Subject: Re: OpenSources "opensourced"
From: Ian Lance Taylor <ian@airs.com>
Date: 22 May 1999 22:58:00 -0400

   Date: Fri, 21 May 1999 04:59:38 -0400
   From: Brian Bartholomew <bb@wv.com>

I'm replying to these messages out of order.  Sorry.

   I do not believe that the term "free software" relates only to the
   rights you have once you are holding a copy.

But that is what we all think ``free software'' means.  Communication
relies on agreed-upon meanings of terms.

If necessary, when you read the term ``free software'' in mail on this
list, please endeavour to translate it into whatever term you think
actually applies.  Criticising people for using words in generally
accepted ways just leads to pointless and meaningless argument.

   I think it also relates
   to the business practices of whoever you got it from.  My position is
   that "free software" should only apply to things that are managed as a
   donation to the world.

A better term for that might be ``charity software.''  I think there
is a place in the world for charity software, but let's not call it
free software.  We'll just get confused.

   Commercial software companies such as Cygnus and Red Hat are
   tremendously less proprietary and offer much more value than other
   commercial software companies, and that's great -- but we shouldn't
   call something they do free software if it isn't managed like a
   donation.  Please don't let "free software" get diluted down in
   popular usage like "open source" was.

I've been in the free software community for a long time now, and free
software never meant charity software.  It's not a matter of dilution;
you are trying to redefine the term to take on a meaning it never had.

The FSF does indeed operate as a charity, but that is not and never
was the economic programming model that they encourage people to
adopt.

   > Are you implying that the Cygnus marketing department is vigorously
   > discouraging the propagation of competitive numbers of gratis copies?
   > Is that really a fair statement, or are you exaggerating for effect?

   I don't see GNUPro up for ftp anywhere, and I would expect to find it.
   Since if everyone who bought GNUPro had downloaded it for gratis you'd
   make less money, I have to believe this is a deliberate outcome.

First I'll note that although you omit attribution you are replying to
a note that I wrote, and you say ``you'd make less money.''  I am not
currently at Cygnus, so my personal money interests are not at stake
here.

Second I'll note that you are merely presuming a conspiracy; perhaps
you simply haven't thought of the real reason.  You say you would
expect to find GNUPro up for FTP, but you don't say why.  What
specific individual do you expect to make it available for FTP?

We agree that Cygnus does not make GNUPro available for FTP, and we
agree that any Cygnus customer could (with the expection of the visual
debugging interface in the most recent GNUPro release).  You assert
that the Cygnus marketing department vigorously discourages this, and
I assert that this is not true.  You have no evidence for your claim,
only an assertion that you suspect a conspiracy.  I do have evidence
for my assertion, namely my past personal association with Cygnus,
which you can easily check.

   That a mailing list with this much industry experience doesn't know
   the correlation between GNUPro components and free software packages
   means somebody has done an effective job of hiding those facts.  If
   Cygnus wanted to make it clear, they could put "We offer the following
   GNUPro products, and you can download them gratis from here" on their
   top web page.

In fact, I think this mailing list quickly resolved that correlation.
You move very quickly from (somebody does not announce the facts) to
(somebody is hiding the facts).  You should watch that sort of thing;
the two cases are quite different.

   Perhaps the vigorous discouragement comes in the form of a high price
   to purchase copies, which discourages companies from giving up the
   proprietary advantage they've purchased:

	   "You've put the software we had to pay lots of dollars for up
	    for ftp where our competitors can download it for gratis?
	    With employees like that we don't need enemies.  <Boot>"

Now you are moving from (Cygnus is vigorously discouraging
distribution) to (companies have a natural reason to not distribute).
Are you really claiming with a straight face that charging a high
price for GNUPro is the same as vigorously discouraging people from
distributing it for free?

   > AFAIK Cygnus never withholds bug fixes, only new features, and those
   > only for a few months.

   Withholding features for a few months could protect a calendar window
   in which to make proprietary revenue.  Does that sound like "free
   software" to you?  This software might be a great value -- and it's
   precisely the mechanism I want to use in my "minimally proprietary"
   software development business plan -- but I don't call it libre.

It meets the definition of ``libre,'' so why don't you call it
``libre?''  What definition are you using, and why is that definition
more correct than the commonly accepted one on this list?

   > Note that since it's embedded, it's not really reasonable to think
   > that the profits come from "software" entirely, but from bundling
   > with hardware.  Then the whole "open source" issue is quite moot.

   Consider the example of a speech-recognition card for the PC.  The
   value was always in the software, not the hardware.  If the speech
   recognition card had been managed as a donation, the source for the
   software could be libre, as could the source for the hardware:
   schematic, pcb patterns, parts list, suppliers, prices.  Since these
   items didn't ship with speech recognition cards, I conclude the cards
   were proprietary.  Libre IP concepts apply quite well to hardware,
   even if the copying cost is higher.

A speech-recognition card for a PC is an example of a software/
hardware product, but it is not an example of an embedded system.  If
we're going to argue on the basis of examples, consider the software
which controls the engine of an automobile.  In that case almost all
the value is in the hardware.

   > Cygnus is scarcely above criticism, but attacking it on FSB as an
   > evil company is akin to the conflict between the Judean People's
   > Front and the People's Front of Judea.

   I would be much happier if "free software" continued to mean "donation
   to the world" like it used to, rather than "a lot less proprietary
   than the competing brand".

``Free software'' never meant ``donation to the world,'' and you are
rewriting history in asserting that it did.  ``Free software'' meant
``no restrictions on copying or use.''

   If the majority of this mailing list believes the techniques Cygnus
   and Red Hat are using to generate profits from gatekeeping software
   don't invalidate their claims to be trafficing in free software, then
   what techniques would?

Changing the license of the software such that it is no longer free.

   Where is the line drawn?

Where the software becomes no longer free.

   How high would the
   price have to go?

There is no limit.

   How long could you withhold features before
   releasing them as a true donation?

There is no limit.

   Could you offer customers a bounty
   not to republish?

Yes, provided they were free to republish with no penalty if they so
chose.

Ian