On Wed, 5 Jun 1991 13:03:01 EDT Valdis Kletnieks <[log in to unmask]> said:
>Please note that anything Rich writes will most likely be covered by the
>GNU Public License, which has been referred to as the "gnu public virus"
>due to some of its "copyleft" features. Please read it *carefully* to
>determine if the conditions are acceptable -
Freedom of software is one of the aspects of Richard Stallman's personal
"quest" against today's most dangerous criminals:
1. People/companies who *sell* software (software should be free).
2. People/companies who work on making computers more secure (computers
should have no "locks").
Your average Richard Stallman program is not unlikely to contain
statements such as '/* Someone put something here for security, I removed
it */'. Feel free to ftp /etc/passwd from the anonymous ftp directory of
the GNU machine, it's been placed there on purpose. Here are a couple
quotes from the EMACS release notes:
-------------------------------------------------------------------------
"GNU Emacs does not support encryption. Down with security!"
;; you can of course turn this off by doing
;; (setq ftp-password-alist 'compulsory-urinalysis)
(defvar ftp-password-alist () "Security sucks")
-------------------------------------------------------------------------
>Other than that, Rich is quite the programmer - both the GNU Emacs
>editor and the GNU gcc compiler are regarded as among the best in their
>respective fields...
I agree with Valdis, except that he forgot to mention that the field in
question is known under the name of "Unix" - an environment where the
"standard" system editor is 'vi' and where C compilers often do not have
any "real" optimizer. I have a copy of GCC on a (quoting RMS's release
notes) Vomit Making System machine and, while the code it produces is
more than decent, it is the slowest compiler I have ever seen. Also, I'm
afraid Ken's C compiler produces less vomit per unit of source code -
unless you ask for a listing, but then everyone knows that compiler
listings and cross-references are intrinsically useless (proof: if they
weren't, GCC would know how to make one).
Now that the intro is over... I'd be delighted to hear that RMS is going
to make a mailing-list server for GNU ("GNU's Not Un*x" - "GNU's Nearly
Un*x" would be more accurate, but for some reason the official name is
the former), so that I can stop having to answer complaints from Un*x
people who are unhappy about the fact that end-users do not have access
to privileged commands, when they can so easily fake the 'From:' field
and my passwords are not secure since they are not crypt(3)'ed anyway -
so why not make the LIST files world writable and set up a 'subscribe'
guest account? However, I'm afraid that the newspaper in question (which
I haven't read) was probably talking about a 'list processor'. Obviously,
a list processor is not a LISTSERV but a compiler or interpreter for the
LISP language - it's so obvious that any self-respecting newspaper
reporter should know this, and there really is no point saying it
explicitly. The LISP language is what EMACS macros are written in - look
at the quote beginning with '(setq' above for an example (oops, I meant
'(setq of course :-) ). My understanding (but I may well have
misunderstood) is that RMS would like LISP to be the system language for
GNU, and now that he's got a C compiler it would make sense for him to
work on a LISP garbage collector (oops, interpreter).
Disclaimer: I am a criminal, I use brain-damaged operating systems on a
daily basis and never liked having to take care of the garbage bags. In
other words, I am hopelessly asinine and you shouldn't make the mistake
of taking me seriously.
Eric
|