RMS and nblug

Mitch Patenaude mrp at sonic.net
Sun Dec 16 09:18:59 PST 2001


On Saturday, December 15, 2001, at 11:00 , Jake wrote:
>> Mitch said:
>> so he always says
>> it should be called GNU/Linux, rather than just Linux.
>
> I think he has a good point, I think I would be pretty bitter in his 
> shoes.
> However I think that he does alot of work and gets only minor credit.

While there is a lot of GNU stuff, there is also a lot of stuff in most 
distributions that isn't GNU.  What about Apache, PERL, Python, PHP, 
XEmacs, KDE (and widgets and tools and apps).

Those are a major part of most dists, but aren't put out or maintained 
by FSF or GNU (though most are released under the GPL).  While RMS's 
contribution to the free software movement is great, he has a reputation 
of being exceedingly difficult to work with, and wants FSF to have the 
last say over all things having to do with Free software.  He even 
objects to the term Open Source, and other licenses, since they dilute 
or destroy "fundamental rights".

I sometimes think that his insistence that it be called GNU/Linux is 
just bitterness about HERD not being a success.  But it easy to see why 
some of his projects had problems.  The one I know well is emacs, having 
fought many battles with it.   Back in the day when most unix-based 
computing was done through dumb terminals, or terminal emulators, and 
those terminals often used ^S as software flow control (XOFF), and ^H as 
backspace, he refused to change the default keymappings of emacs, which 
used ^S as search and ^H for help.  His position was that "it isn't my 
fault if *your* terminal is broken",  even though greater that 90% 
behaved like that.  While it was possible remap all those by writing 
some elisp, it made a big barrier to new users.  Many times when I 
introduced new people to emacs, the first time they tried to delete a 
mistake they got themselves buried in 20 layers of help (The more times 
you pressed ^H the deeper you got into the help system).  many times 
they would try to start a serach according to the tutorial and instead 
would end up (apparently) locking up their terminal.  (Once the remote 
end gets an XOFF (^S), it stops all output until it gets and XON (^Q), 
but most people didn't know that, and would just power-cycle their 
terminals and lose hours of work, and then vow to never use emacs 
again).  Even the fact that it still uses a lisp variant as a language 
shows how recalcitrant RMS can be.  LISP is a throwback to the early 
80's, when it was the "language of the future" because of it's ties  to 
AI.   Would you use a window manager or web server that required you to 
program in Modula-2 or PROLOG?

> The proof is that any linux newbie will not have a clue who RMS is but 
> will
> know of Cox, Linus and perhaps even ESR.

How many remember RMS's somewhat radical policy on privacy, passwords 
and file permissions?  I remember that he used to publicly decry 
passwords and file permissions.  His basic premise was that "data wants 
to be free", and that any attempt to limit access to data was violation 
of fundamental human rights.   He used to set the password on all his 
accounts to 'rms' and publicize that fact.  His accounts became such a 
common place from which to launch attacks that few places would give him 
one.  Even his own institution (MIT) finally threatened to take away his 
accounts unless he started using strong passwords.

Personally, I think calling it GNU/Linux does a disservice to the 
thousands of projects that *aren't* GNU, but are an integral par t of 
the linux experience.

    -- Mitch



More information about the talk mailing list