[NTLUG:Discuss] RMS's Speach

[Christopher Browne]

On Sun, 21 Jan 2001 21:37:20 CST, the world broke into rejoicing as
Steve Baker <sjbaker1 at airmail.net>  said:
> kbrannen at gte.net wrote:
>> For the record, I have no problem with commercial software.  I
>> think it has a place, a small one, but it's not inheritantly evil.

>I agree - programmers need to eat, so there has to be commercial
>software.

I don't want to be blindly following RMS, but I think there needs to
be a _bit_ more of an ethical framework than that.  I could use
largely the same constructs to generate:

 "Torturers need to eat, so there needs to be a commercial sector of
  torture and inquisition."

There needs to be some more useful distinction than just "programmers
need to eat."  After all, inquisitors also need to eat :-).  Music and
movie executives need to eat.  Lawyers need to eat.  None of which
justifies the more aggrevious excesses of those various professions.

> HOWEVER, I don't think that software that forms part of the
> infrastructure of our computer systems should be commercial.

> Hence, I'd be happy to pay for a decent wordprocessor - but I want
> my NFS server to be OpenSourced.  I'll pay for a game - but not an
> OpenGL implementation.

What if someone starts building "document management systems" atop the
decent word processor?  

Microsoft would like nothing more than for that to take place with
Word; they'd doubtless _love_ to see people writing applications that
assume Word as "basic infrastructure."

One of the less-well-articulated things about the "GNU/Linux"
controversy is the fact that there are some interesting extreme
positions.

One of the "GNU Positions" is that "Linux" is merely an operating
system kernel, which requires a whole bunch of "GNU things" in order
to have the user space that makes it useful to you and I.  And
regardless of whether or not you agree with RMS on the "I want credit
for the name" side of things, this position is not ludicrously
unsound.  The software that Linus Torvalds works on is indeed an OS
kernel, and he primarily _doesn't_ work on user space stuff.

The flip side is that the traditional view of "The GNU System" has
been that it was really quite amorphous; it has _never_ been
completely clear how far the OS went, and where that "stops" and the
"application space" starts.

Look at the architecture of GNU Emacs: There's a sort of "kernel" that
is a Lisp interpreter written in C; over the years, the set of Lisp
programs that come with an Emacs install have grown to include pretty
much "too much for anyone to figure out how to configure."  

Which parts of the Elisp are "system"?  Which are "application"?

I think the rhetorical answer is that over time, what's "in the
system" is intended to grow to include whatever stuff proved useful.
And I rather think that this is the intent of "The GNU System;" once
something has proven useful, in a sufficiently general way, it is a
good thing to let the understood scope of the system grow to include
the useful stuff.

Getting practical again, people tend to _expect_ that a Linux install
will include Perl, Apache, and probably even Netscape.  A modern
install of Windows pretty much mandates having IE as a basic system
component.

>> My biggest gripe with commercial software is that many companies
>> price it only for businesses, putting it out of the range of
>> individuals.

> That's a hard problem.  You can make more money selling software for
> $1000 a copy to businesses than for $20 a copy to individuals
> because so much of your costs are in customer support.  Fewer
> customers == less support.

> And games *do* sell for $30 or so - it's only things like C++
> compilers that come in at the $300 to $500 mark...since the vast
> majority of purchasers are commercial users.

By the way, there are some confounding factors here:

-> The business license probably comes via direct relationship between
   vendor and customer, so that while the vendor has to pay for a few
   lunches out of the $1K, they at least _got_ the whole $1K

-> A box sitting on store shelf that sells for $30 likely results in
   $10 getting back to the makers.  Much lower efficiency of "money
   collection" due to there being a bunch of extra layers of
   intermediaries involved.

>> There is a similarity with books.  I always buy paperback books
>> (for pleasure reading) instead of hardbacks; same sort of reasoning
>> (not to mention that what I'd pay for 1 hardback would buy 3-4
>> paperbacks).

> Yes - but realise that cheap paperback novels cost $3 to $5 - but
> technical books - also in paperback and with a similar number of
> pages cost about $50.

Going hardcover tends to add a _bit_ to the price of a technical book,
but not vast amounts...
--
(concatenate 'string "aa454" "@freenet.carleton.ca")
http://vip.hex.net/~cbbrowne/internet.html
Where do you want to Tell Microsoft To Go Today?