The Framework for Accessible Applications: Text-Based Case for Blind People Michael Pozhidaev National Research Tomsk State University, Computer Science Department Lenina Avenue, 36, 634050 Tomsk, Russia [email protected] ABSTRACT still are not able to enjoy all the advantages of this process, This paper offers a Java framework for creating accessible mostly because the accessibility technologies are just add- applications for blind and visually impaired people as part ons to interfaces initially designed for sighted users. This of a proposed general conception based on the maximum kind of solution can be considered only a partial solution use of objects filled with text data only. It offers new types since it offers the general ability to do the same things as ev- of applications more easily recognizable by disabled persons, erybody but sacrifices the efficiency and level of personal helping them to do their work faster and more comfortably. comfort. Strong and weak points are analyzed. The published proto- Each time it is necessary to create an application designed type of the proposed platform is described as well as the con- and highly accessible for blind people developers always cre- clusions of the performed experiments. The prototype is im- ate a usual application for graphical user interface (GUI). plemented on Java SE and wrapped by a GNU/Linux envi- Their main distinguishing characteristics aren’t more than ronment as a bootable ISO-image. features to enable the ability to contrast colours and ad- just font sizes. These functions are only measures for users with partial sight. If somebody has never dealt with GUI Categories and Subject Descriptors as a sighted user the complete and proper understanding D.2.2 [Software Engineering]: Design Tools and Tech- is generally impossible (especially for senior users). Respect- niques—User interfaces ing the increasing requirements for an accessible environ- ment we would like to propose a platform designed com- General Terms pletely for blind people which could be a stable solution for the problems mentioned above. Experimentation Speed, comfort and good understandability usually mean just one thing; keeping things simple. The question is what Keywords kind of user interface (UI) could be simple and at the same time functionally sufficient. The suggested solution to an in- accessibility, blind people, user interfaces, usability, Java, terface for blind people is, on one hand, simple enough and, API on the other hand, could provide the replacement for most GUI widgets. We will discuss the pros and cons, and discover 1. INTRODUCTION how visualization and speech output can mutually supple- During last decade we have seen the process of diversifi- ment each other. The prototype of the proposed platform cation of information technologies which people use in their was published to demonstrate implementation on Java. It everyday lives. If during the mid 2000-s almost all tasks were consists of the core maintaining new interface and launched performed using Microsoft Windows and desktops only, to- applications, as well as a set of Java classes making API day’s solutions are very different, regardless of whether we for creating new applications. This environment could are speaking about software platforms (Mac OS X, iOs, An- be launched on any OS with the Java virtual machine but droid are have a wider distribution and very often can be we would like to see it as a complete OS on the Linux kernel considered as real competitors to Windows) or about ways (published prototypes are prepared in this way). of interaction with PC’s or gadgets (multitouch as a replace- ment for the mouse). In the meantime, blind and visually 2. TEXT-BASED ENVIRONMENT FUNDA- impaired people (from here on in “blind people”) typically MENTALS 2.1 General conception First of all, we would like to describe an alternative solu- Permission to make digital or hard copies of all or part of this work for tion to ensure it is really suitable and comfortable the blind. personal or classroom use is granted without fee provided that copies are There is some prior experience in this area described be- not made or distributed for profit or commercial advantage and that copies low and we would like to respect it. The new environ- bear this notice and the full citation on the first page. To copy otherwise, to ment enable speech output and some visual representation republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. without involving any screen reading software. The picture Copyright 2014 ACM 978-1-4503-2889-0/14/10 ...$15.00. on the screen remains very important because it is used by users with low vision, though it plays only a very supple- is based on a Gnu/Linux distribution. It involves a number mentary role now. We suggest splitting the screen into sev- of applications for various tasks which were wrapped with eral rectangular areas, and filling its entire space with some speech-enabled interactions. Adriane cannot be considered tiles. Each tile shows some textual data related to the cor- as a platform for constructing new accessible applications responding object and is closely associated with speech out- because it hasn’t any strict and flexible core providing API put. Font size and colour should be easily adjustable to suit for developers (although its interface could be considered a particular user’s requirements. as quite consistent) and the components it uses are mostly The method of how to constructing a comfortable and suf- based on the Dialog [23] utility and Bash scripts [14]. ficient speech representation for all objects needed for work is one of the main goals of our research. It is described 2.3 Pros and Cons in detail below but, roughly speaking, we are considering Let us consider all the reasons why our application makes all of the usual GUI widgets in a way where no graphical sense and look at arguments for why it could have some weak and visual data is involved in their representation. For- characteristics. To respond to criticism, which we may ex- tunately, there are only a few cases when it is impossible pect, we will give additional comments for why we don’t take (e.g. the web page cannot be described in words without them seriously and don’t think these criticisms are crucial. lost of entirety). Let’s consider, for example, a text edit box: As a stronger approach, everything is written like it would there is nothing graphical about it, but the selection of some be as a complete OS, and we will also look at why the case part of the text with a mouse is performed through choosing of running in an environment on Microsoft Windows is sig- some other colour and this happens by using visual informa- nificantly weaker. tion. Hence, we must invent something else for text selection The statements supporting our proposal could be as fol- but there is no need to create anything new for text input lows: itself. The environment is designed for distribution with some 1. Users no longer need to struggle with GUI and that standard set of applications prepared and organized saves a lot of time, increasing efficiency. GUI is a very in a new way. At first glance it seems that we will face a unsuitable solution because in complicated cases blind tremendous number of new functions to implement, but ac- users should reference a screen structure which they tually there is a wide range of libraries for Java, distributed don’t see. On desktops and laptops, GUI is useful only under public licenses; so there is no need to do anything with a mouse, which is inaccessible for blind users. from scratch. We should simply prepare a new interface With tablet computers, where multitouch is popu- for them with our new platform. lar, blind users should be able to touch the screen in same position multiple times. That is easy to do having visual information, but in our speech-enabled 2.2 Known experience case it turns out as a very time consuming procedure. Usually blind people use screen readers launched In the meantime, the presence of GUI is a completely on a GUI, but we will not review them as we are interested artificial problem because GUI is created by the hands only in software designed as“audio desktops”. There are cur- of people and isn’t something essential for interaction rently three solutions which can be considered as providing with a PC. a speech environment without screen reading. All of them accomplish this in different ways. They are Emacspeak [20], 2. A new type of interface could be easier to understand Dolphin Guide [2] and Adriane [8]. Emacspeak certainly had for blind users with a lack of experience with com- a major impact on the outcome of this paper. It is an add-on puters, especially seniors. With GUI it is necessary for the popular text editor GNU Emacs [22] which has a lot to have a proper understanding of what it is. That of additional features covering areas significantly wider than is not a problem for users who have previously used just usual text editing, such as file management, mail read- a PC as a sighted user. But for those who have never ing, calendar etc. The main advantage of Emacspeak is that seen a computer screen such understanding becomes with some proper training it can help users to be highly ef- a really serious problem. ficient(almost to the level of a sighted user). We are wanted 3. The general conception and suggested implementation to keep all of the positive parts of that interesting experience.