Subject: Re: street performer protocol
From: Crispin Cowan <crispin@wirex.com>
Date: Wed, 24 May 2000 21:56:06 +0000

Ian Lance Taylor wrote:

> You omitted some other cases in which one FSB supports a product:
> Crynwyr, Netscape, SleepyCat, OpenNMS, InterBase.
>
> I suppose I think the sample size is too small to make any judgement.

I'd agree that it's smaller than I'd like, and lots of additional data points certainly
helps.


> I note something interesting about your examples.  In the cases you
> listed in which multiple FSBs supported a product, none of the FSBs
> actually wrote the main product.  They did some work on it, but they
> didn't write the bulk of what they are providing (Linux providers

That is interesting.  It could indicate at least two things:

   * It might suggest that if an FSB originates a product, the "first mover" and "we
     know this code best" advantages are sufficient to keep other vendors out.  This
is
     very interesting, because the first mover bears a *much* greater portion of the
     cost.  A counter-example:  RT/Linux.  Yodaiken & co. did most of the work, but
I
     hear (2nd hand) that a competing vendor snapped up his open source code and went
     into head-to-head competition with him.
   * It might only suggest that vendors are only motivated to move first if there is
no
     libre product in that category.  Stephen suggested this in his recent post.

> You seem to be arguing that because libre software often becomes
> better than proprietary software, it follows that FSBs often become
> better than proprietary vendors.  But that doesn't follow.

I have seen a *lot* of straw-man arguments going around in the last few days, which
is
why I tried to be very specific in what I'm claiming.  No, the above is not what I'm
claiming:  the difference is that OSS is better for consumers that proprietary software,
not that it is better for the vendor.  On the contrary, OSS seems to accelerate the
slide towards commodization.


> Good libre
> software can kick out the underpinnings of a market, making it hard
> for anybody to make money.

That is the commodity end-game, where margins are slim, and making money is based on
service.  This is what RMS has advocated all along:  you *don't* make money on the
software, because it is free.  You make money on the service.  It would seem that free
software is actually doing what was intended :-)

Crispin
-----
Crispin Cowan, CTO, WireX Communications, Inc.    http://wirex.com
Free Hardened Linux Distribution:                 http://immunix.org
                  JOBS!  http://immunix.org/jobs.html