History of LISP, Emacs, Symbolics (Daniel Weinreb Rebut Richard Stallman)
The following is written by Daniel Weinreb (1959 to 2012). He is co-founder of Symbolics, and wrote the EINE for Lisp Machine, the first Emacs with graphical user interface, and the first emacs written in lisp.
Originally from weinreb's blog at
http://danweinreb.org/blog/rebuttal-to-stallmans-story-about-the-formation-of-symbolics-and-lmi
accessed on 2012-09-08.
Weinreb Died in 2012, and spammer got his website.
〔see Lisp Programer Daniel Weinreb Died (1959 to 2012)〕
Rebuttal to Stallman's Story About The Formation of Symbolics and LMI
by Daniel Weinreb, 2007-11
Richard Stallman has been telling a story about the origins of the Lisp machine companies, and the effects on the M.I.T. Artificial Intelligence Lab, for many years. He has published it in a book, and in a widely-referenced paper, which you can find at http://www.gnu.org/gnu/rms-lisp.html .
His account is highly biased, and in many places just plain wrong. Here's my own perspective on what really happened.
Richard Greenblatt's proposal for a Lisp machine company had two premises. First, there should be no outside investment. This would have been totally unrealistic: a company manufacturing computer hardware needs capital. Second, Greenblatt himself would be the CEO. The other members of the Lisp machine project were extremely dubious of Greenblatt's ability to run a company. So Greenblatt and the others went their separate ways and set up two companies.
Stallman's characterization of this as “backstabbing”, and that Symbolics decided not “not have scruples”, is pure hogwash. There was no backstabbing whatsoever. Symbolics was extremely scrupulous. Stallman's characterization of Symbolics as “looking for ways to destroy” LMI is pure fantasy.
Stallman claims that Symbolics “hired away all the hackers” and that “the AI lab was now helpless” and “nobody had envisioned that the AI lab's hacker group would be wiped out, but it was” and that Symbolics “wiped out MIT”. First of all, had there been only one Lisp machine company as Stallman would have preferred, exactly the same people would have left the AI lab. Secondly, Symbolics only hired four full-time and one part-time person from the AI lab (see below).
Stallman goes on to say: “So Symbolics came up with a plan. They said to the lab, ‘We will continue making our changes to the system available for you to use, but you can't put it into the MIT Lisp machine system. Instead, we'll give you access to Symbolics' Lisp machine system, and you can run it, but that's all you can do.’” In other words, software that was developed at Symbolics was not given away for free to LMI. Is that so surprising? Anyway, that wasn't Symbolics's “plan”; it was part of the MIT licensing agreement, the very same one that LMI signed. LMI's changes were all proprietary to LMI, too.
Next, he says: “After a while, I came to the conclusion that it would be best if I didn't even look at their code. When they made a beta announcement that gave the release notes, I would see what the features were and then implement them. By the time they had a real release, I did too.” First of all, he really was looking at the Symbolics code; we caught him doing it several times. But secondly, even if he hadn't, it's a whole lot easier to copy what someone else has already designed than to design it yourself. What he copied were incremental improvements: a new editor command here, a new Lisp utility there. This was a very small fraction of the software development being done at Symbolics.
His characterization of this as “punishing” Symbolics is silly. What he did never made any difference to Symbolics. In real life, Symbolics was rarely competing with LMI for sales. LMI's existence had very little to do with Symbolics's bottom line.
And while I'm setting the record straight, the original (TECO-based) Emacs was created and designed by Guy L. Steele Jr. and David Moon. After they had it working, and it had become established as the standard text editor at the AI lab, Stallman took over its maintenance.
Here is the list of Symbolics founders. Note that Bruce Edwards and I had worked at the MIT AI Lab previously, but had already left to go to other jobs before Symbolics started. Henry Baker was not one of the “hackers” of which Stallman speaks.
- * Robert Adams (original CEO, California)
- * Russell Noftsker (CEO thereafter)
- * Minoru Tonai (CFO, California)
- * John Kulp (from MIT Plasma Physics Lab)
- * Tom Knight (from MIT AI Lab)
- * Jack Holloway (from MIT AI Lab)
- * David Moon (half-time as MIT AI Lab)
- * Dan Weinreb (from Lawrence Livermore Labs)
- * Howard Cannon (from MIT AI Lab)
- * Mike McMahon (from MIT AI Lab)
- * Jim Kulp (from IIASA, Vienna)
- * Bruce Edwards (from IIASA, Vienna)
- * Bernie Greenberg (from Honeywell CISL)
- * Clark Baker (from MIT LCS)
- * Chris Terman (from MIT LCS)
- * John Blankenbaker (hardware engineer, California)
- * Bob Williams (hardware engineer, California)
- * Bob South (hardware engineer, California)
- * Henry Baker (from MIT)
- * Dave Dyer (from USC ISI)
Tags: [12] LMI, [13] MIT, [14] Symbolics
This entry was posted on Sunday, November 11th, 2007 at 3:15 pm and is filed under [15] Symbolics. You can follow any responses to this entry through the [16] RSS 2.0 feed. Responses are currently closed, but you can [17] trackback from your own site.
Replies
Sp3w Says:
November 12th, 2007 at 9:10 pm
[…] Stallman was wrong […]
Moon Says:
November 15th, 2007 at 9:08 pm
All true, so far as I can remember.
But in all fairness I have to say that Stallman greatly improved Emacs after he “liberated” it from Guy and me.
Also, where you say “Symbolics only hired four full-time and one part-time person from the AI lab (see below).” that was true at the founding. But I think some other people left the AI lab for Symbolics later. Was that Symbolics pull or Stallman push?
Paddy Murphy Says:
November 15th, 2007 at 9:32 pm
“it was part of the MIT licensing agreement, the very same one that LMI signed. LMI's changes were all proprietary to LMI, too.”
You say that as if it excuses everything. RMS's POINT is that restrictive software licensing is evil in itself – “without copyright the GPL would be unenforceable. It would also be unnecessary”. You may not agree it's evil to abide by the law, but if the law is unjust, it can be.
By following a proprietary model, symbolics (and lmi) doomed lisp machines. Doomed in much the same way, though a different market, as Amiga and BeOS. The only surviving lisp machine is RMS' emacs VM.
dlweinreb Says:
November 16th, 2007 at 2:31 am
Paddy Murphy: I see, if you had been Symbolics, you would not have abided by the law. You would have violated the MIT license, been sued, and had your company shut down. THAT would have doomed Lisp machines.
Is it your contention that “following a proprietary model” leads to doom? The reason Amiga and BeOS failed, but Apple and Microsoft succeeded, was because, um, what, again? Actually Symbolics was quite successful for a long time; there were many, many reasons why it eventually failed. Referring to GNU Emacs as a “Lisp machine” is entirely disingenuous.
In 1980, all software was proprietary. The phrases “free software” and “open source” had yet to be coined. In fact, software that was portable from one vendor's computer to another had yet to be invented.
Anyway, I totally reject Stallman's contention that restrictive software licensing is “evil in itself”.
TJIC Says:
November 16th, 2007 at 3:52 am
Great write-up, Dan. I idealized Stallman a bit, before I met him, and engaged him in a few email dialogues. Now that I've got a better feel for the man, I appreciate that much of his narratives are a bit self-serving and simplified. Having chatted a few times with some of the other characters on the list (Tom Knight was around Permabit a fair bit, and he was a thesis adviser for a good friend of mine), that only loads on the “Stallman's story doesn't pass the sniff test” accumulator.
Foo Says:
November 16th, 2007 at 9:04 am
Why Symbolics failed is an interesting question, but I guess it has nothing to do with so-called “free software”.
Maybe this paper is relevant: http://www.sts.tu-harburg.de/~r.f.moeller/symbolics-info/ai-business.pdf
Maybe Dan has some comments on that?
Alexander Shopov Says:
November 16th, 2007 at 10:06 am
Dear Mr. Weinreb,
Thank you very much for your post. It has been an interesting read.
I would like to ask you several questions and hope that you might some more clarity.
1. Why are you telling your side of the story?
2. Why did it take so long for you to do this?
3. I would believe that you will be challenged by many people (I would point to the post of Paddy Murphy). Such challenges frequently escalate. Doubtlessly you have thought about this and could you share you thoughts?
4. Can other people support your side? Actually – I am not so much interested in the dispelling the myth investigative side (find the truth) as much as the way a myth gets created and woven, repeated, changed, etc.
5. What is your current attitude to open source and free software?
6. What has been your attitude towards the open source and free software movements?
Once again: thank you for your post and I hope there would be a follow up.
Kind regards: al_shopov
dlweinreb Says:
November 16th, 2007 at 1:14 pm
In reply to “Foo”: thank you very much for giving me the idea for my next blog posting. I hope you enjoy it.
In reply to Mr. Shopov: I'm telling my side of the story because I feel that I and my friends have been unfairly attacked (defamed, slandered, whatever). It took so long to do it because I only recently established a blog for the first time and got around to it. As for challenges, I'll deal with them as they come. Yes, other people can support my side; see the posting above from Dave Moon, another Symbolics founder.
My attitude toward open source and free software would take a long time to discuss, and I'm not sure I have much to say that's novel. I use lots of open source software and I'm very grateful for its existence; I'd be happy to produce some myself if I get the opportunity; I'm not a supporter of the philosophy behind the GPL although I do see a lot of the points. But the software I write for a living is not going to be open source (in fact we won't sell it at all; we're a service provider) and I don't think there's anything remotely unethical about it or else I would not do it.
Kamen Tomov Says:
November 16th, 2007 at 3:41 pm
Mr. Weinreb,
I am grateful to you that you wrote this article. I never believed that the people in Symbolics were as unethical as Stallman describes them. Usually the talented and creative people have very high moral standards. I hope that your article would clear the name of the company that produced the finest piece of software known. I also hope that your active position would be a driving force in clearing the name of Common Lisp as well.
With respect, Kamen Tomov
Alexander Shopov Says:
November 16th, 2007 at 3:54 pm
Mr. Weinreb,
Thank you for the answers. I might have left you with the impression that I was judgmental but I fully respect your opinion and I can understand the feeling of being singled out and unfairly defamed.
I did not question you ethics but alas it seems that things are quite heated these days and everyone is making similar statements which is the topic I am interested in. Still – your look at those times was quite informative for me.
I really wonder what is the opinion of the other participants in the story that has turned into a myth.
Can someone give an answer to Mr. Moon's question on the Symbolics pull/Stallman's push for the fate of AI lab?
Kind regards: al_shopov
Paddy Murphy Says:
November 16th, 2007 at 4:32 pm
“In 1980, all software was proprietary. ”
Hah. Only in November 1980 was it even decided definitively that software was copyrightABLE. Before that, it was mostly trade-secret (see: unix), and thus typically copyable, patchable, etc. once disclosed – a much better situation for software users/ computer owners. Sure, some vendors might have _tried_ to claim copyright, but it was not clear they had a leg to stand on. Until Nov. 1980, when Congress sided with the copyrightists.
The MIT/Symbolics betrayal was an effective endorsement of the new 1980s regime of software-as-a-definitively-copyrightable-entity: You guys, seeing the proposed legalisation of slavery, chose to become slave-owners, instead of trying to fight slavery.
You might have no ethical problem with copyrighting or even patenting software, but clearly plenty of people do. And I think history has already proven RMS right. The “cambrian extinction event” of microsoft would hardly have happened if disclosed software was freely shareable.
Stavros Macrakis Says:
November 16th, 2007 at 6:01 pm
Paddy Murphy, why do you classify software protected by trade secret law less proprietary than software protected by copyright? Although some vendors shared their source code, many didn't, and in any case, in a time when there was hardly any portable systems software, any use of a vendor's software implied use of its hardware.
I'll ignore the rest of your rantXXX post.
Eric Says:
November 16th, 2007 at 7:49 pm
You might have no ethical problem with copyrighting or even patenting software, but clearly plenty of people do.
That's like saying, “You may not have an ethical problem with eating eggs and dairy, but clearly plenty of people do.” Technically true, but that contrast doesn't demonstrate anything. The implied appeal to popularity is of course a logical fallacy, but it's also absurd because it's an appeal to subcultural popularity – PETA and FSF are fringe organizations, and both vegans and people who straightfacedly equate copyrighting software with slavery are rather tiny minorities.
Bernie Greenberg Says:
November 16th, 2007 at 8:44 pm
My dear friend Dan's statements about the origin and precedence of Lisp Emaces are, of course, 100% accurate. Both Dan and Richard were great fans of what I was doing. Accurate contemporary (1979) (and embarrassingly dated!) reportage about the origin and then-significance of Multics Emacs can be found at http://www.multicians.org/mepap.html
For the record, I agree with all of Dan's observations.
I would also like to add that the notion of “hiring away” adults as though they had no free will or say in the matter has always offended me.
David Chapman Says:
November 17th, 2007 at 4:32 am
I was at the MIT AI Lab and peripherally involved in the Lisp machine development effort at this time. (Hi, Dan!) I contributed a small amount of code (and a lot of bug reports). I never worked for either company, and was friendly with people from both.
Dan's account is accurate.
Also, although I was not a Symbolics employee and have only second-hand knowledge of how it failed, his account in his next posting accords with what I heard at the time.
It is an appalling travesty that the best available software development tools in 2007 are much less good than what we had in 1987. Sure wish I could develop for Zetalisp today.
I had not previously read RMS's account. I found it interesting, too.
Something that RMS does not make explicit, that I think was very important, was that we were working in a taxpayer-funded university lab. Obviously the software we wrote was and should have been made freely available to everyone because everyone had paid for it. The “spirit of freewheeling cooperation” was indeed a great thing; it is (or should be) the way all university research is done, in any field.
There's a bit of a logical missing link in extending that ethos outside the university. It is great when it can and does happen. (I personally write and support GPL FLOSS, as well as of course using gnu emacs and other FLOSS systems every day.) I don't think it follows that all software can or should be developed that way.
(As far as I know, I'm the only person besides Paul Graham who has successfully followed his recipe of building a company on proprietary LISP code and selling it for zillions of dollars.)
I wish that RMS had built a stand-alone LISP into gnu early on, and that it had become the basis for a bunch of other stuff. The world would be a much better place now. Oh, well; I bet he wishes that too.
An importantly true point in RMS's piece: “Why did he want to avoid outside investment? Because when a company has outside investors, they take control.” Dan acknowledges that this is part of why Symbolics failed. Like Greenblatt, I bootstrapped my company using only cashflow from sales for precisely this reason. On the other hand, most hackers don't make good business people. (I thought I would, and I was right, luckily.) And, it's harder to do this for a hardware company than a software one, because of the capital requirements.
Back to a couple points I made earlier. Software development systems, including languages and IDEs, are worse than they were in 1987. That's an anomaly. (I haven't read the book of RPG's that Dan recommends in his next posting, extending his “worse is better” essay; maybe it has the answer.) I write a lot of PHP code these days. What a pathetic excuse for a programming system.
The answer may be, in 1987, most significant software development was done at university computer science departments, which were full of smart people who had actually taken a programming languages course and knew how not to make the same language-design mistakes over and over.
For some reason nowadays CS departments don't seem to develop influential programs. I haven't been near a university in most of 20 years, so I don't know why. Maybe they are producing great stuff that no one uses for some reason, or maybe they don't attract smart people any longer.
Most influential programs now seem to be produced by amateurs. That's weird. It's great, in a way, that you don't need any credentials to create wonderful things millions of people will use. On the other hand, the CS curriculum really is critical. Learning it would save you from inventing PHP, which would be a huge benefit to mankind.
Anyway, thank you Dan for writing this. I still think about those days often.
dlweinreb Says:
November 17th, 2007 at 12:41 pm
Mr Shopov, about what Dave Moon said: I can't remember any other AI Lab employees who later joined Symbolics, although it's been a long time now and I'm not sure. But Stallman's story is really about the initial formation of Symbolics and LMI.
Mr Murphy: Whether the MIT contract used the legal doctrine of “contract” or not, Symbolics was contractually bound not to give away the source code. There was absolutely nothing unusual or novel about that. Your idea that until 1990, all code was copyable and patchable “once disclosed” is wrong. The very example you point out, Unix, demonstrates this clearly. Unix sources had “escaped” but that didn't change the legal ownership of Unix, and some people were successfully sued for using it. To this day, Unix is legally owned. The people who are teaching you history are misleading you badly.
Hi David (Chapman)! Good to hear from you! I'll add that in all fairness, I don't think the investors were to investors, per se. By the time the problems set in, Symbolics had already gone public, and the influence of the initial venture capitalists as such was less strong, except insofar as some of them were still on the board of directors. I do blame the board for having installed incompetent CEO's after Russell Noftsker and John Kulp were gone.
I agree completely that a startup is much better off if it can avoid outside investors. But my (then-modest) salary had to come from somewhere, and the capital to build a hardware company fast enough to get to market while you're still relevant is unavoidable. ICAD is another company that pulled off a great feat of bootstrapping, to their great advantage. Phil Greenspun's Ars Digita story is a classic of how obtaining venture capital can hurt your business. Not that I'm opposed to venture capital! You just have to know what you're getting into when you accept it.
Your point about where influential new software comes from these days is very interesting. I agree that it's great that an uncredentialed (and unaffiliated) amateur can have so much influence. It might be illuminating to do an informal survey. Erlang, as I understand it, was largely developed at Ericcson, although Joe Armstrong's Ph.D. work at the Royal Institute of Technology (Stockholm, Sweden) is also seminal. Does anyone know if “Matz” was affiliated with any institution when he invented Ruby? Groovy was done entirely by amateurs. The World Wide Web was invented by Tim Berners-Lee when he was at CERN, which is basically academic. What examples did you have in mind?
On the other hand, the MIT Lisp machine work wasn't really part of the mission of the AI Lab. It wasn't artificial intelligence research, as we and the Lab saw it. We were just the support group, being responsive to the needs of our user community (“customers”). I sure didn't have any credentials! In fact, when I started on the project, I had never written any software except classroom exercises and little toys for myself. I was incredibly lucky to become apprentice to Dave Moon, and learn from Bernie Greenberg and so many others. During the MIT years (1976-1979, for me) I had no idea that Lisp machines would ever be used beyond MIT.
Dave Says:
November 17th, 2007 at 10:31 pm
Thanks for publishing more details about the deep history of popular computing :-)
However, I think its a bit harsh to lay into Richard over “lies” in a speech; Richard always ad-libs his speeches and so it is certain to be full of errors.
Richard is very up front that his memory of his life isn't very good and he's forgotten most of the details of what has happened, especially for things 10, 20 years ago, as in this case.
The book “Hackers” by Steven Levy is the authoritative book about the AI Lab, and Richard's written article at http://www.gnu.org/gnu/thegnuproject.html about the AI Lab recommends it.
I think you've mistaken incompetence for malice.
Dave Says:
November 17th, 2007 at 10:32 pm
Anyway, I totally reject Stallman's contention that restrictive software licensing is “evil in itself”.”
I think its more factual to say Richard's contention is that it is wrong for software developers to have power over software users. Evil is something of a strange concept, and I shy away from using it in general… I think its inappropriate for merely restrictive software licensing, because it overeggs the cake in these days of software idea patents and anti-circumvention law. Its similar to the difference between racism and apartheid; one is to be criticized and curbed but ultimately tolerated in a free society, and the other is unjust law that needs to be abolished in a free society.
Understanding » Blog Archive » Stallmans Story About The Formation Of Symbolics and LMI Says:
November 17th, 2007 at 11:20 pm
[…] Dan Weinreb recently posted a “rebuttal” that I thought was quite interesting (thanks to Gerry for point it out :-) […]
dlweinreb Says:
November 18th, 2007 at 12:31 am
Dave: Stallman certainly doesn't present his story by saying that he doesn't remember the facts very well. I find it entirely implausible that he “forgot” the Lisp machine version of Emacs (EINE), since he used it frequently and even modified it. Perhaps, as your blog says, Stallman has said that he doesn't remember his childhood well, but that's a different thing entirely.
Levy's “Hackers” isn't authoritative to me. I was there, and I know what happened. Levy obviously interviewed Stallman, as you can tell from reading the book, and evidently he took much of what Stallman said at face value. The chapter in the book is called “The Last of the True Hackers” and is written primarily from Stallman's point of view (although he does quote many other people). It make a good story, the kind that a journalist would naturally be looking for. It's no wonder Stallman's written article recommends the book. This is exactly why I wrote this blog entry in the first place!
Where do you think Levy got the part about “RMS had single-handedly attempted to match the work of over a dozen world-class hackers”? From Stallman himself, or Greenblatt. It's quite misleading. First of all, most of the software effort at Symbolics was new software that Stallman did not match, such as Bernie Greenberg's LMFS file system, new software aimed at the next-generation hardware (the 3600), internal software to support the hardware manufacturing process, and so on. Stallman copied mainly incremental improvements. Secondly, most of the work entailed in making those improvements was their design; Stallman didn't have to do that, since he just copied the design that we spent time working on. The story of Stallman being ten times as good as any of the rest of us is simply self-serving.
Levy says, “But Symbolics, in Stallman's view, had purposely stripped the lab of its hackers in order to prevent them from donating competing technology to the public domain.” Levy doesn't question this. What really happened is much simpler: Symbolics hired lots of good hackers for the same reason that every other company hires lots of good hackers: to get the work done.
The split between Symbolics and LMI was indeed painful for everyone, as Levy says, but he explains pretty well why it happened.
Adrienne Says:
November 18th, 2007 at 2:19 pm
It seems that RMS has seen your blog. In the article to which you refer “My Lisp Experiences and the Development of GNU Emacs” at http://www.gnu.org/gnu/rms-lisp.html (Stallman's speech at ILC on 28 October 2002) footnote (5) states:
(5) Bernie Greenberg says that Dan Weinreb's implementation of Emacs for the Lisp Machine came before Greenberg's implementation for Multics. I apologize for the mistake.
RMS is certainly not being malicious but, Dave, he is hardly incompetent either or he could not have “… single-handedly attempted to match the work of over a dozen world-class hackers” ;-)
Cheers
Kamen Tomov Says:
November 18th, 2007 at 5:33 pm
In fact there is plenty of evidence that supports Mr. Chapman's point that most of the influential programs are produced by amateurs.
Take Altair Basic, MS-DOS, and Windows. Neither Gates nor Allen had formal education in computer science when they released these products. They wouldn't probably worked on them if they had.
Another evidence is Linux. Linus Torvalds started university in 1988, first released Linux in 1991 and got his degree in 1996. He originally developed the Linux kernel as a hobby OS.
Richard Gabriel has a good analysis on that.
Thomas Lord Says:
November 18th, 2007 at 7:08 pm
I first heard and read Stallman's stories as an undergraduate at a different major university — the late 1980s. Stallman was a hot topic around campus because of GCC, GDB, Emacs and GNU and was often described disparagingly but, these stories weren't themselves often discussed (so I assumed, then, that they were basically true).
This isn't the first time I've heard the responses from other people who were there but it is nice to see a lot of those collected in one place. A truer picture is certainly painted here.
I resent the original stories because of the influence they had on important career decisions. I was an employee for the FSF a few times and often did work for the GNU project. Originally, this was motivated by a false belief in Stallman's abilities as a hacker: that with someone of his alleged skill at the helm, the GNU software development project (to build a complete unix system with a core lisp component) could proceed in a rational, efficient way to completion. Instead, I watched over the years as the GNU project pissed away one effort after another until, ultimately, the “complete system” project has been privatized, as have most of the interesting so-called open-source projects.
If the goal of the Free Software Movement has been to free users from the control of vendors, then today's open source systems have achieved that goal in theory only. In practice, users are utterly dependent on a small number of open source vendors and rarely have the realistic option to evade the control of those vendors by just “using the source”. Stallman changed the competitive landscape for large vendors of systems software and that's helped lower the price of some software commodities but he so far hasn't done much to increase the software freedoms directly experienced by all that many users.
The GNU software development project had but two organizational successes, by my accounting: First, it solicited and received contributions of small pieces of code, each implementing some set of features found on traditional unix manual pages. Second, it solicited and received contributions of translations for program messages and documentation. Those are impressive, but they hardly display the leadership prowess one would expect of a hero who can slay 10 symbolics hackers with a single all nighter. Nor did the GNU project ever manage to *package* these developments in a way that would make their utility as free software available to average users: the vendor role emerges from that.
Anyone who had had a little experience, back when I was getting started, would have seen that coming. That is, looking at the technical agenda that the GNU project laid out in support of its political agenda, the experienced eye would have said “Yeah, right,” and not been satisfied with anything less than a fairly clear plan for how all of that would get built and delivered into the hands of users in an orderly, efficient way. But, to the inexperienced eye, there's no problem — because there's a super-hacker at the helm. It's just a small matter of hacking.
So, I think those stories and the corrections to them matter. Thanks to those who were there for collecting their versions.
-t
P.s.: Not meaning to disrupt the solemnity of the recounting but just for the opportunity to address some interesting people from the LISP world — you might like something I've been working on (“XQVM”). It is a virtual machine a bit like an idealized machine for Scheme, but is purely functional and uses XML Data Model instances rather than CONS pairs and ATOMs. If you install it as a web service, it's a bit like having a lisp reader as a web service (which you can then customize to be some more specific service, of course). And then there's additional things you can build on top of that. (It's on my web site.)
Dave Says:
November 18th, 2007 at 8:00 pm
@Adrienne: I don't mean to cuss Richard for being incompetent in general, obviously – I just mean he's human and none of our memories are perfect. Although the mythology around Richard makes him out to be have superhuman hacker powers, “single-handedly matched the work of over a dozen world-class hackers”, “wrote GCC 1.0 on his own,” blah blah, he isn't really superhuman and its not surprising he forgets the fine details of things that happened over 20 years ago.
@Dan: Similarly, its been about a decade since I read Levy's book (and I'm 24), and I'd love to read it again. I bet Levy does retell the story Richard told him, the way Richard told it, without questioning much. But what I mean is, in general, when we want to know who did what and when, we're better off referring to authored and formally published texts, not informal transcriptions of semi-formal speeches. Some texts are more fact-checked than others, and a text on a webpage is almost always worthless.
Really my point is that its too harsh to accuse Richard of lying in his speeches when his facts are wrong. If Richard is talking about the history of copyright and mentions the printing culture of venice hundreds of years ago, I think it would be silly to claim that when he gets the details wrong he's “lying.” What he says isn't totally accurate, but its a speech, not an authored and fact-checked text, and he's telling a mythological ripping yarn with goodies and baddies. Isn't that kind of storytelling run-of-the-mill practice for starting a social movement?
That Richard has clarified what he meant by “hiring away” and even apologised for a factual error in the footnotes recently added to the end of transcript says a lot about him, I think: Not a liar, not superhuman, just human.
Dan, are there any other books about these events, that are more authoritative? :-)
Adrienne Says:
November 19th, 2007 at 12:11 am
Dave:
‘That Richard has clarified what he meant by “hiring away”’
Where might I find this clarification from RMS?
Dave Says:
November 19th, 2007 at 12:17 am
@Adrienne: “Greenblatt's plan, as I understood it, was to hire lab people part time, so that they could continue working at the AI Lab. Symbolics hired them full time instead, so they stopped working at MIT.” is now at the bottom of http://www.gnu.org/gnu/rms-lisp.html
Sherry Finkel Murphy Says:
November 19th, 2007 at 1:50 am
OMG, a friend who has the memory of an elephant pointed me to this blog and I'm smiling as I read it. While I spent a mere year at Symbolics as a technical writer, I assure you that I'm every bit the adoring groupie that I was then. I want to say hi particularly to Bernie (I always wanted to grab your glasses, clean them and hand them back to you, but didn't have the nerve!) and Clark–simply one of the nicest human beings on earth. I apologize that I can't remember the night-owl who wrote the Interlisp conversion tool and whose project I worked on, but I bet you do. And I certainly hope you are all well and happy.
dlweinreb Says:
November 19th, 2007 at 12:47 pm
In light of Stallman's footnote about the precedence of Emacs editors, I have removed my statement in the post, that included the word “lie”. I really should not have said that in the first place; I got carried away in the heat of writing the post. @Dave: Actually, as we saw it, Symbolics was being particularly ethical and scrupulous by hiring people full time! The experienced founders, such as Tom Knight and Jack Holloway, said that they had previously (I don't know exactly where) seen MIT-spinoff companies that took advantage of MIT by using MIT's facilities. We, in contrast, were going to make an absolutely clean break, avoiding any reality or perception that we were unfairly using MIT. I remember once Tom had to borrow an unusual tool (a telephone “punch-down”) from the AI lab for a few hours, and we all felt a bit guilty about it (no kidding). It was a big deal to us at the time.
And I'm sure we could not have accomplished what we did had we been working part-time. We worked very long hours, like many founders of new technology companies, and just managed to succeed in the way we wanted to. There was plenty of work to do, and we needed everyone we could get, full-time. During those days, I wrote documentation, went to trade shows to demo the machine, went to customer sites to do installation and support, and so on, perhaps even more than writing software. It was “all hands on deck” with everybody pitching in to do what needed to be done. (It was lots of fun, though!)
Although Stallman characterizes Symbolics's founding as “aggression against the AI lab” (see footnote 2 of the above-cited paper), I would be surprised to hear that anyone else thought of it that way at the time. I simply don't remember anger at Symbolics from anyone besides Stallman and the LMI people. As far as the AI lab was concerned, they continued to get everything they used to get, including all sources. They could see that the AI lab had been acting as a computer factory (they made lots of original CADR Lisp machines), and that that was inappropriate. Symbolics provided much better hardware and software than the AI lab could have produced itself, since we had a real factory, and a very large number of hackers (many more than the founding group, that is!). I think most people at the AI lab were happy (or indifferent) about the whole situation, although of course everyone on all sides regretted the Symbolics/LMI split.
@Sherry: Hi! Ah, yes, the InterLisp conversion tool. (Yet another piece of software that Stallman did not copy.)
Andreas Davour Says:
November 19th, 2007 at 2:46 pm
Thanks for posting this important historical piece!
I once read Levy's book, and while I knew it was a bit controversial and had RMS as it's primary source for some eras I think it gave some important clues to what happened back in the days. I would actually love to have more books describe the computer revolution like that, in order to get a more complete and varied picture of events.
Now, I have read a lot stuff Richard have written, and while I think it's very good to hear the full story I'm very impressed with what he has achieved. It might not have been a complete GNU system, but the availability of free software for learning have been invaluable for people like myself. I wouldn't work in the computer business if it wasn't for the GNU project. If Richard as a political leader takes a few liberties (faulty memory, or whatever the reason) I guess I can understand why. It doesn't hinder me from wanting to know more about what really did happen!
After reading Levy's book I got the impression that the “AI lab” that felt Symbolics was betraying them was probably just RMS and maybe Greenblatt. Levy gives the impression that in Richard's mind Greenblatt was the lab!
I'm amazed by the amount of truly great minds who have chimed in here and shared their view of things. After reading “Hackers” someone like me really wanted to be like Greenblatt, Gosper, Moon, Knight, Deutsch, Greenberg, Weinreb etc etc. :-)
Now I just want to be myself and hack lisp. Thanks guys!
Kamen Tomov Says:
November 19th, 2007 at 7:26 pm
I am a bit surprised that not only Mr. Stallman, but also you (Mr. Weinreb, Mr. Chapman) use the word “hackers” when talking about the members of the AI lab. Why this word, what's the semantics? It seems that the word gained some new context with the development of open source in the last decade and perhaps it is very different from what it used to be.
dlweinreb Says:
November 20th, 2007 at 3:29 am
Mr Tomov; please see http://www.ccil.org/jargon/jargon_23.html#SEC30 for clarification. At MIT, the word “hacker” was very commonly used in the first seven senses of the definition of “hacker” in the Jargon Dictionary. The use of “hacker” to mean someone malicious came out much later, because it sounded to the greater-world as if it would mean something bad. Also see Steven Levy's book, “Hackers”. I'm sure that the first time someone at the AI lab referred to me as a hacker, I knew it was a compliment.
Kamen Tomov Says:
November 20th, 2007 at 10:54 am
Mr Weinreb, thank you very much for your response and for the references!
It is written (on the same web page that you kindly provided) that hackers subscribe to some version of the hacker ethic. That implies that “hacker ethic” is (as written there): “The belief that information-sharing is a powerful positive good, and that it is an ethical duty of hackers to share their expertise by writing free software and facilitating access to information and to computing resources wherever possible.”. I assume that the members of the AI lab were refereed to as hackers long before the term “free software” was coined by RMS. So I wondered why are you referring to ex-members of the AI Lab as hackers provided that a rule not to share know-how with LMI was introduced?! Regarding the (contemporary) definition you become less hackers once you bound to it. I suppose it is for either of two reasons – historical: once you get the qualification “hacker” you never lose it or contextual: you refer to them only in the context of the AI Lab. On the other hand there is a possibility that you use the word in its old (original) sense so I wondered what it was and that's one of the reasons I asked you why you use it.
With the globalization and with the popularization of “free software” and especially with the wide use of “open source” the semantics of the word “hacker” had changed. While in the context of the AI Lab or similar organizations of that time hacker ethic was arguably beneficial in terms of information-sharing, in the context of companies or the global word it is not longer that simple.
You implied that back then it was a compliment to be referred to as a hacker. It seems to me that although it was cool and prestigious, the success of Symbolics had been much more valuable than that. Symbolics was a business organization and the principles of doing business contradicted with the ones within the AI Lab, which was supported (AFAIK) with taxpayer's money.
In the context of the global world information-sharing in the software domain means sharing software. Sharing software globally has quite a different impact. While it could be beneficial in some cases it could be devastating in others (imagine Google sharing its search engine code, or you sharing the ITA's Lisp engine). In that context hacker ethic seems far from beneficial. With “free software” sharing went out of control and as a result software has been devalued. The values had shifted away from what was once important – it is no longer the creativity that matters most. In fact I'm still trying to understand what's important if software were “free” in Stallman's sense.
So what needs to be done? Shouldn't hacker ethic imply that information-sharing is a powerful positive good only in certain circumstances, e.g. if happening within an organisation or group? Or should we think about hackers as people from the past that lived in idealistic world? Or something else?
Dave Says:
November 20th, 2007 at 12:23 pm
@Dan: Thanks for redacting the “lies” stuff :-)
dlweinreb Says:
November 21st, 2007 at 1:42 am
Mr Tomov: As I say, at MIT the word “hacker” was used in the first seven senses of that dictionary. The “hacker ethic”, described as “an ethical duty to … write free software”, is not what it meant. That's just the “Stallman ethic” and perhaps it is the “Eric Raymond ethic”, as far as I am concerned, since it could apply to open source as well as GPL software. Therefore, there's nothing contradictory about my using the word “hacker” as I do.
You say that the semantics of the word “hacker” has changed. The idea that it has changed to mean “writing non-open software is unethical” is Stallman's idea. Of course the way the word has really changed is that it means “a person who illegally gains access to and sometimes tampers with information in a computer system”, which is what most people mean despite the fact that it's definition #4 in the online Merriam-Webster. Language changes all the time; it's inevitable and there's little point in trying to fight the tide. But on this blog I'm using the perhaps-archaic sense in which I learned the word when I was at MIT. By the way, at ITA, software developers are referred to as “hackers”, all of whom are writing proprietary code.
The AI lab did not have a principle that the software written there was open-source or free. It was paid for by tax-payer money, indeed, and was owned by MIT (and/or the Office of Naval Research and/or DARPA) under terms of the funding contracts. All of the MIT Lisp machine code, prior to Symbolics and LMI, was copyright by MIT.
It's true that software has been “devalued” in the sense that it's harder to make money by selling software products when there are competing free products. It hasn't been “devalued” in the sense that people get less value from it. So it depends in what sense you mean “devalued”.
Also, the vast majority of software in the world does not have any open/free version. It's easy to look at software tools, software infrastructure, operating systems, computer languages, and other things that are targeted at programmers and see all the open/free versions. But there's plenty of software that's much more vertical, aimed at specific markets, in which there is less or no open/free software, and there is (arguably) less likely to be any. I have a Mac on which I run Apple's Final Cut Pro and Mark of the Unicorn's Digital Performer, each of which ran me hundreds of dollars. They're both truly fine tools, and it would take an awful lot of effort to product something equivalent. You can still sell things like that, at least for the foreseeable future, for various reasons.
As for “it is no longer the creativity that matters most”, I'm not sure what you mean. Creativity of the software or of the person? Matters most in the sense of selling, or market uptake, or aesthetic value?
I have absolutely nothing against open source software and software that one need not pay for! And I think it's wonderful that the XO (the One Laptop Per Child computer, see my other post), which is specifically intended as an educational vehicle, lets the user see and modify all the sources. I hope kids will learn a lot from that; perhaps only a minority of the kids, but that's OK, we don't all have to grow up to be hackers!
Kamen Tomov Says:
November 21st, 2007 at 8:22 am
Mr. Weinreb, I apologize for not being clear enough. To me the people from MIT AI Lab and later in Symbolics, LMI and some other lispers as well are the people that can define words like hacker. That is the reason why I asked you to provide your definition and I thank you for doing so. As to the fact that you call that trying to fight with the tide I do not agree. Obviously there are many other people that would listen to what you have to say.
It was interesting to me to learn about the licensing of the software written in the AI lab and to find out that there is no contradiction in terms of the word “hacker” as I mistakenly thought.
Software has been devalued in the sense that it is harder to make money from it as for anything good one implements a free software / open source product is very likely to appear, unless of course that's a huge effort most likely by a powerful corp. As a friend said – you can not compete with a price of zero. In result creativity (innovations in software) does not matter so much these days. I meant the creativity of the person writing software.
OLPC is an amazing project – too bad they chose Python and not Lisp.
dlweinreb Says:
November 21st, 2007 at 11:05 am
Mr Tomov: Actually, you can compete with zero cost. Look at Jetbrains's IntelliJ. And three of the current implementations of Common Lisp are commercial despite the strengths of the seven that are not (more to come in a later post about this).
Meanwhile, I think creativity is just as important as even, even if the competitors are open/free software. Sturgeon's Law applies (http://en.wikipedia.org/wiki/Sturgeon's_law) but the other 10% ranges from interesting to brilliant. We're living in a time of an explosion of creativity in software, in so many areas. Another sign of vibrancy is that after a long period, developers seem to be open to using new computer languages to build real products; the pendulum swings back.
Thanks for all your comments!
Adrienne Says:
November 22nd, 2007 at 2:15 am
Dan Weinreb:
Points raised in your rebuttal are apparently not new, since they appear to have been sufficiently addressed by Stallman in various articles, speeches and so on. Your assertions in regard to Stallman's role in the development of EMACS, and his supposed copying of Symbolics code, are misleading, doing nothing to “[set] the record straight.
1. EMACS ========
First, your statement concerning EMACS aims to diminish Stallman's role to that of mere maintenance, Guy Steele and David Moon having – according to you – first created and designed the first TECO-based Emacs, then got it working and established as the standard text editor at the AI Lab. Guy Steele disagrees with you.
The question of Steele's role in the creation and development of EMACS appears to be an old issue that continues to rear its head, although Stallman addressed this as far back as 1987 in his article “Emacs the Full Screen Editor” [8] and just this year in the comp.lang.lisp thread “teco, rms, gosling, mocklisp” on 28 July: [2, viper-2]
Guy Steel played a role in starting the development of Emacs in 1975. He developed the key bindings, I designed the internal platform, and we worked together for the first night of implementation. After that he dropped out.
Guy Steele concurred: [3, Pitman 8 August]
Except for the minor misspelling of my name :-), I concur with everything in RMS' response. Feel free to post this reply to the discussion thread.
… I can confirm that the EMACS command set was based on (I think) four different command sets that sprang up once programmable control-R mode was available. My role was principally to forge a consensus on a standard command set. I also started an implementation of the EMACS command set, but RMS soon took it over (which was fine with me, because I was trying to get into full-time graduate student mode).
TMACS, which was the project of Dave Moon, Charles Frankston, Earl A. Killian, and Eugene G. Ciccarelli, was not a TECO-based Emacs but one of the 5 precursor command sets on which the EMACS command set was based. Referring to the latter, Steele clarified further:(ibid)
Those precursor command sets were: TECMAC, done by JLK (John Kulp) MODE, done by RVB (Robert Baron) MACROS, done by RBR DOC, done by ???; this was actually rather VI-like, with an INSERT MODE and an ALTER MODE
There was also some infrastructure called TMACS being worked on by CBF, EAK, ECC, and MOON that had to do with representing macros in “Teco pure string format” All this was going on during 1975-76.
The posts cited in the threads “teco, rms, gosling, mocklisp” [2] [3] which continued as “TECO, RMS, Gosling, Mocklisp” on 10 August 2007,[4] addressed former Symbolics employee Kent Pitman's misattribution, to Steele, of Stallman's seminal idea to bind keystrokes to TECO macros.[2, Pitman 20 May] Pitman conceded that he was factually incorrect after receiving responses from both Stallman and Steele.[4, Pitman 12 August]
The precursor command sets of TMACS, TECMAC and others were the progeny of Stallman's idea to transform TECO to a user-programmable WYSIWIG editor following his powerful re-implementation of TECO's control R mode originally authored by Carl Mikkelson. According to Steele “That was the real breakthrough”.[10, at note 6] By giving users the power to redefine their own editors command sets, such as TMACS and TECMAC, “[e]verybody and his brother [began] writing his own collection of redefined screen-editor commands.[8]
Beginning with that first night's joint work with Steele (after which Steele dropped out),[2, viper-2, 28 July] [8] Stallman's development of the precursor command sets, along with other features, realised the original TECO – based Emacs. Mark Crispin (IMAP inventor) who in the summer of 1976 introduced Stallman to XTEC (from which Stallman incorporated features into Emacs) described the emergence of the first EMACS:[5]
Two major sets of TECO macro packages … developed; TECMAC and TMACS. TECMAC was a more real-time editor, while TMACS had a much richer set of functionality including named commands. Just about everybody had their own customizations on top of these packages.
This was the situation when I worked at MIT in the summer of '76. I had brought with me my own favorite TECO-style editor, which, although it had only the functionality of primitive DEC TECO, had two interesting facilities: (1) it compiled all TECO programs (including commands) prior to execution, and (2) it had multi- character register names, which greatly increased the number of possible TECO registers to virtually infinite.
Richard Stallman implemented the latter in TECO as part of the EMACS project, which was originally intended as a replacement for both TECMAC and TMACS. By New Years in 1977 EMACS had made significant inroads against TECMAC/TMACS; and in another year or so the older editors had both succumbed to software rot.
This was the editor that became established as the standard text editor at the AI lab. As Greenberg noted, “The impact of Emacs and Emacs-style editing far outweighs that of any TECO”.[6]
2. DUPLICATION OF SYMBOLICS CODE ================================
Of course you “caught him” looking at Symbolics code. Stallman concurs; footnote (3) of the transcript of his 28 October 2002 speech to the International Lisp Conference refers to his statement being “… misconstrued as saying that [he] never, ever looked at Symbolics code. Actually it says [he] did”. [9, note 3]. As Stallman stated, he was entitled to read the code at MIT, but found that reading Symbolics' code made it more difficult for him to write original code to implement equivalent features.
Every programmer knows that it is much easier to code an independent solution to a problem if you don't first look at someone else's code, and that trying to understand someone else's code can be arduous and time-consuming compared with the effort to simply write your own. As Stallman said, “After a while, I concluded it was better not to even look”.[9, note 3] Sometimes programmers working independently of each other will produce code that is almost identical (if not identical), [1] but more often than not the solutions will be different. It would have been exceptionally difficult for anyone – including Stallman – to achieve the duplication of code for two years by first reading the Symbolics code (written by over a dozen world-class hackers) before writing its equivalent. The making of false accusations against the authorship of one's own original code is a particularly sore point with me. Proof of Stallman's authorship is found in the originality of the duplicated code, which speaks for itself.
3. SUPERHACKER AT THE HELM ===========================
Tom Lord:
Others might view your contributions to FSF/GNU as constituting an impressive part of your resume. It is unfortunate that you claim to have had a false belief in Stallman's abilities as a hacker. In light of the cooperation needed to build an organization like GNU such an attitude as an employee of FSF/GNU would have contributed to, as you put it, “the GNU project piss[ing] away one effort after another”.
The leadership prowess of a hero who can slay 10 Symbolics hackers with a single all-nighter is relevant to leadership in technical matters as opposed to leadership in matters of management. The hacker mindset and that required for management rarely intersect, and the hacker/engineering and management functions may even appear to be mutually exclusive in all but the smallest organizations. RMS undertook an unprecedented, mammoth task in founding GNU, and as a practical matter, with a superhacker at the helm of its technical mission the organization would necessarily have required structured management as well as additional resources to do more.
As you know, the complete GNU/Hurd operating system is a reality actively under development. The Hurd kernel is an ambitious, novel project aiming to achieve powerful functionality well beyond the capabilities of known kernels. The GNU operating system in combination with the kernel Linux (a kernel built more rapidly by re-implementation of conventional techniques) enjoys world-wide acclaim by individuals as well as organizations, various governments being in the process of considering similar provisions to Peru's and Venezuela's move to adopt legislation for free software in public administration.
Richard Stallman's superior abilities are not in question. His impressive repertoire of credits include a growing list of honorary awards that now number six honorary doctorates and two honorary professorships – global recognition for launching the free software movement.
Dan Weinreb:
As is clear from this blog, Stallman has his detractors. I observed on 19 November, that someone rushed to edit the Wikipedia entry for Emacs, inserting a footnote in the first paragraph of the section “Other emacsen” pointing to your now deleted statement [above, dlweinreb 19 November] in the post that included the word “lie” in referring to Stallman's error for which he has now apologized [9, note (5)] concerning your editor EINE, the first Emacs written in Lisp.
The Symbolics hackers, collectively, comprise a reservoir of possibly unmatched talent. The energy used here in vilifying Stallman is misdirected. Perhaps you should ask yourselves just who is attacking, defaming, and slandering, whom?[above, dlweinreb, 16 November]
Sincerely Adrienne
cc Dr. Guy Steele
1. See for example Pitman's and Bourguignon's code in the comp.lang.lisp thread “Catching multiple values” 16 May 2007, at http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/e04b8d3b628af6b6/5a09cfdb13f66692?hl=en&lnk=st&q=#5a09cfdb13f66692
2. Comp.lang.lisp teco, rms, gosling, mocklisp, at http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/5b19382493ae549b/10f05b24f20cdb58?lnk=raot&hl=en#10f05b24f20cdb58
3. Comp.emacs teco, rms, gosling, mocklisp, at http://groups.google.com/group/comp.emacs/browse_thread/thread/5b19382493ae549b/735030587c8ebf80?hl=en&
4. Comp.lang.lisp, TECO, RMS, Gosling, Mocklisp, at http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/ddf6408f137f9ec4/5ce5d9a01969004e?hl=en#5ce5d9a01969004e.
5. Crispin, Mark, email to alt.folklore.computers 19 January 1990, at http://www.djmnet.org/lore/emacs-origin.txt; See also http://www.ibiblio.org/pub/academic/computer-science/history/pdp-11/teco/doc/tecolore.txt .
6. Greenberg, Bernard S. Multics Emacs: The History, Design and Implementation, section II, 15 August 1979, at http://www.multicians.org/mepap.html#secii.
7. Stallman, Richard EMACS: The Extensible, Customizable Display Editor (1981), at http://www.gnu.org/software/emacs/emacs-paper.html.
8. Stallman, Richard, “Emacs the Full Screen Editor” (1987), at http://www.lysator.liu.se/history/garb/txt/87-1-emacs.txt.
9. Stallman, Richard, My Lisp Experiences and the Development of GNU Emacs, Transcript of Speech at the International Lisp Conference, 28 Oct 2002, at http://www.gnu.org/gnu/rms-lisp.html.
10. Williams, Sam, Free as in Freedom: Richard Stallman's Crusade for Free Software (Sebastopol, CA: O'Reilly, 2002), 6, at http://www.oreilly.com/openbook/freedom/ch06.html.
Humpty-Dumpty Moon Says:
November 22nd, 2007 at 2:44 am
A “hacker” is simply a person who has a habit of producing “hacks.” A “hack” is a brilliant, surprising, and original piece of (lower-case-a) art, usually emphasizing cleverness more than technical skill and experience although hacking without any skill is impossible.
A “computer hacker” is a “hacker” whose “hacks” usually involve computers; software or hardware or both. Phone (switch) hackers preceded computer hackers by a decade or so.
All those other definitions you read about are just neologisms.
Clear now?
PS: Employing the Big Bang to create an entire universe with zero net energy, yet endlessly complex content, was certainly The Greatest Hack of All Time.
dlweinreb Says:
November 22nd, 2007 at 3:50 pm
Adrienne:
(1) No, it doesn't. By “maintenance” I simply meant development”, and I suppose I should have used that word to be clearer (sorry). And there is no question that Steele and Moon started Emacs. I was there and watched it happen; both Steele and Moon confirm what I said. It is not true that Stallman was involved from the first day. While I don't remember precisely how many days after the beginning of the implementation that he got involved, I'm sure it was at least a few weeks. I was one of the two original beta-testers and I know who was fixing the bugs that I found. Stallman's claim to sole credit, in [7], is not accurate. Steele and Moon don't really care about getting credit. Since Stallman did the vast majority of the total work on Emacs, which is quite true, they don't mind his taking sole credit. But it has always bothered me. As to who came up with the idea of binding keys to TECO in the first place (control-R mode), I didn't say anything about that. That happened way before Emacs. I only commented on Emacs proper.
(2) I am correcting the impression created by Steven Levy's book, “Hackers”. Many more people have read it than have seen Stallman's footnote. You yourself describe Stallman as “a hero who can slay 10 Symbolics hackers with a single all-nighter”; as I have explained, that is simply not true.
(3) I'm not slandering anyone, because everything I said is actually true.
Andreas Davour Says:
November 22nd, 2007 at 7:33 pm
Adrienne,
I think things that happened that far back is very likely to be remembered a bit differently. Dan has told his story and let us be thankful for that. Richard can probably, if he feel his honor is at stake, say his piece by himself. I doubt he feels that strong about it. His legacy is alive as the meme that software can be something more than a black box. Emacs is just a tool.
Who cares who did what anyway? Nobody doubts the abilities of Stallman, Steele or Moon and we have one hell of an editor as the fruit of their labour. Good work, guys – all of you!
I still think it would be fun to have more books like “Hacker” written.
Thomas Lord Says:
November 23rd, 2007 at 7:26 am
Adrienne,
Richard can speak for himself. I think most people who are reporting their direct and personal experience here did not come here to debate. No public space is safe, it would seem, from The Myth.
I think you partly prove my point, really. Your response to me, in this context, borders on the inhumane — you will recognize my perspective only in so far as it is an object for you to refute on the basis of the “authority” of Stallman's reputation. You aren't honestly inquistive into just what I might be talking about: just eager to be in the League of Friends of the Hero and foreclose all debate. You exemplify the problem, here.
-t
Adrienne Says:
November 24th, 2007 at 4:52 pm
Tom Lord
1. As I said before, RMS has already spoken for himself on these issues. I am merely directing your attention to the relevant references. I can see from http://www.fsf.org/blogs/rms that he doesn't have the time to update his own travel blog; he seems to be tied up spreading the GNU revolution all over the world.
2. If my intent was to foreclose debate, I don't seem to be doing a good job since from your response I have provoked further discussion. Describing my response as “inhumane” is unfair. You shouldn't take my comments personally. I'm simply saying that Stallman needed high levels of cooperation to mitigate GNU's inadequate resources. Under-resourced projects fail everyday. Resources include effective management personnel. It is just not possible for any one person to hack while managing projects, people and resources at the same time. The failure of some projects should have been predictable.
Adrienne Says:
November 25th, 2007 at 4:13 pm
Dan Weinreb:
I contacted Stallman drawing his attention to the comments posted in this blog. With regard to your assertions that Guy L. Steele Jr. and David Moon were the authors of the original TECO-based Emacs, Stallman has requested that I inform you that your claim is false.
In an email message to me dated Thursday, 22 November 2007, Stallman stated:
Weinreb is wrong. Moon was never involved in developing Emacs.
In a further message dated Saturday, 24 November 2007, Stallman reiterated:
Please post that I told you that this claim is false. Steele and I worked together for the first night of writing code for Emacs, and then Steele dropped out. Moon was not involved.
Sincerely Adrienne
cc * Richard M. Stallman * Guy L. Steele, Jr.
dlweinreb Says:
November 25th, 2007 at 5:35 pm
Adrienne: As you can see from his comment above, Dave Moon remembers it the same way I do. I asked Steele about it at the OOPSLA, and he also remember it the way I do. Stallman's memory is inaccurate regarding this.
Douglas Knight Says:
November 25th, 2007 at 10:22 pm
Thomas Lord, Could you elaborate on your view? You go back and forth between two views that seem contradictory to me. On the one hand, you complain that you were tricked by legends about Stallman's hacking abilities, and on the other hand, you seem to say the problem was other skill sets.
Maybe you would have made better career decisions if you had had more accurate beliefs about hacking abilities, but that seems to me to have been a coincidence. Your error was worrying too much about hacking quality and not about other issues. You say anyone with a little experience would known better; your problem was not having experience.
Adrienne Says:
November 26th, 2007 at 1:22 pm
Dan Weinreb:
THE EVIDENCE
1. The evidence in the written record contradicts your claim that Guy Steele and David Moon authored the first TECO-based Emacs.
2. Guy Steele Jr. did not rely on his memory but verified that he consulted his notes before concurring with Stallman. In his email of 7 August 2007 to viper-2, Steele said:[4, viper-2 10 August 2007]
Thanks for the message. In fact, Mr. Pitman contacted me as well and we have had an exchange of email concerning this bit of history, and I consulted some of my old notes on EMACS.
Except for the trivial fact that RMS seems to have accidentally misspelled my name, I agree with everything in his response as quoted below.
Yours, Guy Steele
3. Nowhere in the record does Steele give Moon credit for working on Emacs. Steele, referring to the project Emacs, said “[a]s I like to say, I did the first 0.001 percent of the implementation, and Stallman did the rest”.[10]. As noted above, Steele stated that David Moon worked on the pre-Emacs command set TMACS.[3, Pitman 8 August]
4. By authorising the publication of his email,[3] [4] Steele made his account of his role in Emacs a matter of public record. Steele's written testimony, and not unsubstantiated second-hand accounts, is accordingly the authority on which the public will draw conclusions. Unless and until Steele retracts these published statements – giving unimpeachable reasons for reversing his position – these are the only statements from him that are admissible.
5. Greenberg concurs with Steele and Stallman:[6]
By 1976, several packages of TECO macros [TMACS, TECMAC] had proliferated.
… At this time Richard Stallman coalesced most of the ideas in these packages, and created a unified approach to command set, extensibility, documentation, and integration of these facilities, and created a large, unified, set of macros which came to be known as Emacs. The name is derived from “Editor Macros.
6. In the book “Lisp” 3rd edition by Berthold K. P. Horn and Patrick H. Winston, David Moon is listed in the Acknowledgments as one of 13 people who scrutinised “large parts of an early draft with incredible care”. [Horn, B. K. and Winston, P. H. (1989), LISP 3rd Edition, Addison Wesley] The glossary included an entry for EMACS: [Id, p. 578]
EMACS – A powerful, popular editor, conceived and implemented initially by Richard M. Stallman. Various versions of EMACS were used in preparing the programs and text for this book.
Horn and Winston, who were professors at MIT during the development of Emacs, give Stallman sole credit for the conception and initial implementation of Emacs. Moon's failure to protest this attribution – in a book acknowledging his contribution – infers agreement.
We need not rely on Stallman's – or anyone else's – memory.
Sincerely Adrienne
cc * Guy L. Steele Jr. * Richard M. Stallman
dlweinreb Says:
November 27th, 2007 at 3:28 am
Adrienne: Stallman has repeated his story many times and everybody believes him. Quoting more and more people who have repeated his story is pointless. The fact that Moon didn't bother to correct Horn and Winston's book is completely predictable, since Moon has not bothered to take credit for what he did. Except here (see his comment above).
Horn and Winston were not around when any of this was happening. They were down on the eighth floor; this all happened on the ninth floor. Very few people were involved.
I should point out that at the time Moon and Steele started working on the new editor, they didn't have a name for it. It went under the name “?” during the period of initial development. When Stallman took over, he made up the name “Emacs”. The story behind the name, which I'm pretty sure Stallman has written up, is that our friends at the Stanford AI Lab used an editor named “E”. When you did a “finger” command to see who was logged in at Stanford, it would also show you what program each person was running, and many of them would always be running “E”. At MIT on ITS, it was conventional that the most widely-used applications were given a one-letter nickname, and so Emacs would also be known as “E”, and would show up that way when people at Stanford did a “finger” on us. Stallman's joke was that they'd see us running “E” and be surprised. Anyway, it could be that Stallman is thinking only of the period after he changed the name to Emacs. This may be why Stallman misremembers what happened.
I have removed your most recent post. I won't tolerate insulting name-calling. If you want to do that, start your own blog.
I have explained what's going on, above. No more on this topic.
Adrienne Says:
November 27th, 2007 at 1:07 pm
“Name calling” Dan?
As your readers can see, it was Moon who first referred to himself as “Humpty Dumpty”.
I bet you had a laugh anyway.
Dan Weinreb, I hope that your current project will be very successful. I also hope to hear that you and your extraordinarily talented former Symbolics colleagues have regrouped and are achieving spectacular things, like maybe helping to take us to Mars!!
Sincerely Adrienne
dlweinreb Says:
November 28th, 2007 at 2:35 am
Adrienne: Oh, I see what the confusion was here. We were talking past each other; sorry. No, David Moon did not write the comment labeled “Humpty Dumpty Moon”. I don't know who that was, but it was definitely not David Moon. Please look at the second comment, way at the top, whose author is listed as “Moon”; that's David Moon.
Emacs History, Who Created Emacs
Thomas Lord Says:
December 17th, 2007 at 9:08 am
Somebody up above in the threads, after my last bitter comment about RMS, gave me an excuse to elaborate. I won't bother quoting the thread: I'm posting this late — after the heat of the thread has died down.
As project leader of the GNU project, in my opinion, RMS has behaved idiotically, obnoxiously, self-servingly, and to ill-effect. The GNU project is an embarrassing rubble, in spite of many good efforts by volunteers. I hate RMS for this.
I also love RMS, for two simple reasons:
First, when you deal with him one-on-one, it's OK. I find myself just reflexively respectful and not because of what he's done (much of which bugs me) but, because, he's reasonably sane at a human level (reputation notwithstanding). He's a mensch.
Second, GNU project per se being the train wreck it is — that aside — there's a simple fact:
Around the world — really, stop and think: the whole world — there's a stunning number of programmers who feel some sense of entitlement that, when they work, they can just grab the source of stuff they use, and study it, and extend it, adapt it, repurpose it, etc. And they have no problem with people doing that with what they themselves produce. There are (by official count) a gazillion little dweebs like me, all around the world, who just presume “code is just code, after all” and who expect and/or increasingly insist on open systems and personal freedom.
The existence of a lot of those people in such a beautiful state (at least re software freedom) is probably appropriately blamed on RMS more than anyone else. The GNU project per se is a disaster — but one can kind of see what he threw it overboard for.
Thomas Lord Says:
December 18th, 2007 at 3:05 am
One more thing.
It's a crap shoot that it'll stick — probably it won't. But if we stop after this one, two things follow:
1. There will be 52 comments: a deck of cards. 2. The bottom line, typographically at least, is a love letter.
F.U. RMS. And you deserve “every inch of [that] love.” [zep]
Love, and thank you, -t
Fromer BBDOer Says:
April 14th, 2008 at 7:10 pm
Anyone know whatever happened to the young genius, Howard Cannon?
Harold Ancell Says:
August 19th, 2009 at 5:14 am
Here's Howard's Linked In account.