exit(2); }

/* All outputs are put back into raster for(j=0;j

((DCELL *) outrast)[j] = R[j];

/* j Loop Finished. Figure 6: Performance Evaluation of Serial (r.vi) and MPI (r.vi.mpi) Version if(G_put_raster_row(outfd,outrast,\ When master with one worker node (slave) are data_type_output) < 0) G_fatal_error(_("Cannot write to output\ working together to solve a particular problem, the raster file")); master node will dispatch the whole job to his slave for processing. When there will be 2 slaves, the workload will be distributed among two salves. Simi- host_n++; larly for 3 slaves, the workload will be distributed to 3 slaves equally. So the time performance increase for master with 2 slaves will be 2 times than /* row Loop Finished. master with one slave and the ratio will be 3 times for master with 3 slaves with one slave and so

ISSN 1994-1897 39 OSGeo Journal Porting a GRASS raster module to distributed computing Vol. 2, August 2007 on... This would highlight a perfect parallelism.

DTT = {NR × (VDS + VDR)}/(NBW) (1)

VDS = NB × DTS × (NC + 1) (2) VDR = DTS × (NC + 1) (3) From equation 2 and 3, equation 1 has been derived as

DTT1 = (6 × 8 × (8519 + 1)) + (8 × (8519 + 1)) (4)

DTT2 = 100 × 1024 × 1024 (5)

Figure 7: Parallelization Effect of r.vi.mpi Module with DTT DTT = (7630 × 8 × DTT1)/DTT2 = 277.74Sec (6) To evaluate the parallel version performance Concerning the above philosophy and experi- more precisely, Fig. 9 has been generated. Only the menting on r.vi.mpi module, Fig. 7 and Fig. 9 have execution time (Data Transfer Time has been reduced been generated. In Fig. 7 and Fig. 9, the 2 Slaves curve from the total running time) is concerned here. Due presents the ratio of the total running time between to a constant amount of data need to transform be- master with two slaves and master with one slave. tween master and slaves, the transfer time (DTT) The 3 Slaves curve represents the ratio of the total is derived from the above equations. Fig. 9 shows running time between master node with 3 slaves that the performance in the slaves curves are improv- and master (one slave) with increasing the operation ing and nearly to the desirable values. It concludes numbers. Increasing the operations, reﬂects to in- that r.vi.mpi module is working well in distributed crease the workload in slaves and improving the per- manner. formance. However, in Fig. 7, the performances are not satisfying with the desire values (as perfect parallelism) due to the regular and high communication overhead. In equation 1, 2 and 3 the following terms have been used.

Figure 9: Parallelization Effect of r.vi.mpi module without DTT

Fig. 10 evaluates the performance of the GRASS modules (r.vi, r.vi.mpi, r.vi.grid) in the time domain. Low, middle and high, three testing work- loads have been set in the modules. When the workload is low, serial version (r.vi) is performing best because of the ﬁne-grain parallelism where communication time is bigger than the execution time. To get the coarse-grain, the workload need to be in- creased so that the communication time will be hid- den by the execution time. As the workload in- creased to 1800 and 3000 operations, parallel ver- Figure 8: Terms used and their deﬁnitions sions are performing better then the serial version.

ISSN 1994-1897 40 OSGeo Journal Bibliography Vol. 2, August 2007

ple with Grid and MPI programming for GRASS GIS distributed processing. Additionally, this type of research will merge the Remote Sensing and GIS with High Performance Computing communities.

Acknowledgements

The authors would like to acknowledge specially, to Osawa Kiyoshi San (PhD Candidate in AIDA Lab), SunHao San (Masters Student in AIDA Lab) and Nishimora Motokazu San (Masters Student in AIDA Lab) for their support to create the testbed (Davinchi Cluster) ready for this experiment. Authors also like Figure 10: Performance Evaluation of GRASS Mod- to thanks to all the members in AIDA Lab for the ules (r.vi, r.vi.mpi and r.vi.grid) moral support.

MPI version, r.vi.mpi provides the best performance among the three models. However, MPI is used mainly in Cluster Computers where the nodes are homogeneous in their speciﬁcation. Addition- ally, available nodes in Clusters are limited up to a speciﬁc range and this will be an obstacle for load balancing for big processing jobs. In this case, Grid computing infrastructure is necessary. Indeed, (the environment maintains heterogeneity as well as dis- Bibliography tributed connecting networks). So far Ninf-G is performing better than the serial version for high work load examples. Particularly, in this issue, Ninf-G is not performing better than MPI for the reason that, the communication overhead (to establish the session with remote hosts) in Ninf-G is larger than MPI. Ninf- G is specially made for Grid-Computing environment (not for running inside Cluster computing environment as have been done in above experiments). M. Snir, S. Otto, S. Huss-Lederman, D. Walker and J . Dongarra (1996) MPI: The Complete Reference. Massachusetts Insti- When the processing workload is much higher than tute of Technology. http://www.netlib.org/utk/papers/mpi- the communication workload, the real performance book/mpi-book.html. improvement with Ninf-G platform will show up. r.vi.grid I. Foster and C. Kesselman (1997) Globus: A Metacomputing In- So far, is developed and tested. In near frastructure Toolkit. International Journal of Supercomputer future, the experiment on real Grid testbed will be Applications. highlighted. M. Neteler and H. Mitasova (2004) Open Source GIS: A GRASS GIS Approach. Second Edition. Kluwer Academic Publisher- s/Springer. Conclusion Y. Tanaka, H. Nakada, S. Sekiguchi, T. Suzumarn, S. Matsuoka (2003) Ninf-G: A Reference Implementation of RPC-based Pro- Geographic Resources Analysis Support System gramming Middleware for Grid Computing Journal of Grid (GRASS) has been used for RS and GIS data anal- Computing 1: 41-51. ysis and visualization. Currently, GRASS handles B. Kamble, Y.H. Chemin (2006) GIPE in GRASS Raster Add- large datasets and the performance and capabilities ons. http://grass.gdf-hannover.de/wiki/, GRASSAddOns, of GRASS for large datasets can be greatly improved RasterAdd-ons Internet. by integrating GRASS with parallel and distributed computing. The major objective of this research was MPI(2007) http://www-unix.mcs.anl.gov/mpi/ Internet. to provide the Remote Sensing user a compact exam- Ninf-G(2007) http://ninf.apgrid.org/ Internet.

ISSN 1994-1897 41 Programming Tutorial OSGeo Journal Vol. 2, August 2007

J.Weier and D.Herring. (2007) Measuring Vegetation (NDVI/EVI) Shamim Akhter, Kento Aida http://earthobservatory.nasa.gov/Library/MeasuringVegetation/ Tokyo Institute of Technology Internet. http: // www. alab. ip. titech. ac. jp/ ~shamim shamimakhter AT gmail.com

SUN Grid Engine(2007) http://www.lesc.ic.ac.uk/projects/epic- Yann Chemin gt-sge.html Internet. yann.chemin AT gmail.com

ISSN 1994-1897 42 OSGeo Journal LizardTech: Why We Use and Support Free and Open Source Software Vol. 2, August 2007

Interview

LizardTech: Why We Use and Support Free and Open Source Software

An Informal But Earnest Manifesto Why would a small, proprietary shop like LizardTech use and support free and open source de- by Matthew Fleagle and Michael P. Gerlek, LizardTech velopment? Don’t the advance of open source development and the spread of an ethos of sharing in the One of the authors, Michael Gerlek, has been a Charter software industry represent threats to LizardTech’s Member from the beginning of OSGeo and was highly business model? We believe the contrary, and this ar- inﬂuential in its success. LizardTech ofﬁcially became an ticle aims to articulate why. First we’ll discuss some OSGeo Associate Sponsor in 2006 and has been an active of the free and open source software (FOSS) products user of open source geospatial tools for a long time. It is that we use or have used in the past, and then we’ll with this in mind that they were invited to describe how talk about why it makes sense strategically for small they use open source from both a business and community companies like ours that have proprietary code not perspective. – Editor only to use the products made by the free and open source development community but also to materi- As a small company with a niche-market cus- ally support its work. tomer base originally built on the strength and ubiq- uity of its proprietary technology – the widely used MrSID technology and image format – LizardTech How and Why LizardTech Uses has reason to cherish private ownership. The fact FOSS that we have something that no one else can use without paying us for it has, to say the least, ad- LizardTech has used free and open source software vanced our prospects over the years. Many might packages since its founding in 1992. An informal sur- be unaware, however, not only that we support open vey of LizardTech’s development team revealed just source development efforts in general but that in par- how much we rely on open source tools and libraries: ticular, we have been supporting OSGeo since its in- we quickly were able to list dozens of packages, ception. ranging from the popular (Firefox, Apache, Linux) to

ISSN 1994-1897 43 OSGeo Journal LizardTech: Why We Use and Support Free and Open Source Software Vol. 2, August 2007 the geo-specific (GDAL, PROJ4, OpenLayers). Many in larger fields without taking on the costs of find- of these libraries are built into our shipping, commer- ing and engaging additional personnel. This point cial software products. In fact, by comparison to the is again related to the first benefit, but it highlights following list we use very few closed source, third- the opportunity to widen our customer base and sell party libraries. more products to more people.

Geospatial libraries: GDAL/OGR, GeoTIFF, GEOS, OSSIM, PROJ.4, MapServer, PostGIS, Open- We can leverage well-used, well-tested code. The Layers open source development community is not three ˘ Non-geospatial libraries: Berkeley DB, CppUnit, guys in a garage âA¸Snot anymore anyway. The Crypto++, libcurl, dlmalloc, efence, Expat community actively hammers away at open source gSOAP, libgif, libjpeg, libtiff, OpenSSL, libpng, projects, putting code to the test, finding out where libungif, xalanj, xercesc, xml4c, zlib, zthread, it breaks, discovering bugs and supplying patches boost, littlecms and other solutions in a robust volunteer testing and Other tools and packages: Apache, Tomcat, gcc, feedback loop. The code evolves at the pace of Linux, Perl, Python, Putty, Firefox, Cygwin, real-world use, testing and fixing, not of marketing Wireshark, Emacs, gnuplot, siege, autoit schedules and sales imperatives. As a result, open source projects represent some of the tightest prod- Okay, you say, you’re impressed by the number ucts never sold. of free and open source software packages out there that can be used in making geospatial software, but We enjoy unplanned advantages. Not uncom- you wonder, are there benefits besides the obvious monly, a side benefit will emerge through the use of savings in license fees? By way of answer, we’ll not an open source project, some benefit that we may not only enumerate the benefits; we’ll also give a couple have had any knowledge of beforehand. Intrigued? of examples. Here are the benefits: Read onâA˛e˘

We don’t have to reinvent the wheel. LizardTech is a sleek outfit. We don’t have the resources, in Real-World Examples terms of personnel and their time, to create every- thing from scratch, especially when it would be du- LizardTech wanted to introduce reprojection to its plicated effort. If the work has been successfully popular GeoExpress product. This would give us ac- done already, it makes good business sense not to cess to a wider customer base by better synchroniz- waste time, money and sweat beating a new path ing the functionality of our product to the identifi- up the same mountain. This is good strategy even able pains of potential users. But while we are ex- if you decide to pay for the existing product. When perts in imaging technology, we are not experts in companies start up they deploy OpenOffice.Org or reprojection. Projection systems involve hard, scary license MS Word; they don’t routinely invent new math. Moreover, in many applications people’s lives word-processing programs just because they know depend on their accuracy, so they present the twin they’ll need to send inter-office memos. horrors of being simultaneously mission critical and easy to get wrong. Even if we’d had the personnel We can concentrate on our core competencies. to send off to work on reprojection, we had rathered LizardTech lacks experience in some areas that the not mess with it. open source community has plenty of experience in. Our options, then, were two. We could buy a This point is similar to the first, but it bears men- commercial, proprietary package or we could use an tioning that we save money when our own develop- open source package like the PROJ.4 reprojection li- ers can concentrate on what they do best. Diverting brary. our energies into areas where we are not experts and The second option was the hands-down smarter don’t need to be experts when the experts stand by choice, and not just because of the absense of licens- proffering their wares at little or no cost, is foolhardy. ing fees. What if the corporate product had a bug that affected our project? What if it only worked on We can play in larger arenas. Having access to in- 32-bit systems but we needed it to work on 64-bit sys- tellectual property beyond the sum of our developer tems? We’d have to put in a bug report or a feature parts means that we can equip our products to play request for 64-bit support and then wait six months

ISSN 1994-1897 44 OSGeo Journal LizardTech: Why We Use and Support Free and Open Source Software Vol. 2, August 2007 for the next release of the product, paying for a main- conferences. LizardTech is active in the open source tenance plan the while. And 64-bit support may or community in all of these ways, often as part of our may not be in the box when we finally got to peel off regular workaday world. Here’s how supporting the wrapper. open source benefits us: By contrast, with PROJ.4 we could likely get a bug fix or 64-bit support in short order. It’s If open source projects fail, we fail. It must be re- true that there is no "company" standing behind membered that as a company, LizardTech relies on the PROJ.4 package, but we’ve found (especially in open source products in order to create and improve larger projects) that this is rarely a problem. Unaf- our own products. Seen from this single perspective, flicted by corporate boardroom pressures, the open it makes sense to feed the chickens that lay the eggs, source community’s priority is finding and fixing especially when what you pay in the grand scheme bugs in its code and adding features and function- amounts to ... ahem ... chicken feed. ality to its products. (In fact, we very frequently satisfy a particular customer requests by taking an open source package and porting it to a compiler The return on investment is high. Contributing, such as SunPro on which we previously hadn’t sup- say, $5000 a year to sponsor an organization like ported that package.) If there was something seri- OSGeo, where that money helps keep open source ously wrong with open source code, then we could projects alive and vital, makes good sense when you pay someone in the community to fix it and still come compare the options that would exist in a world out ahead in terms of time and cost. without FOSS. Hiring several additional developers While PROJ.4 gave our product a very impor- would cost several times as much annually, and sup- tant and visible feature, another kind of open source porting open source, which provides us with well- product supports our product invisibly. We use tested code and quick fixes when necessary, is a GDAL, a library that supports dozens of image for- much better use of funds than paying a proprietary mats used in GIS alone, to make sure that our prod- company for software that might be more buggy and uct supports those formats. Rather than supplying harder to get fixes for. For that matter, how does one a feature or functionality, GDAL gives us access to know if a product or library is buggy? Would you the wider GIS ecosystem. More than once we have rather rely on the biased marketing messages from found out through our customers that we support a the proprietary owners of code or access firsthand format that was important to them that we otherwise experience from users and developers? Sponsorship may not have bothered with the unexpected advan- in this regard is like a different and cheaper kind of tage mentioned above. contract labor.

We can effect positive, rapid change in the prod- How and Why LizardTech Supports ucts we need to use. It has never been necessary Open Source Development for us to lobby for or engage in radically changing an open source package, but if a substantial change was It’s true that it just wouldn’t be cricket to take advan- necessary it‘s a cinch that the open source develop- tage of the now abundantly evident benefits of FOSS ment community could turn it around much quicker while refusing to support free and open source de- than a proprietary code-owner whose priorities and velopment and continuing to charge money for our schedules are beholden to the boardroom. own products. But LizardTech’s central claim in this article is not that our use of and support for open Open source conference attendance pays divi- source is moral (another paper, perhaps, submitted dends. Given finite budgets, why would we want to different journal). We’re claiming that it’s strate- to take several of our development crew to the an- gic. nual OSGeo conference instead of to one of the well- There are a number of ways an organization can attended "proprietary" user conferences? Because support open source development in material ways, not only will they get exposed to all the cutting that is, beyond simply using FOSS. They include edge technology that they’ll find at other shows, monetary donations and sponsorship; membership but they’ll also acquire knowledge they can bring and volunteering; feedback, bug fixing and patches; back and use for free. Further, the information and contracting open source developers for work, such knowledge they’ll get is divorced from any corpo- as we did for 12-bit JPEG support; and attendance at rate agenda no one will try to sell them a grand so-

ISSN 1994-1897 45 Interview OSGeo Journal Vol. 2, August 2007 lution or ply them with company-biased informa- proven that the world smiles along with you when tion. Finally, the developers they’ll meet at open you smile, it’s only a matter of time. It cannot but source conferences are more available for follow-up advance the wholly considered interests of any 21st with conversations spawned at the event. A related century company to adopt toward customers, com- way to support open source development is to do- petitors and partners an attitude that reﬂects the val- nate facilities, soda and snacks to a local user group. ues of cooperation, collaboration, community, trust LizardTech hosts a small monthly OSGeo meeting in and humility embodied by open source develop- our Seattle ofﬁces. ment.

Open source plays well with open standards. Matthew Fleagle We have also invested heavily in supporting ISO LizardTech’s technical writer and Open Geospatial Consortium (OGC) standards, http: // www. lizardtech. com which are "open" in a way that is different from the mfleagle AT lizardtech.com way open source is. We’ve found that the open standards and open source development processes are similar enough that many advantageous synergies Michael P. Gerlek emerge between them. LizardTech’s engineering teams manager OSGeo Charter Member It’s the right thing to do. We said above that this http: // www. lizardtech. com was not about morality, but if science hasn‘t yet mpg AT lizardtech.com

ISSN 1994-1897 46 OSGeo Journal Spatial Data Quality and the Open Source Community Vol. 2, August 2007

Sponsor Spotlight

Spatial Data Quality and the Open Source Community

Mike Sanderson, Graham Stickler, Steven Ramage overseeing the Web Feature Service (WFS), Web Map Service (WMS) and Geography Markup Language (GML). Then more recently through the work of the Introduction OSGeo mechanisms now exist to access geospatial data regardless of source through Feature Data Ob- The Open Source Geospatial Foundation (OSGeo) in- jects (FDO). troduces itself on its website, with the words “The As a result a whole new set of issues are cre- Open Source Geospatial Foundation has been created around spatial data quality and fitness for pur- ated to support and build the highest-quality open pose. We want to now extend the paradigm to in- source geospatial software.” It is clear the objective is clude these spatial data quality and reuse issues as to create a peer review community to improve soft- data collected for one purpose are being accessed for ware quality (1). This article addresses the quality increasingly diverse use, more often by a completely element, but from an entirely different perspective. separate entity. As an industry, Google has raised Spatial data are subject to different regimes of the industry profile to the point where it will hurt us quality management, irrespective of whether they all if data quality is suspect. are used in an open source environment or not. Our vision is to provide a set of web-based tools There will always be issues regarding data consis- to enable spatial data quality to be assessed, i.e. con- tency and integrity or fitness for purpose because formance checking. Ideally this will offer improved data are constantly changing. This may be through spatial data management in an open source environ- real world change, the introduction of new technolo- ment. It will support the goal of improving opera- gies for capture, update or organisational change tional efficiency by allowing a framework to be de- that alters the boundary conditions. veloped to aggregate spatial data. Since the mid '90s the International Standards Or- We wish to offer the OSGeo community the possi- ganisation (ISO) and the Open Geospatial Consor- bility of contributing to these initiatives via our Prac- tium (OGC) have worked on standards, creating and titioner Program and through the Open Geospatial

ISSN 1994-1897 47 OSGeo Journal Spatial Data Quality and the Open Source Community Vol. 2, August 2007

Consortium (OGC) Data Quality Working Group. and a formal approach. The flowchart in Figure 1 de- scribes such an approach for reviewing spatial data quality. This is a tried and tested approach used by Background Information 1Spatial for projects with customers and partners - it offers a route for undertaking quantitative assess- By way of supporting information it is necessary to ments. understand the work that has been carried out by the International Standards Organisation (ISO). ISO provides guidelines for putting together spatial data quality management frameworks. ISO 19113 and ISO 19114 describe various mechanisms for determining and measuring data quality. These ISO standards embody principles and evaluation procedures for geographic information (for a useful summary of these principles and procedures, see Chapter 15 in Spatial Data Quality (2). The opening sentence of ISO 19113 sets the tone: 'Geographic datasets are increasingly being shared, interchanged and used for purposes other than their producers' intended ones.' The forerunner to the current ISO standards were applied by them using the Digital Chart of the World (DCW) project as a case study. The DCW was produced in 1992 and there was an effort to work on "fit for purpose" issues. But in 1995, the initiative ran out of steam, because only non-quantitative (i.e. qualitative) assessments of the quality of geographic Figure 1: XXXX. datasets could take place. Quantitative assessments were just not possible for large geospatial datasets due to the lack of processing power available at What is important in this approach is contextual the time. It is quantitative assessment that is re- analysis of the data quality records or spatial data ally valuable for assessing logical consistency and holdings. This involves working through steps 2 to 5 positional accuracy. And as Jakobsson said so elo- using a rules-based approach to determining the con- quently: "Combining data sets that have no quality formance of spatial data against the specifications or information can be very difficult or impossible." (2) business rules. In an Open Source environment this would be accessed using an XML interface to Fea- ture Data Objects (FDO). There are two key elements to this approach. Firstly, it provides a quantitative Quantitative Spatial Data Quality assessment of spatial data quality, i.e. data conformance % against business rules. Secondly, it is an A data quality audit is designed and implemented in independent verification, i.e. not using the GIS tools order to determine the answer to the question, how that create or edit the data – these remain unchanged. bad (or good) is your spatial data? There are three high-level objectives of a data quality effort: Data quality problems are often widespread and originate in the source data itself. They can be ge- • Produce statistically valid data quality mea- ometric, topological or attribute-based. In order to surements of source or master data combat these problems, data custodians, analysts • Investigate, identify, and document leading and stakeholders must understand their source data. data quality causes (and exceptions) This understanding can come from data profiling or • Create an assessment and recommendation re- what can be thought of as self-describing metrics. port The analysts and data managers must understand how the data profile fits the business requirements. However, in order to undertake quantitative as- This is where a business rules-based audit is not only sessments, there must be some guidelines to follow useful, but critical.

ISSN 1994-1897 48 OSGeo Journal Spatial Data Quality and the Open Source Community Vol. 2, August 2007

Data Quality Rules! ering nearby objects. An initial set of spatial rules are proposed and subsequently enhanced to include In order to establish the fitness for purpose for spa- non-spatial elements such as attribute joins, equal- tial data, it is clear that it is first necessary to under- ities and inequalities, correlated to the spatial rela- stand the business rules relating to those data and tionships between the objects sampled. The final set how they should be interpreted. Conducting a busi- of rules is converted to a form that allows them to ness rules-based audit can be critical to combating be stored in a rules repository, which is based on a data quality problems. common language interface that incorporates OGC Figure 2 highlights an example of identified mea- spatial operators. sures for assessing spatial data quality. It refers to It is clear that such a tool, when combined with five key elements that change according to the type FDO and MapGuide Open Source, can enable an as- of spatial data or its application. The objective at sessment of spatial data against business rules. This this stage is to define the business rules that the data conformance checking approach can be invaluable in should obey. It is often difficult to obtain this infor- providing a quantitative assessment of spatial data mation from the user. The original data model spec- quality. ification is often unavailable or hasn’t been updated As a result 1Spatial has already commenced a since inception. Rather than start from a blank piece Practitioner Program to enable their end customers of paper a potential source of rules is the data itself. to validate spatial data across the Web by running Rules may be discovered using an analysis of domi- spatial data audits. Through remote access to Radius nant statistical patterns in the data. Studio, Practitioners are able to use the rule building The current challenge is that there are limited and assertion capabilities to carry out data certifica- toolsets available in the open source community and tion audits on behalf of customers. In return 1Spatial the wider geospatial market place as a whole to carry receives feedback on the product and 1Spatial tools out such tasks and quantitatively measure spatial are recommended for any fix up or ongoing repur- data quality. While we may look to open architec- posing, reuse or re-engineering requirements. tures and standards as defined by Service-Oriented Architectures (SOA), the World Wide Web Consor- tium (W3C) semantic Web framework and the Web Open Source Opportunity Ontology Language (OWL), it is clear that the semantic Web rules language doesn't support the geospa- There is an opportunity here for the open source tial needs. Work done recently by the OGC as part community to engage with Radius Studio to develop of their OWS-4 test bed on a Topology Quantitative this capability and make it widely available. Assessment Service (TQAS) supports this view. By lodging this assessment service within OSGeo could enhance its reach, providing the basis for cooperation between parties to allow the harmoniza- tion of geographic information to take place, and to build a community (3).The opportunity will then exist for the open source community to extend the range of FDO providers and open up more formats that can be subject to a quantitative assessment of spatial data quality. Another reason for placing an assessment service within OSGeo is to create a community that Figure 2: XXXX. will work on defining the rules expression language for creating the quality measures. The Semantic 1Spatial have been addressing this problem and Web Rules Language (SWRL) is currently not mature have developed a web-based tool, Radius Studio, enough, but as history has shown, we in the industry based around this quantitative rules-based data can create a standard. Once we have this, rather than quality paradigm. Using a variation of an artificial having to accept caveat emptor, we can all make our intelligence boosting algorithm adapted to spatial own assessments of whether the data are fit for pur- data mining, it works by initially considering a small pose. This has to be an automated assessment, and sample of objects taken at random from the data store unlike in 1995, the tools and computing power are and then works outwards from these objects, consid- now available to do this. It may need a 14-day free

ISSN 1994-1897 49 Sponsor Spotlight OSGeo Journal Vol. 2, August 2007 trial to make an assessment, if the data are not free, Biobliography but once the assessment is made, it will then be possible to decide whether or not to pay, or how to much 1. The Cathedral & The Bazaar, Raymond, E.S., to pay (value) for those data. Suddenly the free vs. (2001) ISBN 0-596-00108-8 licensing debate becomes irrelevant. 2. Spatial Data Quality, Wenzhong Shi, Pe- ter Fisher, Michael Goodchild, (2002). If the rules expression language can't be created ISBN:0415258359 fast enough, we can move to the geographer's solu- 3. Improving Operational Efﬁciency with Geo- tion, the "pseudo-quantitative expression." Follow- graphic Information, Finnish Ministry of Agri- ing the Amazon tradition of peer group review, the culture & Forestry, (2006). ISBN 952-453-301-4 user community could assess the nominal value of a spatial data set for completeness, logical consis- Mike Sanderson tency, and positional, temporal and thematic accu- CEO, 1Spatial racy. The Open Source community works well under yourhomepage the same model of peer group review and so we be- info AT 1spatial.com lieve it should be interested in this approach. Graham Stickler If you are interested in extending the OSGeo into Product and Marketing Director, 1Spatial spatial data quality assessment or playing a role in yourhomepage standards development relating to spatial data qual- info AT 1spatial.com ity we urge you to contact 1Spatial direct or through the OGC Working Group on Data Quality at: Steven Ramage Business Development Director, 1Spatial http://www.opengeospatial.org/projects/ yourhomepage groups/dqwg info AT 1spatial.com

ISSN 1994-1897 50 OSGeo Journal Vol. 2, August 2007

Local Reports

Francophone Local Chapter Update

by the local chapter members local board which has been elected at the end of june (see subsection below). Introduction During this time some small projects has been en- gaged : The creation of the Francophone Local Chapter (OSGeo-fr) started in October 2006, following the • OSGeo got a booth in an Open Source vil- FOSS4G 2006 meeting in Lausanne. As of July 2007, lage at the main geospatial tradeshow in we are close to its creation and about to send a pro- France (see http://wiki.osgeo.org/index. posal to the OSGeo Board (expected in July or Au- php/GeoEvenement2007) gust 2007). • we translated a document which presents OS- Since October, a huge work in the Francophone Geo missions and projects, mailing list 43 has been performed in order to share • we ran a poll to discover people’s motiva- ideas about the missions of the local chapter. The tions and to know whether they would be con- Francophone local chapter missions are now sep- sumers or actors in the upcoming Francophone arated into three main committees : organisation, local chapter, open geodata and open source software. For each • we initiated the translation work of the OSGeo committee, a chair will manage communication be- Journal Vol. 1. tween contributors and will send information to the local chapter about the progress of the project. Election Some other important details need to be clarified, i.e. how OSGeo-fr is organized (only the main struc- As of 6th of July, the first round of the election is fin- ture is currently drawn), how to manage and hold ished and a second round to elect the Data Commit- elections? etc. This is the main purpose of the new tee chair will start the 9th of July. The first round 43http://lists.osgeo.org/mailman/listinfo/francophone

ISSN 1994-1897 51 OSGeo Journal Activity Report of OSGeo Taiwan Vol. 2, August 2007 results are : The reasonably large amount of answers to this poll allows us to see the creation of the future Fran- • Local chapter president : Yves Jacolin cophone local chapter positively. Once this will be • Organisation Committee chair : Gwenael done, the visibility of the OSGeo-fr will be greatly Bachelot improved and we expect to see more contributors • Software Committee chair : Vincent Picavet joining us. Finally, the translation of OSGeo Journal Vol. 1 The next step aims at writing the proposal for the showed us that we are ready to manage projects and Francophone Local Chapter creation as well as writ- this will help to improve OSGeo-fr’s visibility. ing an RFC about the organization of elections and how to manage them. In other words, we need to formalize the structure of the Francophone Local Chap- OSGeo Journal Vol. 1 Translation ter in order to be efficient and start on solid bases. A french translation of OSGeo Journal Vol. 1 is, at the time of writing this article, about to be published. Poll Only two articles need to be finalized (proof read- ers). 7 persons have contributed to this project and A 15 days poll was published at the beginning of have translated 27 articles (70 pages). 5 persons have June. 54 persons answered the poll. 37 persons out proof read the translated articles. Some of them both of the 54 had subscribed to the Francophone mailing have translated and read articles. list. This is the biggest project OSGeo-fr managed by From the poll results, we realized that commu- itself so far, and some details of the process need to be nication of the OSGeo missions will be a key point clarified. Managing 27 articles in a process of trans- of the future Chapter : what is OSGeo for? Which lating, proof reading, and back to the translator for project people can be interested in? This could be finalization appeared to not be a so easy task. The one of the main purpose of the Francophone Local working process has to be clarified and a RFC has Chapter organization committee. been drafted for this purpose.

Activity Report of OSGeo Taiwan

At the 5th Taipei International Digital Earth June 29 to August 3 at Academia Sinica Computer Symposium (TIDES-07) and FOSS4G.tw Summer Center. Training 2007

Dongpo Deng A Session for Open Source GIS in TIDES-07 OSGeo Taiwan, founded on March this year, is a new and vibrant social network in Taiwan. OSGeo Tai- The FOSS4G session in TIDES-07 was conducted on wan actively participated in local GIS-related con- 18th May, during the morning session. There are four ferences and meetings to advocate open geo-data, talks and a panel discussion in this session. The aims promote Open Source GIS and introduce OSGeo to of talks are to deliver introduction of FOSS4G and Taiwanese GIS communities. OSGeo Taiwan further demonstrate how the FOSS4G works. First of all, collaborate with Academia Sinica Computer Cen- Dr. Victor T.S. Horng, general manager of OSS In- ter GIS Team, and Chinese Culture University Dig- tegral Institute Co., Ltd., gave the attendees a clear ital Earth Research Center, to organize a session and interesting summary for the beneﬁts of using for Free/Open Source Software for Geoinformatics FOSS. Even the advantages of using OSS is obvious, (FOSS4G) at the 5th Taipei International Digital Earth the intellectual property rights issue of the OSS is of- Symposium (TIDES-07) which is held at Chinese ten doubted in Taiwan GIS communities. Ms. Tung- Culture University from May 15 to 18. Moreover, Mei Ko, project manager of Open Source Software OSGeo Taiwan started to conduct Summer Training Foundry (OSSF), is familiar with legal topics of FOSS. courses for Open Source GIS on each Friday during In her talk, she delivered an overview of FOSS

ISSN 1994-1897 52 OSGeo Journal Nascent UK Local Chapter Update Vol. 2, August 2007 characteristics and the FOSS authorization model. wan should make concerted efforts to lobby the gov- Her experience is useful for encouraging the use and ernment releasing the geospatial data and providing development of FOSS4G in Taiwan. useful resources, and co-developing tutorials, techni- Mr. Hsiung-Ming Liao, initial member of OSGeo cal notes, training Taiwan and project instructor of GIS Term of Com- courses, workshops and conferences to assist puting Centre of Academia Sinica, demonstrated novices of OSGeo & FOSS4G to influence & support a web-based geospatial application constructed by local geospatial professional communities. For in- MapGuide Open Source and Web 2.0 techniques. He stance, Taiwanï£¡s Ministry of The Interior Informa- considered the development of Geo Web cannot lack tion Center promised to start releasing map data to FOSS4G and more successful experience of FOSS4G facilitate the learning and exchange among geospa- is important to FOSS4G progress in Taiwan. How- tial communities & industry. ever, the prevailing commercial GIS solutions ab- sorbed the public attention toward alternatives like FOSS4G. There is urgent need to provide more al- ??? FOSS4G.tw Summer Training ternatives to deal with geospatial information in Tai- wan. The final speaker in this session is Mr. Dongpo ??? Deng who is coordinating and also the initial core member of OSGeo Taiwan. He introduced OSGeo 29 to August 3 at Academia Sinica Computer Cen- Foundation, global rapidly forming OSGeo commu- ter, Taipei. The free courses aim to provide technical nities, and OSGeo Taiwan, with further explana- and practical introduction for beginners, especially tory discussion of crucial software aspect of FOSS4G. students and research assistants. We hope the at- The talk gave the attendees a roadshow of current tendees will use these tools in their homework or geospatial open source software applications, guid- projects, gaining a deeper understanding, and we ing them with the freedom of choices of appropriate anticipate them to be seeds for future promotion tools for their own applications. The final panel dis- of FOSS4G. The serial courses cover several heavy- cussion focused on ?? weighted applications such as GRASS, PostgreSQL/- Development of OSGeo Activities in Taiwan. PostGIS, MapServer, GeoTool and MapGuide, for ex- Panelists not only shared their experiences of FOSS ample. Every instructor has his/her own rich expe- and FOSS4G, but also provided some suggestions for riences for FOSS4G, and can apply to local scenar- future development of FOSS4G in Taiwan. Most pan- ios to best demonstrate the usage of it. The series elists agreed, Open Data is the first and most urgent of courses is sponsored by Institute of Information task on the list of priorities of OSGeo Taiwan. En- Science, Academia Sinica and Academia Sinica Com- thusiastically discussed by panelists, the OSGeo Tai- puter Center.

Nascent UK Local Chapter Update

Jo Cook and Leif Isaksen, 5th July 2007 UK who are new to the Open Source Geospatial movement After returning from FOSS4G in 2006, several UK • To raise awareness of the beneﬁts of public ac- attendees felt there was an interest in setting up a cess to geodata in the UK UK local chapter for OSGEO. The following mission • To work with OGC members in the UK to pro- statement was derived: mote the implementation and development of open standards for geodata exchange • To provide a forum for discussion and promo- • To, maybe, one day, act as focal point for a UK- tion of Open Source Geospatial Software in the hosted and organised FOSS4G conference UK • To help more UK organisations discover the To date we have had 13 expressions of interest on opportunity of open source geospatial tools the wiki page (http://wiki.osgeo.org/index.php/United_- • To support the development of a fully featured Kingdom). This may not seem like a lot of peo- UK SDI ple, but they are from a range of different organisa- • To provide a starting point for people from the tions and disciplines. There has been attendance

ISSN 1994-1897 53 OSGeo Journal Spanish Local Chapter Update Vol. 2, August 2007

Figure 1: Dongpo introduced OSGeo Taiwan and gave a tour of FOSS4G at a number of functions and conferences, most with attendance at geospatial conferences. Individ- notably the Open Knowledge Foundation's Open ual members are planning attendance at FOSS4G Knowledge 1.0 conference in March 2007. There 2007 and have submitted ideas for presentations, is some overlap with the OpenStreetmap Project, awaiting the peer vote in mid July. which is growing in popularity and holding it's Individual members have also been involved in a own conference in Manchester 14-15th July 2007 number of initiatives that relate to the speciﬁed mis- (http://www.stateofthemap.org/), and discussion sions: Oxford Archaeology (thehumanjourney.net) between the two groups has been interesting! In have begun to release data on the location of their the UK there is also a campaign being run by the excavations via WMS/WFS, and are working on a Guardian newspaper on free access to data of all number of software ideas such as GIS on a USB stick. kinds (http://www.freeourdata.org.uk/), and again They have also begun to arrange training courses there is a degree of overlap of interests that will be for their staff in GRASS and QGIS as part of a pro- worth investigating further. gramme of evaluating open source geospatial tools within their day to day work. Overall- more work is needed to build on these Activity promising beginnings and to gain more expressions of interest, along with a specialised forum for discus- Other than gathering expressions of interest, a spe- sion. Many of the issues that a local OSGEO chap- cialised forum for discussion has not yet been im- ter would hope to raise are gaining public awareness plemented, and interaction is mainly informal, via within the UK, and we should seek to be more in- related discussion groups and mailing lists, along volved in this.

Spanish Local Chapter Update

Jorge Gaspar Sanz Salinas, Lorenzo Becchi, Lluís Vicens, sically identiﬁed in its mailing list. Marc Compte, Pedro-Juan Ferrer

The Spanish Local Chaper is not yet ofﬁcial but recently started moving its ﬁrst steps. Actually it is ba-

ISSN 1994-1897 54 OSGeo Journal Spanish Local Chapter Update Vol. 2, August 2007

Introduction section of the book, or you think that new contents should be added, please, feel free to join the mailing The Spanish Local Chapter started on March 2007 in list (http://lists.osgeo.org/mailman/listinfo/ Girona within the first Spanish FOSS4G Meeting (I libro_sig). Jornadas SIG Libre44). An unplanned meeting was organized by Luis W. Sevilla (gvSIG development Chief45) and Lorenzo Becchi (Ominiverdi.org46) and OSGeo portal translation many meeting participants joined the debate. Since May 2007, the main web pages of the OS- After this first meeting a mailing list has been cre- Geo site are translated into Spanish. This effort ated and some activities started as an effort to have from the Spanish Local Chapter tries to bring the a Spanish language community with the main ob- Foundation to Spanish spoken people. If you want jective to spread the word about FOSS4G and share to help with the translation, improving pages, or ideas, technical problems, etc. translating new ones, contact with Jorge Sanz. You The Spanish Chapter is not yet an official organi- can see the progress of the translation at the Portal zation. Translation Status (http://wiki.osgeo.org/index. php/Portal_Translation_Status) page of the OS- Activities Geo wiki.

Mailing list II Jornadas de SIG Libre The creation of the list, at the beginning of March The GIS & Remote Sensing Service (SIGTE49) of the 2007, was the first activity of the starting group and University of Girona (in Spain) is working to orga- will serve as a point of discuss for Spanish spoken nize and prepare the II Spanish FOSS4G Meeting FOSS4G users and developers. At this moment the that will be held in Girona during March 3th, 4th list counts 134 subscribers and we hope, as other and 5th 2008. The previous meeting 50 was with- activities become better known, the Spanish spoken out doubt an unprecedented success, pointing out community will get more involved in the list. the high level of participation and also the high level and quality of the communications presented during Spanish GIS Book the event. One of the main purposes of this meetings is to The biggest effort has been focused to create a present and show new or existing Open Source GIS Free GIS Book in Spanish language. At this software (and its capabilities or improvements). We time a PSC (Project Steering Commetee47) has be- are also really interested in showing real applications ing defined and some important steps have be- and experiences using any Open Source GIS & Re- ing accomplished. There is an almost stable table mote Sensing software, 3D visualizations tools, ... of contents (http://wiki.osgeo.org/index.php/ and their degree of presence in professional world, Contenido) for the first version of the book, a small pointing out advantages and disadvantages. group of people (http://wiki.osgeo.org/index. Moreover, during this II Spanish FOSS4G Meet- php/AsignaciUn_de_cap˚ Stulosˇ ) contributed, led by ing51, we are trying to emphasize the discussion Víctor Olaya (SEXTANTE48) and some technological about free geodata, open standards for open source facilities provided by OSGeo (a specific mailing list software, the role that Public Administration could and a SVN repository for the book sources). (or should) play in the adoption of GIS Open Source The book will be written in LaTeX and the Software solutions. timetable is expected to be followed. Taking advantage of Tyler Mitchell’s attendance Of course, authors are always welcomed. If you during the meeting, an official presentation of the have some Spanish materials that fit some chapter or OSGeo will be done. In addition, some discussion 44http://www.sigte.udg.es/jornadassiglibre2007/ 45http://www.gvsig.gva.es 46http://www.ominiverdi.org 47http://wiki.osgeo.org/index.php/Libro_SIG_PSC_candidaturas 48http://sextantegis.com/ 49http://www.sigte.udg.es 50http://www.sigte.udg.es/jornadassiglibre2007 51http://www.sigte.udg.es/jornadassiglibre/

ISSN 1994-1897 55 Local Reports OSGeo Journal Vol. 2, August 2007 group sessions regarding to the Spanish Local Chap- ticipation of the community is the key to keep the ter are also planned. wheel moving on. We are exploring some ideas and the active par-

Ottawa Local Chapter

Jeff McKenna, Dave Sampson focused on the National Capital Region around Ot- tawa with face to face meetings providing network- ing, support and educational opportunities to the lo- Background cal members. The Ottawa Local Chapter of OSGeo was formed in July 2006 in Ottawa, Ontario, Canada. The group is a union of two previously existing groups: the Ottawa Chapter Co-Chairs GRASS Users Group (OGUG) formed in June 2005, and the Ottawa MapServer Users Group (OMSUG) * Dave Sampson, Natural Resources Canada * Jeff formed in winter 2003. Both groups were already McKenna, DM Solutions Group very active in the FOSS4G community, with OGUG continuing development and use of the Public Geo- data License in english (http://cemml.carleton. Current Activities ca:8080/OGUG/pgl), and OMSUG hosting the 2004 Open Source GIS Conference in Ottawa (http:// Meeting approximately once a month, the Chapter www.omsug.ca/osgis2004/). The Ottawa OSGEO regularly enjoys presentations and discussions on Chapter was formed by the combination of those various FOSS4G subjects, as well as refreshments. original two groups, but also includes active lo- Recent presentation topics include the Nunaliit At- cal users and developers of at least the following las Framework (http://nunaliit.org/), as well as OSGeo projects: GDAL/OGR, GRASS, Mapbuilder, GeoServer’s functionality (http://www.geoserver. MapGuide, and MapServer. org/).

Mission Important Links

The OSGeo Ottawa Chapter exists to provide sup- * Mailing list - http://lists.osgeo.org/mailman/ port for local users of Open Source Geospatial soft- listinfo/Ottawa_users * Chapter home page ware, and to promote the goals and projects of the - http://wiki.osgeo.org/index.php/Ottawa_ Open Source Geospatial Foundation. The Chapter is Chapter

ISSN 1994-1897 56 OSGeo Journal So simple to install Vol. 2, August 2007

Developer Announcements

So simple to install deegree releases 2.1 demos as WAR archives Release management in deegree2

Dr. Markus Lupp (aka Markus Müller) With the 2.1 releases a number of changes took place concerning code management in deegree: About deegree • deegree moved from CVS to SVN at http://deegree.wald.intevation.org/ • a stable branch was introduced that is uses as deegree (http://www.deegree.org) is a Java-based the base for releases Open Source / Free Software framework for the im- • a testing process was installed that results in plementation of Spatial Data Infrastructures (SDI). more stable releases It contains the services needed for SDI (deegree • demos (packages containing a specific, pre- Web Services) as well as portal components (deegree configured component that is part of deegree) iGeoPortal), mechanisms for handling security and are managed in the SVN and are published as access control issues (deegree iGeoSecurity) and stor- WAR-archives. age / visualization of 3D geodata (deegree iGeo3D). deegree is conceptually and interface-wise based All these measures aim at a more professional and on the standards of the Open Geospatial Consor- open development of the project. tium (OGC) and ISO / TC 211. At the time of writing it is the most comprehensive implementation of those standards in one Open Source framework. The deegree 2.1 rc1 demo releases framework is component-based to a high degree, al- On June 11th (in time for deegree day) the 2.1 release lowing the flexible creation of solutions for a wide candidates for the following demos were published: variety of use cases. deegree is the official reference implementation of • deegree Web Map Service (the designated offi- the OGC for the Web Map Service and Web Coverage cial Reference Implementation of the OGC for Service standards. It is published under the GNU WMS 1.3 and WMS 1.1.1) Lesser General Public License. • deegree Web Feature Service (WFS 1.1.0)

ISSN 1994-1897 57 Developer Announcements OSGeo Journal Vol. 2, August 2007

• deegree Web Coverage Service (the designated For installation of the Catalogue Service, setup official Reference Implementation of the OGC of a PostgreSQL/PostGIS is necessary. For the for WCS 1.0) WPVS/WTS an OpenGL-enabled graphics card is • deegree iGeoPortal necessary. • deegree Catalogue Service – Web (CS-W 2.0 Great effort was also put into creation of docu- ISO application profile) mentation for this six demo releases. • deegree Web Perspective View Service / Web The demo releases can be accessed online at: Terrain Service (WPVS/WTS) http://demo.deegree.org/ and downloaded from The publication as WAR-archives including pre- http://www.deegree.org. configured data sources makes the installation and deployment of the demos very easy. WMS, WFS, Dr. Markus Lupp (aka Markus Müller) WCS and iGeoPortal can be installed within seconds http://www.lat-lon.de Aennchenstr 19 (!) assuming that a Servlet Engine like Tomcat is al- 53177 Bonn, Germany ready installed. mueller AT lat-lon.de or markus.lupp AT lat-lon.de

ISSN 1994-1897 58 Developer Announcements OSGeo Journal Vol. 2, August 2007

Editor in Chief: This work is licensed under the Creative Commons Attribution- No Derivative Works 3.0 License. To view a copy of this licence, tmitchell AT osgeo.org Tyler Mitchell - visit: http://creativecommons.org/licenses/by-nd/3.0/ or send a Editor, News: letter to Creative Commons, 171 Second Street, Suite 300, San Jason Fournier Francisco, California 94105, USA. Editor, Case Studies: Micha Silver

Editor, Project Spotlights: All articles are copyrighted by the respective authors. Please Martin Wegmann use the OSGeo Journal url for submitting articles, more details concerning submission instructions can be found on the OSGeo Editor, Integration Studies: homepage. Martin Wegmann Journal online: http://www.osgeo.org/journal Editor, Programming Tutorials: OSGeo Homepage: http://www.osgeo.org Landon Blake Mail contact through OSGeo, PO Box 4844, Williams Lake, Acknowledgements British Columbia, Canada, V2G 2V8 Various reviewers & the GRASS News Project

The OSGeo Journal is a publication of the OSGeo Foundation. The base of this journal, the LATEX 2εstyle source has been kindly provided by the GRASS and R News editorial board.

ISSN 1994-1897

ISSN 1994-1897 59