The Rails Wheels licencing system and Open Source

[Michael Tiemann]

On Thu, Sep 4, 2008 at 4:58 PM, Mark James <mrj at advancedcontrols.com.au>wrote:

> Michael Tiemann wrote:
>
>  The Red Hat model is not the only commercially successful model.  The
>> company I founded, Cygnus, was based largely on developing free (GPL)
>> software for $$$.  We built that company to $25M/year in annual revenue
>> before being acquired by Red Hat (who also had approx $25M/year in revenue
>> at the time).  You can read how we made Cygnus work in the O'Reilly book
>> Open Sources.  My chapter is freely downloadable, and gives revenue figures,
>> profits, and a detailed analysis of what worked and what didn't.
>>
>
> I found and read that page yesterday Michael
>
>  http://oreilly.com/catalog/opensources/book/tiemans.html
>
> Very interesting and inspiring. I think I need to find another founder.


Heh--my advice is to find somebody complementary to you.  Too similar and
you'll fight like brothers!


>
>
>  The Red Hat model has proved itself to be very scalable ($500M+/year and
>> counting), but we still bid and win GNU development contracts, and funded
>> development is still a part of Red Hat's business.
>>
>
> Without Red Hat and Cygnus, I don't think OSS would have the
> respect and spread that it does.


Thanks--I'm sure you mean that as a compliment.  I honestly believe that
sooner or later /somebody/ would have done it successfully.  Well, maybe not
sooner: Sun had all the stars in their favor (including the source code to
BSD Unix), and they completely and comprehensively blew it (although not
without so much success along the way that their failure wasn't apparent
until perhaps only 8 years ago).

M


>
>
> Do you insist that all development work be committed to the
> public repository?
>

I don't insist on anything other than that people obey the terms of their
licenses.  In the case of the GPL, of course a thousand programmers can fork
the code a thousand ways, but economics and the network effect so strongly
favor upstream participation that only the most clueless or the most
courageous don't practice good upstream hygiene (and commit their code to
the public repository).  Sometimes the upstream repository won't accept J
Random Submissions, in which case code forks are like storage tanks for new
ideas.  In the case of Linux and GCC, those tanks can hold good code for
years, but the waiting pays off.

M
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20080904/cb2bf3bd/attachment.html>