Comments on: Rethinking the Linux distribution business model http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/ on emerging platforms, the open source business opportunity, and the commoditization of software Mon, 01 Dec 2008 22:17:06 +0000 http://wordpress.org/?v=2.3.2 By: Ryan Dooley http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-28 Ryan Dooley Wed, 31 Dec 1969 17:00:00 +0000 http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-28 Actually, why isn't Debian in the Enterprise. :-) What I tend to think is: Progeny needs to do for the Debian community as RH has done with Oracle, Maya, whoever, gaining certifications for their commercial products to be supported on. Actually, why isn’t Debian in the Enterprise. :-)

What I tend to think is: Progeny needs to do for the Debian community as RH has done with Oracle, Maya, whoever, gaining certifications for their commercial products to be supported on.

]]> By: Scott Smith http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-29 Scott Smith Wed, 31 Dec 1969 17:00:00 +0000 http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-29 Someone needs to do for Linux what Apple did for BSD. Don't just glue the bazaar together. Make a tight, attractive product which users and developers will be excited to use. Someone needs to do for Linux what Apple did for BSD. Don’t just glue the bazaar together. Make a tight, attractive product which users and developers will be excited to use.

]]> By: Ryan Dooley http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-30 Ryan Dooley Wed, 31 Dec 1969 17:00:00 +0000 http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-30 I don't know if you have seen this (it was on OSNews today) but the text is form Miguel de Icaza (of Ximian/Mono fame)... <a href="http://primates.ximian.com/~miguel/texts/linux-developers.html">http://primates.ximian.com/~miguel/texts/linux-developers.html</a> He talks about ABI/API stability in Linux and in particular GNU/Debian. I think his point is valid and it does suggest another reason why Debian is not considered to be "Enterprise". Debian needs to offer an API/ABI stable release for most users. I don’t know if you have seen this (it was on OSNews today) but the text is form Miguel de Icaza (of Ximian/Mono fame)…

http://primates.ximian.com/~miguel/texts/linux-developers.html

He talks about ABI/API stability in Linux and in particular GNU/Debian. I think his point is valid and it does suggest another reason why Debian is not considered to be “Enterprise”.

Debian needs to offer an API/ABI stable release for most users.

]]> By: xipmix http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-31 xipmix Wed, 31 Dec 1969 17:00:00 +0000 http://ianmurdock.com/2003/10/31/rethinking-the-linux-distribution-business-model/#comment-31 Icaza's posting is twaddle. Debian stable is _so_ stable it gets out of date. No one in their right mind would run an organization with more than three machines on testing or unstable. Reminder: stable means "no new features". How, then, can the API be broken? The actual problem is that Ximian's developers are using the latest toys, er, versions of everything. They charge ahead with new versions of things that have yet to be incorporated in a stable debian release. This is partly because of the amazing amount of flux in the GNOME apis. So their actual problem is that Debian's release cycle is about twice as long as it should be, to keep up in this space. I have to say the length of time dismays me too. Annually would be great. A second problem is the bug fixes to packages in 'stable' frequently end up being added to the version in 'unstable', unless they are security holes or the bug causes major havoc like data deletion etc. Personally I think this is one of Debian's main problems for enterprise use. Users report bugs in the current version of the package, and they want that version fixed, not the latest version which will not appear on their 'stable' box for many months. It's interesting to see him say that GNOME has learned about ABI stability. By making every possible mistake they could in the process, I would say. Thank goodness Sun took to them with the clue-bat. Who do we get to thump the Genuine Cowboy Compiler team? IBM? Icaza’s posting is twaddle.
Debian stable is _so_ stable it gets out of date.
No one in their right mind would run an organization
with more than three machines on testing or unstable.
Reminder: stable means “no new features”.
How, then, can the API be broken?

The actual problem is that Ximian’s developers are
using the latest toys, er, versions of everything.
They charge ahead with new versions of things that
have yet to be incorporated in a stable debian
release. This is partly because of the amazing
amount of flux in the GNOME apis.

So their actual problem is that Debian’s release
cycle is about twice as long as it should be, to
keep up in this space. I have to say the length
of time dismays me too. Annually would be great.

A second problem is the bug fixes to packages in
’stable’ frequently end up being added to the
version in ‘unstable’, unless they are security holes
or the bug causes major havoc like data deletion etc.
Personally I think this is one of Debian’s main
problems for enterprise use. Users report bugs in
the current version of the package, and they want
that version fixed, not the latest version which
will not appear on their ’stable’ box for many
months.

It’s interesting to see him say that GNOME has
learned about ABI stability. By making every possible
mistake they could in the process, I would say.
Thank goodness Sun took to them with the clue-bat.
Who do we get to thump the Genuine Cowboy Compiler
team? IBM?

]]>