What's new? This is just a rewritten version of one side of the Emacs history. Some parts of that are remembered 'slightly' different by other involved persons.
Example: 'He took his revenge by going on a two year coding binge to reproduce every advance that was made by the team of hackers at Symbolics, and match it feature for feature on behalf of Greenblatt and LMI.'
'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.'
But worse, again Weinreb:
'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.'
I was sort of there (in the larger social circle, worked at LMI, was a roommate of RMS after that, etc. etc.).
If you'd spend a little time with both in that period you'd have absolutely no doubt who's more credible.
As for the specific claims:
I can't speak to how he did it and if he looked at Symbolics source code, but I do know that RMS only created a fraction of the software artifacts that the Symbolics team did on top of the existing code base (RMS is/was back then good, but only so good ... and very possibly below the average for the core AI Lab Lisp Machine team (note that average was very high indeed)).
His effort concentrated on covering the very most important new core stuff ... the last thing I remember talking with him that he was doing before I exited the Lispm community was something like an exception system, but it's been way too long to remember any details.
As for who "invented" TECO EMACS, everyone I talked to at the time said it was primarily Guy Steele, who then turned it over to others after a very short period. At the time, the effort was mostly to get the community to agree on a common set of key bindings, macros, etc. There were something like three popular sets which meant it could be very difficult to sit down at someone else's console to help them. I can well believe David Moon lent a significant hand there, he was one of the very top wizards (e.g. KL-10 (for ITS) and Lispm microcode) but one of the most quiet ones.
Interesting. I recall working a bit on Statice, which I believe Weinreb and Lamb? wrote. There was a phenomenal amount of code coming out of that lab, so it's not surprising that RMS was considered average within that group.
I was as TJ Watson then, in math sciences. I recall IBM lawyers not letting RMS come speak at the lab, when one of the guys had invited him to talk about emacs. At the time IBM was quite shy of this "free software" business.
I think the account in this article largely comes from the Sam Williams biography of Stallman.
I actually own stock in Symbolics, if the company still exists. Last time I saw a statement it was worth $0.02 I was never able to get certificates, which I was hoping to have as a momento, to go with my symbolics mousepad.
Thanks for sharing. This sheds a little more light on this piece of history.
The people I talked to, from Symbolics, did not tell me the Stallman version. So at least I have a lot of doubt about the story that the article linked tells.
Especially since Stallman and the FSF surrounds a lot of propaganda. Essentially Stallman was working for a government lab paid for mostly by the military (DARPA). DARPA wanted to see their investments applied to military software and created the commercial structure to do that. Stallman was living in his dream world thinking that all that money was coming for free and for the best of mankind and he could play there forever, working on editors and other stuff - when in fact the military wanted all kinds of advanced software for their projects, which then lead to 'Strategic Computing Project (SCP)' and 'SDI'. DARPA paid the whole show and had the goals - intelligent software for weapons, battle management, pilot training, logistics, etc.
The above is interesting, but I never felt it was a driver of any significance for RMS, especially since the AI Lab was so very far removed from the sharp end of the spear.
A much better explanation someone told me was that one day RMS realized he no longer had anyone who'd go out to lunch or dinner with him (since they'd almost all gone to work at Symbolics and [you can fill in the rest]).
Now, there were people working on that sort of thing, e.g. I did some customer support for an LMI CADR that was sold to MITRE in Virginia (as I recall) where they were trying to do what you might call "battlespace SIGINT with memory", i.e. help get around the switching your transmitter on and off and hopping frequencies fun. Very interesting problem, but it was very early level research, I think more a proof of concept than anything else. I.e. unclassified research, at least at the level I've described it (I certainly didn't have a clearance).
And then there was the company that had Lispm's driving bulldozers (for maximum efficiency), but I was only told that 2nd or 3rd hand.
I don't know if Stallman was aware that he was working as a programmer in an early stage (basic research, tool building, ...) of a multi-year military project.
All the stuff was basically paid for by DARPA (and related). Symbolics sold much of their initial production to SDI projects. It was thought that the SDI systems were extremely complex and that many Expert Systems were needed to design, deploy and maintain the space-based weapons and their infrastructure. The SCP projects were more down to earth: battle management, logistics, pilot training, etc. SCP was spending one billion dollar in ten years on AI - add to that the money that was spend on the AI-related parts of SDI.
Please ... we all knew the money came from DARPA and that they weren't interested in improving the state of the art in plush toy design.
SDI officially started in March 1983 (although BMD had been a hot topic for decades) ... and that was about the time RMS started transitioning out of Lispm work. We were roommates in September 1983 when he officially started the GNU Project.
We certainly had debated BMD/SDI (before I started working for LMI I was the editor of the ARM-D mailing list), and I can't remember him ever saying GNU had anything to do with that sort of thing. It had to do with end user freedom, as he has explained at length, with a singular purpose of mind and consistency.
If this was a major motivator, shouldn't he have put in a "you can't use this software for evil" clauase in the GPL? His GNU EMACS and GCC work has certainly enabled much more development in weapons and the like than anything he did at the lab, and I assure you he's intelligent enough to have realized that, and at least in the '80s he had enough intellectual honesty to realize that ... we all had to ask ourselves that question (and I wouldn't let anyone weasel out of it if they tried :-).
Hmmm ... maybe he claimed this to others, but he knew it wouldn't fly with me. But surely only as a minor aspect of why he was going from proprietary Lispm work to free software.
Example: 'He took his revenge by going on a two year coding binge to reproduce every advance that was made by the team of hackers at Symbolics, and match it feature for feature on behalf of Greenblatt and LMI.'
That's a myth.
http://danweinreb.org/blog/rebuttal-to-stallmans-story-about...
Weinreb::
'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.'
But worse, again Weinreb:
'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.'