Chapter 2: The GNU Manifestoi I 31 2 The GNU Manife$to

The GNU Manifesto was writte at the beginning of the GNU Project, to ask for participation and support. r the first few years, it was updated in minor ways to accountfor developments,but now it seemsbest to leave it unchangedas mostpeople haveseen it. Since t t time, we have learned about certain common misunderstandingsthat different w rding could helpavoid, andfootnotes have been r- addedover the years to explain the e misunderstandings. It t,

)e GNU, which stands for Gnu'~ ot Unix, is the name for the complete Unix- ot compatible software systemwhi~h I am writing so that I can give it away free to ~n everyonewho can use it.! StverW ther volunteers are helping me. Contributions of time, money,programs and equi ment are greatly needed. re So far we have an text ditor with Lisp for writing editor commands, U a source-leveldebugger, a yacc-co patible parser generator,a linker, and around 11, 35 utilities. A shell (command int rpreter) is nearly completed. A new portable optimizing C compiler has compile itself and may be releasedthis year. An initial kernel exists but many more feature are neededto emulateUnix. When the kernel and compiler are finished, it will be ossibleto distribute a GNU systemsuitable for program development.We will use TEX as our text formatter, but an nroff is being worked on. We will use the free, p able X window systemas well. After this we nd will add a portable CommonLisp, Empire game,a spreadsheet,and hundredsof Je- otherthings, plus on-line doc1limenttion. We hope to supply,eventually, everything ~lf, useful that normally comes with a nix system,and more. GNU will be able to run Unix p ograms, but will not be identical to Unix. We :at- will make all improvements that e convenient, based on our experience with for other operatingsystems. In particu ar, we plan to have longer file names,file ver- me sion numbers, a crashproof file sy tern, file name completion perhaps, terminal-

ci -I I ea II!! The wording here was careless,The inte tion was that nobodywould haveto pay for permission .-~- to use the GNU system. But the words do not make this clear, and people often interpret them itly as saying that copies of GNU should a1 ays be distributed at little or no charge. That was never I' ,., the intent; later on, the manifesto menti ns the possibility of companiesproviding the service of m't (\i! f" distribution for a profit. SubsequentlyI h ve learnedto distinguish carefully between"free" in the i to I.!~ ~i: I senseof freedomand "free" in the sens of price. is software that usershave the freedomto distribute and change. Someusers may obtain copies at no charge,while others pay

L to obtain copies-and if the funds help upport improving the software,so much the better. The ~" '*;ir importantthing is that everyonewho has a copy has the freedomto cooperatewith others in using ~eo . it. I !. I ,'- f Originally written in 1984,this versiollis part of reeSoftware. Free Society: SelectedEssays of Richard M. Sta/l- [I R man,2002, GNU Press(http://www.gnupress.org ; ISBN 1-882114-98-1. f : Verbatim copying and distribution of this entire article is i permitted in any mediu , provided this notice is preserved

le.1 32 Free Software, Frpe Society: SelectedEssays of Richard M. Stallman

independent display support, nd perhaps eventually a Lisp-based window sys- tem through which severalLis programsand ordinary Unix programscan sharea screen. Both C and Lisp will e available as systemprogramming languages.We will try to supportuucP, MIT haosnet,and Internetprotocols for communication. GNU is aimed initially at m chines in the 68000/16000class with virtual mem- ory, becausethey are the easi st machinesto make it run on. The extra effort to make it run on smaller mac~i es will be left to someonewho wants to use it on them. To avoid horrible confusiqn, pleasepronounce the 'G' in the word 'GNU' when it is the nameof this project.,

Why I Must Write G~U

I consider that the golden Ie requires that if I like a program I must share it with otherpeople who like it. S ftware sellers want to divide the usersand conquer them, making eachuser agree ot to sharewith others. I refuse to break solidarity with other users in this way. cannot in good conscience sign a nondisclosure agreementor a softwarelicense agreement. For years I worked within the Artificial Intelligence Lab to resist such endenciesand other inhospitalities, but eventually they had gone too far: I could ot remain in an institution where such things are done for me againstmy will. So that I can continue to use computerswithout dishonor, I have decided to put togethera sufficient body of fre softwareso that I will be able to get along without any software that is not free. I h ve resignedfrom the AI lab to deny MIT any legal excuseto preventme from givi g GNU away.

Unix is not my ideal sYstem~ut it is not too bad. The essential features of Unix seem to be good ones, and I .nk I can fill in what Unix lacks without spoiling them. And a system compatib e with Unix would be convenient for many other people to adopt. I

~~~; How GNU Will Be Available

GNU is not in the public dOE ain. Everyone will be permitted to modify and r~distribut~ GNU, but no ~istrib tor ~ill b.eallo,,:,ed to restrict its further redistribu- tion. That IS to say, propnetary modIficatIons wIll not be allowed. I want to make I"",\;11:MY?~i ...,,"'~, sure that all versions of GNU re ain free. ~tii":i;q: ~~!\'*~;,! .'-;I,~{t! ~ " !'I~ ~' ~ ,.IE I;;O ,~1(1'~ I have found many other pro~ ammers who are excited about GNU and want to help. !t}.~(; ~J Many programmersare unh ppy about the commercialization of system soft- Ware. It may enable them to ake more money, but it requires them to feel in

34

Once GNU is written, eve one will be able to obtain good system software free, just like air.3 c This means much more th just saving everyonethe price of a Unix license. i; It means that much wasteful duplication of system programming effort will be I, avoided. This effort can go in tead into advancingthe stateof the art. Complete systemsources ill be availableto everyone. As a result, a userwho needschanges in the system ill always be free to make them himself, or hire any available programmer or com any to make them for him. Users will no longer be at the mercy of one programm r or companywhich owns the sourcesand is in sole position to make changes. Schools will be able to pro ide a much more educational environment by en- couraging all studentsto stud and improve the systemcode. Harvard's computer lab used to have the policy th t no program could be installed on the systemif its sourceswere not on public d.splay, and upheld it by actually refusing to install certain programs. I was very uch inspired by this. Finally, the overheadof con idering who owns the systemsoftware and what one is or is not entitled to do with i will be lifted. Arrangementsto make peo Ie pay for using a program, including licensing of copies, always incur a tremen ous cost to society through the cumbersomemech- anisms necessaryto figure out how much (that is, which programs) a person must pay for. And only a police stat can force everyoneto obey them. Considera space station where air must be man factured at greatcost: charging each breatherper liter of air may be fair, but we .ng the meteredgas mask all day and all night is intolerable evenif everyonec afford to pay the air bill. And the TV camerasev- erywhere to seeif you ever tak the mask off are outrageous.It's better to support the air plant with a headtax an chuck the masks. Copying all or parts of a pro ram is as naturalto a programmeras breathing,and as productive. It ought to be as free. Some easily rebuttedobjecti ns to GNU's goals: "Nobody will use it if lit s free, because that means they can't rely on any support." "You have to chargefor programto pay for providing the support." If people would rather payor GNU plus service than get GNU free without service, a companyto provide ust service to people who have obtained GNU free ought to be profitable. We must distinguishbetwee supportin the fonn of real programming work and mere hand-holding. The fonn r is somethingone cannot rely on from a software vendor. If your problem is not hared by enoughpeople, the vendor will tell you to get lost.

3 This is anotherplace I failed to diS ~ngUiShcarefully betweenthe two different meaningsof "free." The statementas it standsis not f se-you can get copies of GNU software at no charge, from your friends or over the Internet. B t it does suggestthe wrong idea.

~ , Chapter2: The GNU Manifesto I'" 35 Ii .Ii ,,' 1).1 f' If your businessneeds to be able to rely on support, the only way is to have all II :",", the necessarysources and tools. Then ou can hire any availableperson to fix your :1" "'1I' Ii I problem; you are not at the mercy of individual. With Unix, the price of sources !1if, i puts this out of considerationfor most usinesses.With GNU this will be easy. It is still possible for there to be no availabl competentperson, but this problem cannot be blamed on distribution arrangemen .GNU does not eliminate all the world's problems, only some of them. Meanwhile, the users who know n thing about computersneed hand-holding: doing things for them which they coul easily do themselvesbut don't know how. Such servicescould be provided b companiesthat sell just hand-holding and repair service. If it is true that users uld rather spend money and get a product with service, they will also be willin to buy the service having got the product for free. The service companieswill ompete in quality and price; users will not be tied to any particular one. Meanw "Ie, those of us who don't need the service should be able to use the program with ut paying for the service. "You cannot reach many people ~ithout advertising, and you must charge for the program to support that." I "It's no use advertising a program peoplecan get free." There are various forms of free or ve cheappublicity that can be usedto inform numbersof computerusers aboutsome .ng like GNU. But it may be true that one can reachmore microcomputerusers w. h advertising. If this is really so, a business which advertisesthe service of copyin and mailing GNU for a fee ought to be successfulenough to pay for its adve ising and more. This way, only the users who benefit from the advertising pay fo it. On the other hand, if many people g t GNU from their friends, and such com- panies don't succeed,this will show at advertising was not really necessaryto spreadGNU. Why is it that free marketadvocates don't want to let the free market decidethis4? "My companyneeds a proprietary perating systemto get a competitive edge." GNU will remove operating systems ftware from the realm of competition. You will not be able to get an edge in this ar a, but neither will your competitorsbe able to get an edge over you. You and they ill competein other areas,while benefiting mutually in this one. If your busines is selling an , you will not like GNU, but that's tough on you. If your businessis somethingelse, GNU can save you from being pushed into e expensivebusiness of selling operating systems.

4 The Free Software Foundationraises most of its funds from a distribution service, although it is a charity ratherthan a company.If no one ch sesto obtaincopies by orderingthem from the FSF, it will be unable to do its work. But this doe not mean that proprietaryrestrictions are justified to force every user to pay. If a small fractio of all the users order copies from the FSF, that is sufficient to keep the FSF afloat. So we ask sersto chooseto supportus in this way. Have you done your part? 36

I would like to see GNU dev lopment supportedby gifts from many manufac- turers and users,reducing the co t to each.5 "Don't programmersdese e a reward for their creativity?" If anything deservesa reward it is social contribution. Creativity can be a social contribution, but only in so far s society is free to use the results. If programmers deserveto be rewarded for cre 'ng innovative programs, by the same token they deserveto be punishedif they r strict the use of theseprograms. "Shouldn't a programmer e able to ask for a reward for his creativity?" There is nothing wrong with anting pay for work, or seekingto maximize one's income, as long as one does n t use means that are destructive. But the means customary in the field of softw e today are basedon destruction. Extracting money from user of a program by restricting their use of it is de- structive becausethe restriction reducethe amountand the ways that the program can be used, This reducesthe amount of wealth that humanity derives from the program. When there is a deli erate choice to restrict, the harmful consequences

areThe deliberate reasonadestruction, good citizen d s not use suchdestructive means to becomewealth- ier is that, if everyonedid so, we would all become poorer from the mutual de- structiveness.This is Kantian thics; or, the Golden Rule, Since I do not like the consequencesthat result if eve one hoards information, I am required to consider it wrong for one to do so, Specfically, the desireto be rewardedfor one's creativity does not justify depriving the orld in generalof all or part of that creativity. "Won't programmerss " I could answer that nobod is forced to be a programmer. Most of us cannot manageto get any money for tanding on the streetand making faces. But we are not, as a result, condemnedto spend our lives standingon the streetmaking faces, and starving. We do somethin else. But that is the wrong answe becauseit acceptsthe questioner'simplicit assump- tion: that without ownership f software, programmerscannot possibly be paid a

cent.The Supposedly real reasonit programm is all or n thing.rs will not starve is that it will still be possible for them to get paid for progr 'ng; just not paid as much as now, Restricting copying is not e only basis for businessin software. It is the most common basis becauseit bri gs in the most money. If it were prohibited, or re- jected by the customer,softw e businesswould move to other basesof organiza- tion which are now usedless often. There are always numerousways to organize

anyProbably kind of business.programming wi I not be as lucrative on the new basis as it is now. But that is not an argument agai st the change. It is not considered an injustice that sales clerks make the salarie that they now do. If programmers made the same, that would not be an injustice either. (In practice they would still make considerably

more than that,) -;-- A group of computer companies recently pooled funds to support maintenance of tlie GNU C " Compiler. Chapter2: The GNU Manifesto 37

"Control over the use of one's deas" really constitutes control over other peo- ple's lives; and it is usually used t make their lives more difficult. People who have studiedthe iss e of intellectual property rights carefully (such as lawyers) say that there is no in .nsic right to intellectual property. The kinds of supposedintellectual property rig ts that the governmentrecognizes were created by specific acts of legislation for s cific purposes. For example, the patent systemwas establishedto encourageinventors to dis- close the details of their invention. Its purpose was to help society rather than to help inventors. At the time, the Ii span of 17 years for a patent was short com- pared with the rate of advanceof e state of the art. Since patents are an issue only among manufacturers,for wh m the cost and effort of a license agreementare small comparedwith setting up pro uction, the patentsoften do not do much harm. They do not obstruct most individu s who usepatented products. The idea of copyright did not e .st in ancient times, when authors frequently copied other authors at length in orks of non-fiction. This practice was useful, and is the only way many authors' orks have survived evenin part. The copyright systemwas createdexpressly for th purposeof encouragingauthorship. In the do- main for which it was invented-b oks, which could be copied economically only on a printing press-it did little h , and did not obstruct most of the individuals who read the books. All intellectual property rights ar just licensesgranted by societybecause it was thought, rightly or wrongly, that soc ety as a whole would benefit by granting them. But in any particular situation,we h ve to ask: are we really betteroff granting such license? What kind of act are we lic nsing a personto do? The caseof programs todayis ve different from that of books a hundred years ago. The fact that the easiestwa to copy a program is from one neighbor to another, the fact that a program ha both source code and object code which are distinct, and the fact that a program is used rather than read and enjoyed, combine to createa situation in which a pers n who enforcesa copyright is harming society as a whole both materially and sp. .tually; in which a person should not do so If regardlessof whetherthe law enabl s him to.

;t "Competition makes things get Idone better." ~- The paradigm of competition is a race: by rewarding the winner, we encourage l- everyoneto run faster. When capita ism really works this way, it does a good job; :e but its defendersare wrong in assu 'ng it always works this way. If the runners forget why the reward is offered and ecomeintent on winning, no matterhow, they lt may find other strategies-such as, trucking other runners. If the runners get into at a fist fight, they will all finish late. Proprietary and secretsoftware is the moral equivalentof runners in a fist fight, Sad to say, the only referee we've ot does not seemto object to fights; he just regulates them ("For every ten yard you run, you can fire one shot"). He really ought to break them up, and penaliz runners for eventrying to fight. c "Won't everyonestop program~ng without a monetary incentive?" 38 Free Software,Free S~ciety: SelectedEssays of Richard M. Stallman

Actually, manypeople will prog with absolutelyno monetaryincentive. Pro- gramming has an irresistible fasci ation for some people, usually the people who are best at it. There is no shortageof professional musicians who keep at it even though they haveno hope of maki g a living that way. But really this question, though commonly asked, is not appropriateto the sit- uation. Pay for programmerswill not disappear,only become less. So the right question is, will anyone program ith a reducedmonetary incentive? Myexperi- ence shows that they will. For more than ten years, many f the world's best programmersworked at the Artificial Intelligence Lab for far I ss money than they could have had anywhere else. They got many kinds of non monetaryrewards: fame and appreciation, for example. And creativity is also fu , a reward in itself. Then most of them left when 0 ered a chanceto do the sameinteresting work for a lot of money. What the facts show is that peo Ie will program for reasonsother than riches; but if given a chanceto make a lot of money as well, they will come to expectand demand it. Low-paying organiza .ons do poorly in competition with high-paying ones, but they do not have to do b ly if the high-paying ones are banned. "We need the programmersd sperately. If they demand that we stop helping our neighbors,we have to ob y." You're neverso desperatethat y u haveto obeythis sort of demand.Remember: millions for defense,but not a cen for tribute! "Programmersneed to make living somehow." In the short run, this is true. owever, there are plenty of ways that program- mers could make a living without selling the right to use a program. This way is customary now becauseit brings rogrammersand businessmenthe most money, not becauseit is the only way to ake a living. It is easyto find other ways if you want to find them. Here are a numberof examples .A manufacturerintroducing new computerwill pay for the porting of oper- ating systemsonto the new h dware. .The sale of teaching, hand- lding and maintenanceservices could also em- ploy programmers. .People with new ideasc04ld distribute programsas freeware,asking for dona- tions from satisfiedusers, or selling hand-holding services. I have met people who are alreadyworking thi way successfully. .Users with relatedneeds can form users' groups,and pay dues. A groupwould contract with programming companies to write programs that the group's

memberswould like to use. ~ All sorts of developmentcan b funded with a SoftwareTax: .Suppose everyonewho buy a computerhas to pay x percentof the price as a software tax. The governm nt gives this to an agencylike the NSF to spend on softwaredevelopment. Chapter 2: The GNU Manifesto 39

.But if the computerbuyer mak s a donationto software developmenthimself, he can take a credit against th tax. He can donateto the project of his own choosing--often, chosenbeca se he hopesto use the results when it is done. He can take a credit for any ount of donation up to the total tax he had to pay. .The total tax rate could be deci ed by a vote of the payersof the tax, weighted according to the amount they .11be taxed on. The consequences: .The computer-usingcommuni supportssoftware development. .This community decideswhat 1 vel of supportis needed. .Users who care which projects ell shareis spenton can choosethis for them- selves. In the long run, making program free is a step toward the post-scarcityworld, where nobody will haveto work ve hard just to makea living. Peoplewill be free to devotethemselves to activities tha are fun, suchas programming, after spending the necessaryten hours a week on r quired tasks suchas legislation, family coun- seling, robot repair, and asteroidpro pecting. There will be no need to be able to make a living from programming. We have already greatlyreduced e amountof work that the whole society must do for its actual productivity, but onl a little of this hastranslated itself into leisure for workers becausemuch nonprod ctive activity is required to accompanypro- ductive activity. The main causes0 this are bureaucracyand isometric struggles againstcompetition. Free software ill greatly reduce~ese dr.ain~in the ar~a.of softwareproduction. We must do .s, in order for techmcalgaIns In prodUCtIVIty to translateinto less work for us.