From fare@tunes.org Mon, 15 Jan 2001 00:06:42 +0100 Date: Mon, 15 Jan 2001 00:06:42 +0100 From: Francois-Rene Rideau fare@tunes.org Subject: historical clarification Dear Richard, sorry to bother you, but while discussing about LISP Machines with friends (my project is to make an OS based on a LISPy high-level language, and I've just bought one), I have stumbled on a minor historical question about your LISP activity just before the GNU project started. As documented in Steven Levy's "Hackers" (darn, only the first part of it is available online, without the appendix where it is discussed, and I haven't got a copy home, so I can't verify what exactly is said there), while you were still at MIT, you spent your time reverse-engineering Symbolics' proprietary code from its published documented interface, and porting it in a way such that it could notably be used by competitors from LMI. Can you tell me the full story about that code? What dialect of LISP and/or assembler was your code written in? What systems could it run on? Did you write it on MACLISP? On CADR machines? What was the compatibility between the code and the various existing LISP systems? What was the mutual compatibility between these systems? What was the license for that code and who owned it? As I understand it, the MIT owned that code, and let LMI use it; did LMI eventually buy it and turn it into proprietary software? Does this code still exist for peruse, somewhere? When exactly did these events happen, and for how long? How did they impact the eventual birth of GNU, both politically (free software vs proprietary software) and technically (C vs LISP)? Sorry to raise these old stories once more. Maybe we'll meet in Paris in a few weeks. May the LAMBDA be will you. PS: did the FSF make an official stance against the new tax on blank media that was recently voted in France? [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] [ TUNES project for a Free Reflective Computing System | http://tunes.org ] Those who will not reason, are bigots, those who cannot, are fools, and those who dare not, are slaves. -- George Gordon Noel Byron (1788-1824), [Lord Byron] From rms@gnu.org Mon, 15 Jan 2001 00:38:11 -0700 (MST) Date: Mon, 15 Jan 2001 00:38:11 -0700 (MST) From: Richard Stallman rms@gnu.org Subject: historical clarification As documented in Steven Levy's "Hackers" (darn, only the first part of it is available online, without the appendix where it is discussed, and I haven't got a copy home, so I can't verify what exactly is said there), while you were still at MIT, you spent your time reverse-engineering Symbolics' proprietary code from its published documented interface, No reverse engineering was involved. I wrote simply code to implement these specifications. Reading a spec is NOT reverse engineering. and porting it in a way such that it could notably be used by competitors from LMI. No porting was involved. The code I wrote was part of the MIT Lisp Machine system. What dialect of LISP and/or assembler was your code written in? Lisp Machine Lisp. What systems could it run on? Only on Lisp machines. Did you write it on MACLISP? On CADR machines? The CADR is the name of the model of Lisp machine that was in use at the time. All the Lisp Machines which existed at the time were CADRs. As I understand it, the MIT owned that code, and let LMI use it; did LMI eventually buy it and turn it into proprietary software? I don't think so, but I am not sure what happened later. Does this code still exist for peruse, somewhere? I don't know. When exactly did these events happen, and for how long? Symbolics issued its ultimatum on March 16, 1982, which is known as Microwave Day because my first reaction was to disconnect their microwave link. (By coincidence, it was my birthday.) The war began when Greenblatt and I decided to resist the occupation of the AI Lab by Symbolics. How did they impact the eventual birth of GNU, both politically (free software vs proprietary software) Symbolics's destruction of my community created the situation in which GNU was necessary. My success in fighting Symbolics gave me the technical confidence that I could eventually develop GNU, and the determination necessary to launch the project. and technically (C vs LISP)? There was very little influence. GNU is Unix-compatible. PS: did the FSF make an official stance against the new tax on blank media that was recently voted in France? This is the first I hard of it. It's too bad nobody talked with me BEFORE the decision. From fare@tunes.org Tue, 16 Jan 2001 01:11:11 +0100 Date: Tue, 16 Jan 2001 01:11:11 +0100 From: Francois-Rene Rideau fare@tunes.org Subject: historical clarification Dear Richard, thanks a lot for your reply. > No reverse engineering was involved. I wrote simply code to implement > these specifications. Reading a spec is NOT reverse engineering. Oops. My mistake. > [Code written in] Lisp Machine Lisp. [Ran] Only on Lisp machines. > The CADR is the name of the model of Lisp machine that was in use at > the time. All the Lisp Machines which existed at the time were CADRs. Thanks for these details. Do you know about documentation for these systems? I'll dig into MIT AI lab archives... > As I understand it, the MIT owned that code, and let LMI use it; > did LMI eventually buy it and turn it into proprietary software? > I don't think so, but I am not sure what happened later. OK. > Does this code still exist for peruse, somewhere? > I don't know. Last time I contacted people at MIT, I had no answer, but maybe I wasn't precise enough. > When exactly did these events happen, and for how long? > Symbolics issued its ultimatum on March 16, 1982, which is known as > Microwave Day because my first reaction was to disconnect their > microwave link. (By coincidence, it was my birthday.) The war began > when Greenblatt and I decided to resist the occupation of the AI Lab > by Symbolics. Ahem. Are these events documented somewhere? What was the contents of the ultimatum? What was the occupation? Did Greenblatt actually decide to found LMI afterwards, or were those things already in preparation? At the time, what were the legal stances of Symbolics, the MIT, LMI, you, and other Lisp hackers, about the distribution of Lisp Machine code? [BTW, I wish you a happy un-birthday.] > How did they impact the eventual birth of GNU, both politically > (free software vs proprietary software) > > Symbolics's destruction of my community created the situation in which > GNU was necessary. My success in fighting Symbolics gave me the > technical confidence that I could eventually develop GNU, and the > determination necessary to launch the project. If you feel it was a success, what factors made you switch to such a completely different system as UNIX? Do you feel my .signature (below) is correct? Today, what do you (technically) miss most about those LISP systems? What do you (technically) miss less about these systems? What recommendations would you make to free software LISP hackers? > and technically (C vs LISP)? > There was very little influence. GNU is Unix-compatible. There seems to have been some vague influence, with elisp being part of GNU Emacs, with GCC internals being remindful of a LISPy constraint solver, with the GNU recommendations of "no arbitrary limits". I supposed there could also have been some reverse-influence, with your avoiding in GNU things that you learnt from LISP experience were bad? In any case, I realize that the choice of UNIX and C was very political, and little technical, UNIX being somewhat technically inferior in many ways, but also the mostest community-maker thanks to its having the least proprietary license. I'm not sure how much you realized at the time (or if you agree today) about the natural link between those two facts (UNIX having been for some time the "least proprietary" and its having gathered a somewhat large user-base). Maybe somehow GNU kept the free UNIX tradition alive where it would eventually have been killed, too, by the movement of its licenses toward a more and more proprietary model. > PS: did the FSF make an official stance against the new tax on blank > media that was recently voted in France? > This is the first I hard of it. It's too bad nobody talked with me > BEFORE the decision. I don't know if it has actually been voted yet or at all, but a minister has proudly announced it as if it were something that will be done and for which no discussion is necessary. I thought you would have been told by your more permanent FSF correspondants from april.org, and I encourage you to contact them ASAP. Yours freely, [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] [ TUNES project for a Free Reflective Computing System | http://tunes.org ] The LISP community lost the celestial mandate the day Symbolics hired out most MIT AI hackers to make a proprietary system, with utter despise for those not able or willing to purchase expensive licenses only valid on their very expensive specialized hardware. What was an cooperative hacker community fragmented into a bunch of uncooperating vendors; LISP thus lost most of its user base, and began withering as surely as would a tree most of whose roots were severed. As says my mother, LISP hackers were punished for their greed and their arrogance. From rms@gnu.org Wed, 17 Jan 2001 13:18:28 -0700 (MST) Date: Wed, 17 Jan 2001 13:18:28 -0700 (MST) From: Richard Stallman rms@gnu.org Subject: historical clarification Ahem. Are these events documented somewhere? What was the contents of the ultimatum? What was the occupation? See the book Hackers, by Steve Levy. Did Greenblatt actually decide to found LMI afterwards, LMI had existed for around 2 years already. At the time, what were the legal stances of Symbolics, the MIT, LMI, you, and other Lisp hackers, about the distribution of Lisp Machine code? MIT had licensed it to both LMI and Symbolics, as proprietary software. This is one of the reasons why I stopped working on it and started the GNU Project instead. So your signature is not entirely accurate. Aside from that, the idea of "mandate of heaven" supposes there are gods that direct the world based on an idea of justice. This absurd idea leads people to figure that the gods will make sure justice happens, so people can relax. There is no basis for this belief. If you feel it was a success, what factors made you switch to such a completely different system as UNIX? My work was a success for the goal I had in mind: punishing Symbolics. But it did nothing to create a new free software community. For one thing, the Lisp machine software was not free. But even if it had been, it only ran on Lisp machines. What recommendations would you make to free software LISP hackers? At this point, I think it is best to use Scheme rather than other dialects of Lisp. The more free software programs use Scheme (or other Lisp dialects) for extensibility, the more the love of Scheme will spread. There seems to have been some vague influence, with elisp being part of GNU Emacs, with GCC internals being remindful of a LISPy constraint solver, with the GNU recommendations of "no arbitrary limits". Yes, in details like that. I supposed there could also have been some reverse-influence, with your avoiding in GNU things that you learnt from LISP experience were bad? Right. My experience with the Lisp machine convinced me that object-oriented programming is not really as wonderful a thing as it is cracked up to be.