Reported Errors

"IBM, the company once regarded as the most powerful in the computer hardware industry..."

On the one hand, he recalls a man whose piercing insight and moral clarity prompted him to study French a full year before the U.S. entry into World War II, just so he could be more helpful to the Allied cause. On the other hand, he also recalls a harsh parent whose mental gifts only helped to magnify the impact of his emotional outbursts.

At the same time that Emacs was being born, the Lisp Machine project was getting started at the AI Lab. During the spring of 1976, at the end of my freshman year at MIT, Richard Greenblatt invited me to join the AI lab, under the auspices of the Undergraduate Research Opportunities Program. I was delighted and accepted immediately. In the summer of 1976, the Lisp Machine project really got started, and Greenblatt, Moon, and I started writing software in earnest.

The Lisp Machine was an entirely new computer system and needed all of its software written from scratch. Among other things, Greenblatt had started writing the rudiments of a text editor, based on the pre-Emacs experience with TECO macro packages. He assigned me to work on developing this into a complete text editor.

At the same time, Steele and Moon and Stallman were creating Emacs, and I was one of the first two users (other than the implementors). It was quite clear that Emacs was going to be the concensus editor at the AI lab. Since the Lisp Machine was being designed to be used by the AI lab staff, faculty, and students, it was obvious that its text editor should behave exactly as did Emacs on the PDP-10 time-sharing system.

I therefore wrote my text editor to mimic the behavior of Emacs as closely as I could. I named it EINE (Eine Is Not Emacs), copying the self-referential acronym idea, which was originated by Ted Anderson at the Architecture Machine Group (now part of the MIT Media Lab), who had written TINT (Tint Is Not Teco).

First, you say that I "[took] software from the community chest, alter[ed] it, and slap[ped] a new name on the resulting software". This is entirely untrue. EINE was not "altered" from TECO/Emacs. In fact, EINE was written entirely in Lisp, whereas TECO was written in PDP-10 assembly language and Emacs was written in TECO. There was no code in common whatsoever.

Second, you say that this was "a stunning lack of courtesy". Well, if I had done that, I guess it would have been. But I didn't and it wasn't.

Nobody has ever accused my development of EINE to be discourteous. Not even Stallman has ever said or implied anything of the sort/ Certainly at the time, nobody at the AI lab had any problems with EINE, including Stallman. He was perfectly aware of what I was doing and he never made any complaint at all.

He certainly didn't threaten legal harassment!! In fact, there would have been no grounds for any legal harassment at all.

In fact, Stallman particiated in the development of EINE. He even wrote some of it himself.

By the way, ZWEI was also written by me (and Mike McMahon); it was a complete rewrite of EINE. SINE was another editor written from scratch that implemented a subset of the Emacs command set.

Not everybody accepted membership within the commune but many saw the value of Emacs as a convenient starting point for WYSIWYG editor design. Daniel Weinreb, an MIT undergraduate who started hacking at the AI Lab in the summer of 1976, was an early participants in both the macros standardization effort and the simultaneous early development of the LISP machine. Assigned to write a new text editor for the LISP machine, Weinreb decided to mimic the behavior of Emacs. When it came time to name the new editor, Weinreb, both as a tribute to Stallman and as a sly reference to Ted Anderson's recursively named TINT editor [short TINT is not Teco], named his spin-off EINE: EINE is Not Emacs.

"Since the Lisp Machine was being designed to be used by the AI lab staff, faculty, and students, it was obvious that its text editor should behave exactly as did Emacs on the PDP-10 time-sharing system," says Weinreb

Future Emacs, recursive names, spin-offs such as Sine -- short for SINE is not Emacs -- and perhaps the most cleverly named of all: ZWEI, a follow-up to EINE developed by Weinreb and Mike McMahon. This time the recursive initials stood for Zwei was Eine initially.

In a few cases Stallman himself helped the hackers write and debug these spin-off versions. Other times, programmers modified or built their own editors, using Emacs as a design model. How far the commune concept extended beyond the walls of the AI Lab is hard to tell. Even within the lab, the concept had limited impact. Neither Weinreb and Steele remember Stallman mentioning it. At the same time, however, both credit Emacs for pushing the hacker lust for tinkering and collaboration to new heights.

Apple shipped the Apple II Reference manual with every machine. This manual had the complete source listings for the Monitor (which you call a built-in OS). (They also included complete schematics for the machine.)

You may be confusing this Monitor (which was built-in) with DOS. It is true that the source code to DOS was not published by Apple. Apple only published 'entry points' into the code for programmers to use. Third-party publishers *did* produce complete documentation of the system. But DOS was never 'built-in'; you loaded it from disk (the DiskII drive was a separate product). The loader program on the DiskII Interface card *was* documented in source form.

Apple did not provide source code for Applesoft BASIC, which was licensed from Microsoft under terms preventing disclosure. But that's not really part of the Monitor. Source code for Woz's Integer BASIC *was* included.

The time of the Apple II and the Commodore 64 was a time of transition between the hobbyist and home user markets. Hobbyists *insisted* on detailed technical documentation; home users didn't. Even the original IBM PC had ROM listings and schematics available (you had to buy the "Technical Reference Manual"); PC-DOS did not have source code available because it (like Applesoft BASIC) was licensed from Microsoft.

Nowhere was this state of affairs more evident than in the realm of personal computer systems, where, for the first time, computer companies were selling directly to mainstream consumers. Although companies such as Apple Computer, Commodore, and even IBM recognized the importance of courting technical users through the inclusion of software documentation and hardware schematics, the pressure to do so was decidedly less than in both the mainframe and microcomputer marketplaces. For many first time users, the personal computer was simply an expensive appliance with all the inherent cultural taboos against unauthorized tampering and unwarranted exploration. Although a few anarchic adherents of the hacker ethic took delight in "cracking" and distributing copyright-protected software programs and computer games, the marketplace seemed eager to reward programmers speedy enough to write programs for the new platform and savvy enough to attach a copyright notice.

There is an error; the "[sic]" should be there. The phrase "in so far as" *reads* correctly, but it isn't *spelled* correctly. Checking a dictionary (e.g.http://www.yourdictionary.com reveals that the correct spelling of this phrase (which yourdictionary reminds me is a conjunction) is: "insofar as".

So the O'Reilly copy-editor wins. :-)

As an example of this informality, Gilmore cites a mid-1980s copryight notice for trn, a news reader program written by Larry Wall, a hacker who could go onto later fame as the creator of both the Unix "patch" utility and the Perl scripting language. In the hope of striking a balance between common hacker courtesy and a creator's right to dictate the means of commercial publication, Wall used the program's accompanying copyright notice as an editorial sounding board.

Copyleft is a general method for making a program free software and requiring all modified and extended versions of the program to be free software as well. [source: http://www.gnu.org/copyleft/copyleft.html]

Stallman's growing stature as a software programmer, however, was balanced by his struggles as a project manager. Although the GNU Project moved from success to success in creation of developer-oriented tools, its inability to generate a working kernel, the central "traffic cop" program that determines which devices and applications get access to the microprocessor in all Unix systems, was starting to elicit grumbles even before the 1980s came to a close.

As with most GNU Project efforts, Stallman had started kernel development by looking for an existing program to modify. A review of GNU Project "GNUsletters" of the late 1980s reveals that this approach, like the initial attempt to build GCC out of Pastel, was less than optimal. A January, 1987 GNUsletter reports that the GNU Project was working to overhaul TRIX, a Unix kernel developed at MIT. By February of 1988, according to a newsletter published that month, the GNU Project had shifted its design focus to Mach, a lightweight "micro-kernel" developed at Carnegie Mellon. All told, official GNU Project kernel development wouldn't commence until 1990.

RSI is an umbrella term that covers a variety of nerve, ligament, and muscle disorders caused by repetitive motion such as that involved in typing. Stallman probably didn't have carpal tunnel syndrome, but he certainly *did* have some form of RSI. (The description of his symptoms is vague, but cubital tunnel syndrome might fit.)

I think I must have said "timing bugs", but they are more often called "timing errors".

"Timing errors" occur in an asynchronous system where jobs done in parallel can theoretically occur in any order, and one particular order leads to problems.

Imagine that program A does X, and program B does Y, where both X and Y are short routines that examine and update the same data structure. Nearly always the computer will do X before Y, or do Y before X, and then there will be no problem. On rare occasions, by chance, the scheduler will let program A run until it is in the middle of X, and then run B which will do Y. Thus, Y will be done while X is half-done. Since they are updating the same data structure, they will interfere. For instance, perhaps X has already examined the data structure, and it won't notice that there was a change. There will be an unreproducible failure, unreproducible because it depends on chance factors (when the scheduler decides to run which program and how long).

The way to prevent such a failure is to use a lock to make sure X and Y can't run at the same time. Programmers writing asynchronous systems know about the general need for locks, but sometimes they overlook the need for a lock in a specific place or on a specific data structure. Then the program has a timing error.

I contend that your summary of the entire thing is misleading, since the Qt issue was in fact one of the harshest recent controversies RMS got into. A brief recap might go as follows: in 1996 the KDE project is started. KDE shows immediately great promise as the future desktop environment of Linux. Stallman cries foul, because KDE is built on the "non-free" Qt toolkit (then licensed for free to free software developers, but available for a price to closed software developers).

RMS has accepted this compromise on other circumstances, but the graphical environment is too big of a thing to be non-completely-Free by his standards. So he launched the GNOME project (as an alternative desktop), and the Harmony project (aimed at producing a free implementation of Qt -- project now dead.

Trolltech compromised in many ways: they had RMS draft the QPL to see him backing away from it later, they put the Qt sourcecode in the trust of a foundation, so that if their company were bought, the availability of the source for "Free" purposes would be guaranteed, and finally they released it under the GPL (the *NIX version of the toolkit only).

After this. RMS blasted KDE, which was the main target of his attacks, once more, with the famous "beg for forgiveness" editorial on linuxtoday.com. And finally the controversy subsided, getting to where Torvalds had suggested it should have always been: whoever writes the better code, with the best features, should prevail.

RMS has since said that multiple licensing of Qt under GPL/QPL and commercial is acceptable to him, and he confirmed that to me directly recently.

As you can see, it is hard to accept that RMS had done "little" to bring about the license change. He fought very harshly and delivered several blows that about 50% of the observers (the KDE ones, obviously), would define as "below the belt", such as the beg-for-forgiveness nonsense, where he forgave a licensing violation of FSF code that never occurred, given that KDE has NO FSF CODE at all (while FSF-GNOME does borrow cospicuously from KDE).

You can choose to avoid detailing this exchange of insults between Richard and others, as you have done with most others, as "they will be forgotten". But you should not lead the reader to believe that RMS did not fight for that change because, boy he did! And he won, too ;-)

I can provide you with my personal notes and list of lins to the major elements of the controversy if that helps you. I also have a summary of it on one of my talk presenations (http://www.lucifredi.com/talks/linuxexpo-milan.pdf).