[nSLUG] Richard Stallman in Halifax

Daniel Morrison draker at gmail.com
Thu Jan 29 23:53:43 AST 2009


2009/1/29 Michael Crawford <mdcrawford at gmail.com>:
> Quite a long time ago, I really clued in to the fact that few people
> understood what I meant when I said that software was libre and not
> gratis, and those few who did understand, hated me for using the two
> terms at all.  I'm not completely clear as to why...

Because it's foreign. Call me a bigot for saying this, but I find many
North American unilingual anglophones are scared of foreigners, their
language and their culture.

> So for many years I have adopted the convention that I capitalize Free
> when it's as in Freedom, but use lower case when it's free as in beer:

Excellent idea.

> You may be interested to read my following essay, which got me banned
> from Webmasterworld within minutes of my posting it in their copyright
> forum, for what they claimed - I feel incorrectly - was a violation of
> their Terms of Service.

Well I don't know their terms of service, but your essay _is_ about
copyrights, and it is not offensive or vulgar, so it's hard to imagine
what the problem is...

>   Why I'm Proud To Be A Dirty GNU Hippy
>   http://www.goingware.com/notes/dirty-gnu-hippy.html
>
> I should say that it was long ago pointed out to me that there are
> some errors in the piece that need to be corrected.  I just haven't
> gotten to them yet.

I didn't notice any in a quick read. Generally I agreed with most of it.

I completely agree that for most works of art, a license that allows
derivation is the wrong kind of license. Only if the art is _intended_
to be a cooperative project from the start can this sort of license
apply.  It might be argued also that FSF software is _intended_ to be
a cooperative project also.

For music... <sigh>.  There is a difference between 'stealing' a tune,
releasing a song without crediting the original author, and
'paraphrasing' a melody for inclusion in your own work.  In the same
way that I could write an essay that argues the same points as yours,
but in my own words, and that would not be a violation of your
license, nor would it reduce the impact of your essay (so long as I
didn't try to pass off my work as your original). I think it's
reasonable to write my own music, putting forth a similar or
derivative melody to yours, but my own interpretation.

And in fact this is how both music and literature worked for millennia
before the introduction of "intellectual property" in the last
eyeblink of human history.

> (Anyone familiar with Search Engine Optimization methods will readily
> see how I hijacked the top spot for "dirty gnu hippy" - ahead even of
> Stallman Himself!)

Share the secrets with us, do.

> But after quite a bit of reflection, I decided that, for certain kinds
> of software at least, the Cathedral serves the needs of the *user* in
> a way that the Bazaar does not.  That's because by *refusing* to
> release software at all until it's good and ready for the user to
> actually *use*, you avoid subjecting them to incomplete and
> potentially - often, actually - buggy products.

Can't agree with you any more here. It's a free market. Buyer beware.
Just because there are piles of crap free software out there, doesn't
mean that there isn't piles of crap commercial software out there too.
And I think RMS is right -- there is relatively more good free
software than there is good commercial software.  Or are you going to
argue that MS Windows, MacOS X, and all the crapola from Adobe works
better than glibc, gcc, make, bash, and most of BSD?

Imposing a Cathedral model of development will not fix the problem:
that some people simply have poor standards and release poor software,
under whatever license.

(You do say: "for certain kinds of software".  Perhaps you are talking
about mission-critical software such as vehicle, medical, or power
plant control?)

> And my main complaint with most - not all, but definitely most - Open
> Source products is that despite being a technically savvy user myself,
> I have many, many times had to deal with buggy crapware, where I would
> have been far better off had it never been released, so that I
> wouldn't have had to suffer even trying it out at all.

No, no, no. That's like saying it would be far better to have only
large, strong, healthy trees in the forest.  Try to create a forest
like that, and you'll quickly find that it is far more sick than a
forest with a wide variety of trees.

The Free software world is an ecosystem of projects, and yes, the vast
majority of them are unfinished, buggy, useless, or simply not
interesting.  But every project started off this way once, and we need
all of them to exist, lest we miss the forest for the trees.

> Open Source yes, often Free as in Freedom as well - but often
> completely unusable, and sometimes actually harmful, for example with
> bugs that cause the loss or corruption of end-user data.

If it breaks, you get to keep both pieces. You do keep backups of your
data, right?

> Did anyone here use Mozilla 1.0?

Yes. Oh wait, I thought you said Mosaic 1.0. That too.

> How long did it take them until they
> produced a browser that was ready for anyone to actually *use*"?

Um, they haven't yet. Well, they did produce a couple of good versions
(I was quite happy with Firefox 1.5.x) but for some reason they decide
to end support for it and practically force everyone to upgrade to a
bigger, buggier, and less friendly product.

I almost hate the Mozilla foundation more than Microsoft.  Microsoft
_started_ with proprietary software, and the attitude that putting
users on the upgrade treadmill would be the core of their business.
Mozilla, on the other hand, started with Free jewels of software
(albeit in "rough cut" form), and have regressed into an upgrade
treadmill mentality, with a cathedral structure in which well-meaning
bug reports and feature requests are ridiculed if not ignored
outright.  Whereas one was always evil, and knows only evil, the other
was good, and yet chose evil. Bastards.

> I get a lot of heat about it - quite often, people hurl insults at me,
> claiming that it's just vaporware, but that's why my own Free - and
> not Open - Software Ogg Frog hasn't been released, and *won't* be at
> all, until it's entire planned 1.0 feature set is *completely*
> implemented, and is as rock-solid and as bug free as I can possibly
> make it:
>
>   http://www.oggfrog.com/free-music-software/   <--- Ain't No
> Downloads To Be Found

Sorry.  I agree that it's frustrating and annoying to download
half-finished free project after half-finished free project and
finding that none of them work, but the one thing I hate even more is:
finding a project that looks like it will meet my needs, claiming to
be Free software, but refusing to allow downloads because "it's not
finished yet".  So I won't be visiting your website.

I don't doubt that your project exists, and is not vapourware, but it
is NOT free software, because 0) I can't run it 1) I can't study the
source code and modify it 2) I can't redistribute it 3) I can't modify
it and redistribute my modified version.  Failed on all four counts.

> Everyone urges me to take advantage of the Many Eyeballs to be had by
> Releasing Early and Releasing Often.  But my position is that I just
> don't see the point of asking anyone else to apply their Eyeballs
> until the two I've got myself are unable to find any more bugs - and I
> am masterful both as a debugger and as a QA engineer.

The point is 1) it will be done faster if more people do it 2) there
will be more interesting ideas contributed 3) no matter how masterful,
you WILL miss some bugs that others may not 4) even with some bugs, it
may be very useful to some users (or to put it another way, your
definition of "completely implemented feature set, rock-solid
operation" may be wildly more strict than someone else's) 5)
cooperating with others to code and smash bugs enriches not just you
and your project, but others and their abilities to learn from your
coding experience -- in short, it's good for the community 6) a
partially finished project that attracts additional developers helps
focus energies, whereas making it unavailable merely encourages other
(possibly less capable) coders to create their own, likely less
capable and buggier versions

OK I've stopped with 6 reasons off the top of my head, but if they
don't convince you, I'm sure I can come up with more.

> I can see how Release Early, Release Often would make sense for
> developer tools - but not for the sort of non-technical teeny-boppers
> that are Ogg Frog's target market.

Teeny-boppers?  Hmm.  Quick way around that: "My software is not ready
for general release, so there are no downloads here, but if you want
to check out the work in progress, get it from
cvs/svn/git://oggfrog.com/ogg-frog"

> For one thing, I've never felt that it was right to say GNU/Linux when
> so many components of a Linux distro these days have nothing at all to
> do with the FSF or GNU - but I wasn't really appreciating that so many
> of the components of any distribution were developed by the FSF, even
> Stallman Himself, but aren't readily apparent even to
> technically-savvy end users like me.

Just check the list:

ftp://prep.ai.mit.edu/pub/gnu/

a2ps, aspell (and dictionaries), auto{conf,gen,make}, bash, bc,
binutils, bison, coreutils, diffutils, emacs, enscript (with a2ps
means BOTH major ascii->ps packages!), fdisk!, findutils, flex (with
bison means BOTH major lexical parser packages!), fontutils, g++, g77,
gawk, gcc, gdb, gdbm, gettext, ghostscript, ghostview, gimp, glibc,
gnome!!, grep, groff, gtk, gzip, halifax (!), less, libg++, libstdc++,
lilypond (for your music, check it out, it's beautiful), m4, mailman,
make, nano, ncurses, readline, screen, sed, sharutils, tar, termcap,
termutils, wget, zlibc.

And I only picked the ones I immediately recognized as essential or
very common (except halifax which I threw in for a laugh; it's a
fax-modem package. And lilypond, because you're a musician and after
reading http://lilypond.org/web/about/automated-engraving/big-page I
resolved to never again use any piece of music notation software that
wasn't _at_least_ as serious as the lilypond guys at making it look
good. Fond memories I have of flipping through my mother's century-old
Edition Peters hand engraved sheet music...).

I mean, where would we be with tar and gzip? Without grep? What about
this nSLUG list, it runs on mailman! sed and awk? No man pages without
groff! Command-line completion and history in most programs wouldn't
exist without readline. Before GUIs we would have been screwed without
ncurses, which itself would be useless without termcap!

I've already mentioned glibc, gcc, and make, but I should say again:
it's really the ENTIRE development toolchain:
flex/bison - lexical tools
make - coordinates everything
gcc - compile to assembly
binutils - includes assembler to binary format
libc, libstdc++ - for all your library calls
groff - so you can 'man' the library calls and work out how to use them
auto{conf,gen,make} - for portability
gdb - for debugging afterwards
diff - for patching bugs...

And then I haven't even mentioned the quality documentation that is
considered a core part of every official GNU project.

Sorry, I went a little over the top here, but even I'm getting
astounded that RMS hasn't burnt Linus Torvalds in effigy for stealing
so much glory, writing a single project (albeit a very important one)
on the basis of such a huge base of GNU software.

:)

-D.



More information about the nSLUG mailing list