1 The GNU Project 2 3 by Richard Stallman 4 5 originally published in the book "Open Sources" 6 7 The first software-sharing community 8 9 When I started working at the MIT Artificial Intelligence Lab in 1971, 10 I became part of a software-sharing community that had existed for many 11 years. Sharing of software was not limited to our particular community; 12 it is as old as computers, just as sharing of recipes is as old as 13 cooking. But we did it more than most. 14 15 The AI Lab used a timesharing operating system called ITS (the 16 Incompatible Timesharing System) that the lab's staff hackers (1) had 17 designed and written in assembler language for the Digital PDP-10, one 18 of the large computers of the era. As a member of this community, an AI 19 lab staff system hacker, my job was to improve this system. 20 21 We did not call our software "free software", because that term did not 22 yet exist; but that is what it was. Whenever people from another 23 university or a company wanted to port and use a program, we gladly let 24 them. If you saw someone using an unfamiliar and interesting program, 25 you could always ask to see the source code, so that you could read it, 26 change it, or cannibalize parts of it to make a new program. 27 28 (1) The use of "hacker" to mean "security breaker" is a confusion on 29 the part of the mass media. We hackers refuse to recognize that 30 meaning, and continue using the word to mean, "Someone who loves to 31 program and enjoys being clever about it." 32 33 The collapse of the community 34 35 The situation changed drastically in the early 1980s when Digital 36 discontinued the PDP-10 series. Its architecture, elegant and powerful 37 in the 60s, could not extend naturally to the larger address spaces 38 that were becoming feasible in the 80s. This meant that nearly all of 39 the programs composing ITS were obsolete. 40 41 The AI lab hacker community had already collapsed, not long before. In 42 1981, the spin-off company Symbolics had hired away nearly all of the 43 hackers from the AI lab, and the depopulated community was unable to 44 maintain itself. (The book Hackers, by Steve Levy, describes these 45 events, as well as giving a clear picture of this community in its 46 prime.) When the AI lab bought a new PDP-10 in 1982, its administrators 47 decided to use Digital's non-free timesharing system instead of ITS. 48 49 The modern computers of the era, such as the VAX or the 68020, had 50 their own operating systems, but none of them were free software: you 51 had to sign a nondisclosure agreement even to get an executable copy. 52 53 This meant that the first step in using a computer was to promise not 54 to help your neighbor. A cooperating community was forbidden. The rule 55 made by the owners of proprietary software was, "If you share with your 56 neighbor, you are a pirate. If you want any changes, beg us to make 57 them." 58 59 The idea that the proprietary-software social system--the system that 60 says you are not allowed to share or change software--is antisocial, 61 that it is unethical, that it is simply wrong, may come as a surprise 62 to some readers. But what else could we say about a system based on 63 dividing the public and keeping users helpless? Readers who find the 64 idea surprising may have taken proprietary-software social system as 65 given, or judged it on the terms suggested by proprietary software 66 businesses. Software publishers have worked long and hard to convince 67 people that there is only one way to look at the issue. 68 69 When software publishers talk about "enforcing" their "rights" or 70 "stopping piracy", what they actually *say* is secondary. The real 71 message of these statements is in the unstated assumptions they take 72 for granted; the public is supposed to accept them uncritically. So 73 let's examine them. 74 75 One assumption is that software companies have an unquestionable 76 natural right to own software and thus have power over all its users. 77 (If this were a natural right, then no matter how much harm it does to 78 the public, we could not object.) Interestingly, the US Constitution 79 and legal tradition reject this view; copyright is not a natural right, 80 but an artificial government-imposed monopoly that limits the users' 81 natural right to copy. 82 83 Another unstated assumption is that the only important thing about 84 software is what jobs it allows you to do--that we computer users 85 should not care what kind of society we are allowed to have. 86 87 A third assumption is that we would have no usable software (or would 88 never have a program to do this or that particular job) if we did not 89 offer a company power over the users of the program. This assumption 90 may have seemed plausible, before the free software movement 91 demonstrated that we can make plenty of useful software without putting 92 chains on it. 93 94 If we decline to accept these assumptions, and judge these issues based 95 on ordinary common-sense morality while placing the users first, we 96 arrive at very different conclusions. Computer users should be free to 97 modify programs to fit their needs, and free to share software, because 98 helping other people is the basis of society. 99 100 There is no room here for an extensive statement of the reasoning 101 behind this conclusion, so I refer the reader to the web page, 102 http://www.gnu.org/philosophy/why-free.html. 103 104 A stark moral choice. 105 106 With my community gone, to continue as before was impossible. Instead, 107 I faced a stark moral choice. 108 109 The easy choice was to join the proprietary software world, signing 110 nondisclosure agreements and promising not to help my fellow hacker. 111 Most likely I would also be developing software that was released under 112 nondisclosure agreements, thus adding to the pressure on other people 113 to betray their fellows too. 114 115 I could have made money this way, and perhaps amused myself writing 116 code. But I knew that at the end of my career, I would look back on 117 years of building walls to divide people, and feel I had spent my life 118 making the world a worse place. 119 120 I had already experienced being on the receiving end of a nondisclosure 121 agreement, when someone refused to give me and the MIT AI lab the 122 source code for the control program for our printer. (The lack of 123 certain features in this program made use of the printer extremely 124 frustrating.) So I could not tell myself that nondisclosure agreements 125 were innocent. I was very angry when he refused to share with us; I 126 could not turn around and do the same thing to everyone else. 127 128 Another choice, straightforward but unpleasant, was to leave the 129 computer field. That way my skills would not be misused, but they would 130 still be wasted. I would not be culpable for dividing and restricting 131 computer users, but it would happen nonetheless. 132 133 So I looked for a way that a programmer could do something for the 134 good. I asked myself, was there a program or programs that I could 135 write, so as to make a community possible once again? 136 137 The answer was clear: what was needed first was an operating system. 138 That is the crucial software for starting to use a computer. With an 139 operating system, you can do many things; without one, you cannot run 140 the computer at all. With a free operating system, we could again have 141 a community of cooperating hackers--and invite anyone to join. And 142 anyone would be able to use a computer without starting out by 143 conspiring to deprive his or her friends. 144 145 As an operating system developer, I had the right skills for this job. 146 So even though I could not take success for granted, I realized that I 147 was elected to do the job. I chose to make the system compatible with 148 Unix so that it would be portable, and so that Unix users could easily 149 switch to it. The name GNU was chosen following a hacker tradition, as 150 a recursive acronym for "GNU's Not Unix." 151 152 An operating system does not mean just a kernel, barely enough to run 153 other programs. In the 1970s, every operating system worthy of the name 154 included command processors, assemblers, compilers, interpreters, 155 debuggers, text editors, mailers, and much more. ITS had them, Multics 156 had them, VMS had them, and Unix had them. The GNU operating system 157 would include them too. 158 159 Later I heard these words, attributed to Hillel (1): 160 161 If I am not for myself, who will be for me? 162 If I am only for myself, what am I? 163 If not now, when? 164 165 The decision to start the GNU project was based on a similar spirit. 166 167 (1) As an Atheist, I don't follow any religious leaders, but I 168 sometimes find I admire something one of them has said. 169 170 Free as in freedom 171 172 The term "free software" is sometimes misunderstood--it has nothing to 173 do with price. It is about freedom. Here, therefore, is the definition 174 of free software: a program is free software, for you, a particular 175 user, if: 176 * You have the freedom to run the program, for any purpose. 177 * You have the freedom to modify the program to suit your needs. (To 178 make this freedom effective in practice, you must have access to 179 the source code, since making changes in a program without having 180 the source code is exceedingly difficult.) 181 * You have the freedom to redistribute copies, either gratis or for a 182 fee. 183 * You have the freedom to distribute modified versions of the 184 program, so that the community can benefit from your improvements. 185 186 Since "free" refers to freedom, not to price, there is no contradiction 187 between selling copies and free software. In fact, the freedom to sell 188 copies is crucial: collections of free software sold on CD-ROMs are 189 important for the community, and selling them is an important way to 190 raise funds for free software development. Therefore, a program which 191 people are not free to include on these collections is not free 192 software. 193 194 Because of the ambiguity of "free", people have long looked for 195 alternatives, but no one has found a suitable alternative. The English 196 Language has more words and nuances than any other, but it lacks a 197 simple, unambiguous, word that means "free", as in 198 freedom--"unfettered" being the word that comes closest in meaning. 199 Such alternatives as "liberated", "freedom", and "open" have either the 200 wrong meaning or some other disadvantage. 201 202 GNU software and the GNU system 203 204 Developing a whole system is a very large project. To bring it into 205 reach, I decided to adapt and use existing pieces of free software 206 wherever that was possible. For example, I decided at the very 207 beginning to use TeX as the principal text formatter; a few years 208 later, I decided to use the X Window System rather than writing another 209 window system for GNU. 210 211 Because of this decision, the GNU system is not the same as the 212 collection of all GNU software. The GNU system includes programs that 213 are not GNU software, programs that were developed by other people and 214 projects for their own purposes, but which we can use because they are 215 free software. 216 217 Commencing the project 218 219 In January 1984 I quit my job at MIT and began writing GNU software. 220 Leaving MIT was necessary so that MIT would not be able to interfere 221 with distributing GNU as free software. If I had remained on the staff, 222 MIT could have claimed to own the work, and could have imposed their 223 own distribution terms, or even turned the work into a proprietary 224 software package. I had no intention of doing a large amount of work 225 only to see it become useless for its intended purpose: creating a new 226 software-sharing community. 227 228 However, Professor Winston, then the head of the MIT AI Lab, kindly 229 invited me to keep using the lab's facilities. 230 231 The first steps 232 233 Shortly before beginning the GNU project, I heard about the Free 234 University Compiler Kit, also known as VUCK. (The Dutch word for "free" 235 is written with a V.) This was a compiler designed to handle multiple 236 languages, including C and Pascal, and to support multiple target 237 machines. I wrote to its author asking if GNU could use it. 238 239 He responded derisively, stating that the university was free but the 240 compiler was not. I therefore decided that my first program for the GNU 241 project would be a multi-language, multi-platform compiler. 242 243 Hoping to avoid the need to write the whole compiler myself, I obtained 244 the source code for the Pastel compiler, which was a multi-platform 245 compiler developed at Lawrence Livermore Lab. It supported, and was 246 written in, an extended version of Pascal, designed to be a 247 system-programming language. I added a C front end, and began porting 248 it to the Motorola 68000 computer. But I had to give that up when I 249 discovered that the compiler needed many megabytes of stack space, and 250 the available 68000 Unix system would only allow 64k. 251 252 I then realized that the Pastel compiler functioned by parsing the 253 entire input file into a syntax tree, converting the whole syntax tree 254 into a chain of "instructions", and then generating the whole output 255 file, without ever freeing any storage. At this point, I concluded I 256 would have to write a new compiler from scratch. That new compiler is 257 now known as GCC; none of the Pastel compiler is used in it, but I 258 managed to adapt and use the C front end that I had written. But that 259 was some years later; first, I worked on GNU Emacs. 260 261 GNU Emacs 262 263 I began work on GNU Emacs in September 1984, and in early 1985 it was 264 beginning to be usable. This enabled me to begin using Unix systems to 265 do editing; having no interest in learning to use vi or ed, I had done 266 my editing on other kinds of machines until then. 267 268 At this point, people began wanting to use GNU Emacs, which raised the 269 question of how to distribute it. Of course, I put it on the anonymous 270 ftp server on the MIT computer that I used. (This computer, 271 prep.ai.mit.edu, thus became the principal GNU ftp distribution site; 272 when it was decommissioned a few years later, we transferred the name 273 to our new ftp server.) But at that time, many of the interested people 274 were not on the Internet and could not get a copy by ftp. So the 275 question was, what would I say to them? 276 277 I could have said, "Find a friend who is on the net and who will make a 278 copy for you." Or I could have done what I did with the original PDP-10 279 Emacs: tell them, "Mail me a tape and a SASE, and I will mail it back 280 with Emacs on it." But I had no job, and I was looking for ways to make 281 money from free software. So I announced that I would mail a tape to 282 whoever wanted one, for a fee of $150. In this way, I started a free 283 software distribution business, the precursor of the companies that 284 today distribute entire Linux-based GNU systems. 285 286 Is a program free for every user? 287 288 If a program is free software when it leaves the hands of its author, 289 this does not necessarily mean it will be free software for everyone 290 who has a copy of it. For example, public domain software (software 291 that is not copyrighted) is free software; but anyone can make a 292 proprietary modified version of it. Likewise, many free programs are 293 copyrighted but distributed under simple permissive licenses which 294 allow proprietary modified versions. 295 296 The paradigmatic example of this problem is the X Window System. 297 Developed at MIT, and released as free software with a permissive 298 license, it was soon adopted by various computer companies. They added 299 X to their proprietary Unix systems, in binary form only, and covered 300 by the same nondisclosure agreement. These copies of X were no more 301 free software than Unix was. 302 303 The developers of the X Window System did not consider this a 304 problem--they expected and intended this to happen. Their goal was not 305 freedom, just "success", defined as "having many users." They did not 306 care whether these users had freedom, only that they should be 307 numerous. 308 309 This lead to a paradoxical situation where two different ways of 310 counting the amount of freedom gave different answers to the question, 311 "Is this program free?" If you judged based on the freedom provided by 312 the distribution terms of the MIT release, you would say that X was 313 free software. But if you measured the freedom of the average user of 314 X, you would have to say it was proprietary software. Most X users were 315 running the proprietary versions that came with Unix systems, not the 316 free version. 317 318 Copyleft and the GNU GPL 319 320 The goal of GNU was to give users freedom, not just to be popular. So 321 we needed to use distribution terms that would prevent GNU software 322 from being turned into proprietary software. The method we use is 323 called "copyleft".(1) 324 325 Copyleft uses copyright law, but flips it over to serve the opposite of 326 its usual purpose: instead of a means of privatizing software, it 327 becomes a means of keeping software free. 328 329 The central idea of copyleft is that we give everyone permission to run 330 the program, copy the program, modify the program, and distribute 331 modified versions--but not permission to add restrictions of their own. 332 Thus, the crucial freedoms that define "free software" are guaranteed 333 to everyone who has a copy; they become inalienable rights. 334 335 For an effective copyleft, modified versions must also be free. This 336 ensures that work based on ours becomes available to our community if 337 it is published. When programmers who have jobs as programmers 338 volunteer to improve GNU software, it is copyleft that prevents their 339 employers from saying, "You can't share those changes, because we are 340 going to use them to make our proprietary version of the program." 341 342 The requirement that changes must be free is essential if we want to 343 ensure freedom for every user of the program. The companies that 344 privatized the X Window System usually made some changes to port it to 345 their systems and hardware. These changes were small compared with the 346 great extent of X, but they were not trivial. If making changes were an 347 excuse to deny the users freedom, it would be easy for anyone to take 348 advantage of the excuse. 349 350 A related issue concerns combining a free program with non-free code. 351 Such a combination would inevitably be non-free; whichever freedoms are 352 lacking for the non-free part would be lacking for the whole as well. 353 To permit such combinations would open a hole big enough to sink a 354 ship. Therefore, a crucial requirement for copyleft is to plug this 355 hole: anything added to or combined with a copylefted program must be 356 such that the larger combined version is also free and copylefted. 357 358 The specific implementation of copyleft that we use for most GNU 359 software is the GNU General Public License, or GNU GPL for short. We 360 have other kinds of copyleft that are used in specific circumstances. 361 GNU manuals are copylefted also, but use a much simpler kind of 362 copyleft, because the complexity of the GNU GPL is not necessary for 363 manuals.(2) 364 365 (1) In 1984 or 1985, Don Hopkins (a very imaginative fellow) mailed me 366 a letter. On the envelope he had written several amusing sayings, 367 including this one: "Copyleft--all rights reversed." I used the word 368 "copyleft" to name the distribution concept I was developing at the 369 time. 370 371 (2) We now use the GNU Free Documentation License for documentation. 372 373 The Free Software Foundation 374 375 As interest in using Emacs was growing, other people became involved in 376 the GNU project, and we decided that it was time to seek funding once 377 again. So in 1985 we created the Free Software Foundation, a tax-exempt 378 charity for free software development. The FSF also took over the Emacs 379 tape distribution business; later it extended this by adding other free 380 software (both GNU and non-GNU) to the tape, and by selling free 381 manuals as well. 382 383 The FSF accepts donations, but most of its income has always come from 384 sales--of copies of free software, and of other related services. Today 385 it sells CD-ROMs of source code, CD-ROMs with binaries, nicely printed 386 manuals (all with freedom to redistribute and modify), and Deluxe 387 Distributions (where we build the whole collection of software for your 388 choice of platform). 389 390 Free Software Foundation employees have written and maintained a number 391 of GNU software packages. Two notable ones are the C library and the 392 shell. The GNU C library is what every program running on a GNU/Linux 393 system uses to communicate with Linux. It was developed by a member of 394 the Free Software Foundation staff, Roland McGrath. The shell used on 395 most GNU/Linux systems is BASH, the Bourne Again Shell(1), which was 396 developed by FSF employee Brian Fox. 397 398 We funded development of these programs because the GNU project was not 399 just about tools or a development environment. Our goal was a complete 400 operating system, and these programs were needed for that goal. 401 402 (1) "Bourne again Shell" is a joke on the name ``Bourne Shell'', which 403 was the usual shell on Unix. 404 405 Free software support 406 407 The free software philosophy rejects a specific widespread business 408 practice, but it is not against business. When businesses respect the 409 users' freedom, we wish them success. 410 411 Selling copies of Emacs demonstrates one kind of free software 412 business. When the FSF took over that business, I needed another way to 413 make a living. I found it in selling services relating to the free 414 software I had developed. This included teaching, for subjects such as 415 how to program GNU Emacs and how to customize GCC, and software 416 development, mostly porting GCC to new platforms. 417 418 Today each of these kinds of free software business is practiced by a 419 number of corporations. Some distribute free software collections on 420 CD-ROM; others sell support at levels ranging from answering user 421 questions, to fixing bugs, to adding major new features. We are even 422 beginning to see free software companies based on launching new free 423 software products. 424 425 Watch out, though--a number of companies that associate themselves with 426 the term "open source" actually base their business on non-free 427 software that works with free software. These are not free software 428 companies, they are proprietary software companies whose products tempt 429 users away from freedom. They call these "value added", which reflects 430 the values they would like us to adopt: convenience above freedom. If 431 we value freedom more, we should call them "freedom subtracted" 432 products. 433 434 Technical goals 435 436 The principal goal of GNU was to be free software. Even if GNU had no 437 technical advantage over Unix, it would have a social advantage, 438 allowing users to cooperate, and an ethical advantage, respecting the 439 user's freedom. 440 441 But it was natural to apply the known standards of good practice to the 442 work--for example, dynamically allocating data structures to avoid 443 arbitrary fixed size limits, and handling all the possible 8-bit codes 444 wherever that made sense. 445 446 In addition, we rejected the Unix focus on small memory size, by 447 deciding not to support 16-bit machines (it was clear that 32-bit 448 machines would be the norm by the time the GNU system was finished), 449 and to make no effort to reduce memory usage unless it exceeded a 450 megabyte. In programs for which handling very large files was not 451 crucial, we encouraged programmers to read an entire input file into 452 core, then scan its contents without having to worry about I/O. 453 454 These decisions enabled many GNU programs to surpass their Unix 455 counterparts in reliability and speed. 456 457 Donated computers 458 459 As the GNU project's reputation grew, people began offering to donate 460 machines running UNIX to the project. These were very useful, because 461 the easiest way to develop components of GNU was to do it on a UNIX 462 system, and replace the components of that system one by one. But they 463 raised an ethical issue: whether it was right for us to have a copy of 464 UNIX at all. 465 466 UNIX was (and is) proprietary software, and the GNU project's 467 philosophy said that we should not use proprietary software. But, 468 applying the same reasoning that leads to the conclusion that violence 469 in self defense is justified, I concluded that it was legitimate to use 470 a proprietary package when that was crucial for developing a free 471 replacement that would help others stop using the proprietary package. 472 473 But, even if this was a justifiable evil, it was still an evil. Today 474 we no longer have any copies of Unix, because we have replaced them 475 with free operating systems. If we could not replace a machine's 476 operating system with a free one, we replaced the machine instead. 477 478 The GNU Task List 479 480 As the GNU project proceeded, and increasing numbers of system 481 components were found or developed, eventually it became useful to make 482 a list of the remaining gaps. We used it to recruit developers to write 483 the missing pieces. This list became known as the GNU task list. In 484 addition to missing Unix components, we listed added various other 485 useful software and documentation projects that, we thought, a truly 486 complete system ought to have. 487 488 Today, hardly any Unix components are left in the GNU task list--those 489 jobs have been done, aside from a few inessential ones. But the list is 490 full of projects that some might call "applications". Any program that 491 appeals to more than a narrow class of users would be a useful thing to 492 add to an operating system. 493 494 Even games are included in the task list--and have been since the 495 beginning. Unix included games, so naturally GNU should too. But 496 compatibility was not an issue for games, so we did not follow the list 497 of games that Unix had. Instead, we listed a spectrum of different 498 kinds of games that users might like. 499 500 The GNU Library GPL 501 502 The GNU C library uses a special kind of copyleft called the GNU 503 Library General Public License(1), which gives permission to link 504 proprietary software with the library. Why make this exception? 505 506 It is not a matter of principle; there is no principle that says 507 proprietary software products are entitled to include our code. (Why 508 contribute to a project predicated on refusing to share with us?) Using 509 the LGPL for the C library, or for any library, is a matter of 510 strategy. 511 512 The C library does a generic job; every proprietary system or compiler 513 comes with a C library. Therefore, to make our C library available only 514 to free software would not have given free software any advantage--it 515 would only have discouraged use of our library. 516 517 One system is an exception to this: on the GNU system (and this 518 includes GNU/Linux), the GNU C library is the only C library. So the 519 distribution terms of the GNU C library determine whether it is 520 possible to compile a proprietary program for the GNU system. There is 521 no ethical reason to allow proprietary applications on the GNU system, 522 but strategically it seems that disallowing them would do more to 523 discourage use of the GNU system than to encourage development of free 524 applications. 525 526 That is why using the Library GPL is a good strategy for the C library. 527 For other libraries, the strategic decision needs to be considered on a 528 case-by-case basis. When a library does a special job that can help 529 write certain kinds of programs, then releasing it under the GPL, 530 limiting it to free programs only, is a way of helping other free 531 software developers, giving them an advantage against proprietary 532 software. 533 534 Consider GNU Readline, a library that was developed to provide 535 command-line editing for BASH. Readline is released under the ordinary 536 GNU GPL, not the Library GPL. This probably does reduce the amount 537 Readline is used, but that is no loss for us. Meanwhile, at least one 538 useful application has been made free software specifically so it could 539 use Readline, and that is a real gain for the community. 540 541 Proprietary software developers have the advantages money provides; 542 free software developers need to make advantages for each other. I hope 543 some day we will have a large collection of GPL-covered libraries that 544 have no parallel available to proprietary software, providing useful 545 modules to serve as building blocks in new free software, and adding up 546 to a major advantage for further free software development. 547 548 (1) This license is now called the GNU Lesser General Public License, 549 to avoid giving the idea that all libraries ought to use it. 550 See http://www.gnu.org/philosophy/why-not-lgpl.html. 551 552 Scratching an itch? 553 554 Eric Raymond says that "Every good work of software starts by 555 scratching a developer's personal itch." Maybe that happens sometimes, 556 but many essential pieces of GNU software were developed in order to 557 have a complete free operating system. They come from a vision and a 558 plan, not from impulse. 559 560 For example, we developed the GNU C library because a Unix-like system 561 needs a C library, the Bourne-Again Shell (bash) because a Unix-like 562 system needs a shell, and GNU tar because a Unix-like system needs a 563 tar program. The same is true for my own programs--the GNU C compiler, 564 GNU Emacs, GDB and GNU Make. 565 566 Some GNU programs were developed to cope with specific threats to our 567 freedom. Thus, we developed gzip to replace the Compress program, which 568 had been lost to the community because of the LZW patents. We found 569 people to develop LessTif, and more recently started GNOME and Harmony, 570 to address the problems caused by certain proprietary libraries (see 571 below). We are developing the GNU Privacy Guard to replace popular 572 non-free encryption software, because users should not have to choose 573 between privacy and freedom. 574 575 Of course, the people writing these programs became interested in the 576 work, and many features were added to them by various people for the 577 sake of their own needs and interests. But that is not why the programs 578 exist. 579 580 Unexpected developments 581 582 At the beginning of the GNU project, I imagined that we would develop 583 the whole GNU system, then release it as a whole. That is not how it 584 happened. 585 586 Since each component of the GNU system was implemented on a Unix 587 system, each component could run on Unix systems, long before a 588 complete GNU system existed. Some of these programs became popular, and 589 users began extending them and porting them---to the various 590 incompatible versions of Unix, and sometimes to other systems as well. 591 592 The process made these programs much more powerful, and attracted both 593 funds and contributors to the GNU project. But it probably also delayed 594 completion of a minimal working system by several years, as GNU 595 developers' time was put into maintaining these ports and adding 596 features to the existing components, rather than moving on to write one 597 missing component after another. 598 599 The GNU Hurd 600 601 By 1990, the GNU system was almost complete; the only major missing 602 component was the kernel. We had decided to implement our kernel as a 603 collection of server processes running on top of Mach. Mach is a 604 microkernel developed at Carnegie Mellon University and then at the 605 University of Utah; the GNU HURD is a collection of servers (or ``herd 606 of gnus'') that run on top of Mach, and do the various jobs of the Unix 607 kernel. The start of development was delayed as we waited for Mach to 608 be released as free software, as had been promised. 609 610 One reason for choosing this design was to avoid what seemed to be the 611 hardest part of the job: debugging a kernel program without a 612 source-level debugger to do it with. This part of the job had been done 613 already, in Mach, and we expected to debug the HURD servers as user 614 programs, with GDB. But it took a long time to make that possible, and 615 the multi-threaded servers that send messages to each other have turned 616 out to be very hard to debug. Making the HURD work solidly has 617 stretched on for many years. 618 619 Alix 620 621 The GNU kernel was not originally supposed to be called the HURD. Its 622 original name was Alix--named after the woman who was my sweetheart at 623 the time. She, a Unix system administrator, had pointed out how her 624 name would fit a common naming pattern for Unix system versions; as a 625 joke, she told her friends, "Someone should name a kernel after me." I 626 said nothing, but decided to surprise her with a kernel named Alix. 627 628 It did not stay that way. Michael Bushnell (now Thomas), the main 629 developer of the kernel, preferred the name HURD, and redefined Alix to 630 refer to a certain part of the kernel--the part that would trap system 631 calls and handle them by sending messages to HURD servers. 632 633 Ultimately, Alix and I broke up, and she changed her name; 634 independently, the HURD design was changed so that the C library would 635 send messages directly to servers, and this made the Alix component 636 disappear from the design. 637 638 But before these things happened, a friend of hers came across the name 639 Alix in the HURD source code, and mentioned the name to her. So the 640 name did its job. 641 642 Linux and GNU/Linux 643 644 The GNU Hurd is not ready for production use. Fortunately, another 645 kernel is available. In 1991, Linus Torvalds developed a 646 Unix-compatible kernel and called it Linux. Around 1992, combining 647 Linux with the not-quite-complete GNU system resulted in a complete 648 free operating system. (Combining them was a substantial job in itself, 649 of course.) It is due to Linux that we can actually run a version of 650 the GNU system today. 651 652 We call this system version GNU/Linux, to express its composition as a 653 combination of the GNU system with Linux as the kernel. 654 655 Challenges in our future 656 657 We have proved our ability to develop a broad spectrum of free 658 software. This does not mean we are invincible and unstoppable. Several 659 challenges make the future of free software uncertain; meeting them 660 will require steadfast effort and endurance, sometimes lasting for 661 years. It will require the kind of determination that people display 662 when they value their freedom and will not let anyone take it away. 663 664 The following four sections discuss these challenges. 665 666 Secret hardware 667 668 Hardware manufacturers increasingly tend to keep hardware 669 specifications secret. This makes it difficult to write free drivers so 670 that Linux and XFree86 can support new hardware. We have complete free 671 systems today, but we will not have them tomorrow if we cannot support 672 tomorrow's computers. 673 674 There are two ways to cope with this problem. Programmers can do 675 reverse engineering to figure out how to support the hardware. The rest 676 of us can choose the hardware that is supported by free software; as 677 our numbers increase, secrecy of specifications will become a 678 self-defeating policy. 679 680 Reverse engineering is a big job; will we have programmers with 681 sufficient determination to undertake it? Yes--if we have built up a 682 strong feeling that free software is a matter of principle, and 683 non-free drivers are intolerable. And will large numbers of us spend 684 extra money, or even a little extra time, so we can use free drivers? 685 Yes, if the determination to have freedom is widespread. 686 687 Non-free libraries 688 689 A non-free library that runs on free operating systems acts as a trap 690 for free software developers. The library's attractive features are the 691 bait; if you use the library, you fall into the trap, because your 692 program cannot usefully be part of a free operating system. (Strictly 693 speaking, we could include your program, but it won't run with the 694 library missing.) Even worse, if a program that uses the proprietary 695 library becomes popular, it can lure other unsuspecting programmers 696 into the trap. 697 698 The first instance of this problem was the Motif toolkit, back in the 699 80s. Although there were as yet no free operating systems, it was clear 700 what problem Motif would cause for them later on. The GNU Project 701 responded in two ways: by asking individual free software projects to 702 support the free X toolkit widgets as well as Motif, and by asking for 703 someone to write a free replacement for Motif. The job took many years; 704 LessTif, developed by the Hungry Programmers, became powerful enough to 705 support most Motif applications only in 1997. 706 707 Between 1996 and 1998, another non-free GUI toolkit library, called Qt, 708 was used in a substantial collection of free software, the desktop KDE. 709 710 Free GNU/Linux systems were unable to use KDE, because we could not use 711 the library. However, some commercial distributors of GNU/Linux systems 712 who were not strict about sticking with free software added KDE to 713 their systems--producing a system with more capabilities, but less 714 freedom. The KDE group was actively encouraging more programmers to use 715 Qt, and millions of new "Linux users" had never been exposed to the 716 idea that there was a problem in this. The situation appeared grim. 717 718 The free software community responded to the problem in two ways: GNOME 719 and Harmony. 720 721 GNOME, the GNU Network Object Model Environment, is GNU's desktop 722 project. Started in 1997 by Miguel de Icaza, and developed with the 723 support of Red Hat Software, GNOME set out to provide similar desktop 724 facilities, but using free software exclusively. It has technical 725 advantages as well, such as supporting a variety of languages, not just 726 C++. But its main purpose was freedom: not to require the use of any 727 non-free software. 728 729 Harmony is a compatible replacement library, designed to make it 730 possible to run KDE software without using Qt. 731 732 In November 1998, the developers of Qt announced a change of license 733 which, when carried out, should make Qt free software. There is no way 734 to be sure, but I think that this was partly due to the community's 735 firm response to the problem that Qt posed when it was non-free. (The 736 new license is inconvenient and inequitable, so it remains desirable to 737 avoid using Qt.) 738 739 [Subsequent note: in September 2000, Qt was rereleased under the GNU 740 GPL, which essentially solved this problem.] 741 742 How will we respond to the next tempting non-free library? Will the 743 whole community understand the need to stay out of the trap? Or will 744 many of us give up freedom for convenience, and produce a major 745 problem? Our future depends on our philosophy. 746 747 Software patents 748 749 The worst threat we face comes from software patents, which can put 750 algorithms and features off limits to free software for up to twenty 751 years. The LZW compression algorithm patents were applied for in 1983, 752 and we still cannot release free software to produce proper compressed 753 GIFs. In 1998, a free program to produce MP3 compressed audio was 754 removed from distribution under threat of a patent suit. 755 756 There are ways to cope with patents: we can search for evidence that a 757 patent is invalid, and we can look for alternative ways to do a job. 758 But each of these methods works only sometimes; when both fail, a 759 patent may force all free software to lack some feature that users 760 want. What will we do when this happens? 761 762 Those of us who value free software for freedom's sake will stay with 763 free software anyway. We will manage to get work done without the 764 patented features. But those who value free software because they 765 expect it to be techically superior are likely to call it a failure 766 when a patent holds it back. Thus, while it is useful to talk about the 767 practical effectiveness of the "cathedral" model of development (1), 768 and the reliability and power of some free software, we must not stop 769 there. We must talk about freedom and principle. 770 771 (1) It would have been clearer to write `of the "bazaar" model', since 772 that was the alternative that was new and initially controversial. 773 774 Free documentation 775 776 The biggest deficiency in our free operating systems is not in the 777 software--it is the lack of good free manuals that we can include in 778 our systems. Documentation is an essential part of any software 779 package; when an important free software package does not come with a 780 good free manual, that is a major gap. We have many such gaps today. 781 782 Free documentation, like free software, is a matter of freedom, not 783 price. The criterion for a free manual is pretty much the same as for 784 free software: it is a matter of giving all users certain freedoms. 785 Redistribution (including commercial sale) must be permitted, on-line 786 and on paper, so that the manual can accompany every copy of the 787 program. 788 789 Permission for modification is crucial too. As a general rule, I don't 790 believe that it is essential for people to have permission to modify 791 all sorts of articles and books. For example, I don't think you or I 792 are obliged to give permission to modify articles like this one, which 793 describe our actions and our views. 794 795 But there is a particular reason why the freedom to modify is crucial 796 for documentation for free software. When people exercise their right 797 to modify the software, and add or change its features, if they are 798 conscientious they will change the manual too--so they can provide 799 accurate and usable documentation with the modified program. A manual 800 which does not allow programmers to be conscientious and finish the 801 job, does not fill our community's needs. 802 803 Some kinds of limits on how modifications are done pose no problem. For 804 example, requirements to preserve the original author's copyright 805 notice, the distribution terms, or the list of authors, are ok. It is 806 also no problem to require modified versions to include notice that 807 they were modified, even to have entire sections that may not be 808 deleted or changed, as long as these sections deal with nontechnical 809 topics. These kinds of restrictions are not a problem because they 810 don't stop the conscientious programmer from adapting the manual to fit 811 the modified program. In other words, they don't block the free 812 software community from making full use of the manual. 813 814 However, it must be possible to modify all the *technical* content of 815 the manual, and then distribute the result in all the usual media, 816 through all the usual channels; otherwise, the restrictions do obstruct 817 the community, the manual is not free, and we need another manual. 818 819 Will free software developers have the awareness and determination to 820 produce a full spectrum of free manuals? Once again, our future depends 821 on philosophy. 822 823 We must talk about freedom 824 825 Estimates today are that there are ten million users of GNU/Linux 826 systems such as Debian GNU/Linux and Red Hat Linux. Free software has 827 developed such practical advantages that users are flocking to it for 828 purely practical reasons. 829 830 The good consequences of this are evident: more interest in developing 831 free software, more customers for free software businesses, and more 832 ability to encourage companies to develop commercial free software 833 instead of proprietary software products. 834 835 But interest in the software is growing faster than awareness of the 836 philosophy it is based on, and this leads to trouble. Our ability to 837 meet the challenges and threats described above depends on the will to 838 stand firm for freedom. To make sure our community has this will, we 839 need to spread the idea to the new users as they come into the 840 community. 841 842 But we are failing to do so: the efforts to attract new users into our 843 community are far outstripping the efforts to teach them the civics of 844 our community. We need to do both, and we need to keep the two efforts 845 in balance. 846 847 "Open Source" 848 849 Teaching new users about freedom became more difficult in 1998, when a 850 part of the community decided to stop using the term "free software" 851 and say "open source software" instead. 852 853 Some who favored this term aimed to avoid the confusion of "free" with 854 "gratis"--a valid goal. Others, however, aimed to set aside the spirit 855 of principle that had motivated the free software movement and the GNU 856 project, and to appeal instead to executives and business users, many 857 of whom hold an ideology that places profit above freedom, above 858 community, above principle. Thus, the rhetoric of "open source" focuses 859 on the potential to make high quality, powerful software, but shuns the 860 ideas of freedom, community, and principle. 861 862 The "Linux" magazines are a clear example of this--they are filled with 863 advertisements for proprietary software that works with GNU/Linux. When 864 the next Motif or Qt appears, will these magazines warn programmers to 865 stay away from it, or will they run ads for it? 866 867 The support of business can contribute to the community in many ways; 868 all else being equal, it is useful. But winning their support by 869 speaking even less about freedom and principle can be disastrous; it 870 makes the previous imbalance between outreach and civics education even 871 worse. 872 873 "Free software" and "open source" describe the same category of 874 software, more or less, but say different things about the software, 875 and about values. The GNU Project continues to use the term "free 876 software", to express the idea that freedom, not just technology, is 877 important. 878 879 Try! 880 881 Yoda's philosophy ("There is no `try'") sounds neat, but it doesn't 882 work for me. I have done most of my work while anxious about whether I 883 could do the job, and unsure that it would be enough to achieve the 884 goal if I did. But I tried anyway, because there was no one but me 885 between the enemy and my city. Surprising myself, I have sometimes 886 succeeded. 887 888 Sometimes I failed; some of my cities have fallen. Then I found another 889 threatened city, and got ready for another battle. Over time, I've 890 learned to look for threats and put myself between them and my city, 891 calling on other hackers to come and join me. 892 893 Nowadays, often I'm not the only one. It is a relief and a joy when I 894 see a regiment of hackers digging in to hold the line, and I realize, 895 this city may survive--for now. But the dangers are greater each year, 896 and now Microsoft has explicitly targeted our community. We can't take 897 the future of freedom for granted. Don't take it for granted! If you 898 want to keep your freedom, you must be prepared to defend it. 899 900 Copyright (C) 1998 Richard Stallman 901 902 Verbatim copying and distribution of this entire article is permitted 903 in any medium, provided this notice is preserved. 904