Kurt B. Kaiser kbk at
Sat Aug 2 18:47:30 PDT 2003

Matthew Dillon <dillon at xxxxxxxxxxxxxxxxxxxx> writes:

>     I, and most people on this list, are obviously strongly in the
>     UNIX/open-source camp.  I believe that it is extremely important
>     to have control over all the supporting software used to develop
>     any major project because without that control one risks
>     depending on things which often become obsolete over time and
>     wind up no longer being supported.  Software tends to become
>     obsolete fairly quickly (within a few years).  For example, MS
>     just recently stopped supporting NT.  Of course, this happens in
>     the open source world too but support tends to continue for a
>     far longer period of time because, being open source, those
>     people with an interest and an investment in a piece of old
>     software can get the code and form a loose association to help
>     maintain it long beyond its mainstream lifetime.  There are
>     open-source programs that are over 20 years old still being
>     supported, something you simply cannot find in the commercial
>     world outside of a military contract.

Well, I think MS expects the customers to move to XP, the current
version of NT.  Take RedHat.  I got a nice four year run out of RH6.2
with good errata/security support.  Now they have reduced support to
one year after release (unless you want to pony up for the enterprise
edition, then you get five years :).  They are ending all support for
everything before RH9 by the end of this year, and RH9 ends 30Apr04!
They claim that certain "popular" releases may be supported for a
longer period but AFAIK none have been identified as yet.

OpenBSD is supported for two releases, so that's one year also.

To me, one of the key points about an OS is how well the package
management system works and how much time is going to be spent keeping
the system current with the errata.  I can understand why it's
necessary to drop support on older releases, but how easy is it to
upgrade to the next release and then keep it patched?  Right now too
much time is spent with mergemaster grovelling over /etc.  Why should
I have to merge passwd and group files because a system user or group
was added at the new release?  Why not have separate files for system
and user?

Most Windows users don't upgrade, they just buy a new computer.
Most RH users seem to re-install rather than attempt an upgrade
across a major release.  With BSD and Debian, there's hope.


More information about the Kernel mailing list