Thoughts on legacy...

Chris Spencer cspencer at cait.org
Thu Jan 15 14:56:57 UTC 2004


I wanted to post some comments to this list from someone who is part of
producing the full releases and legacy on them.  Comments below.

-Chris



>When will people understand that a community project needs more than a
>"one man gang" who just tries to push it forward? Warren Togami's
problem
>is, he can make a project look as if it is alive and happy when in fact
it
>is in serious lack of contributors. Well, at fedora.us, for instance,
it
>is not a problem if some 50+ packages for educational programming
>languages [and other stuff with a very special target group] seem to
stay
>in the queue forever. But with Fedora Legacy, the project should have
>tried to be ready 1-2 months ago already and know exactly how to get an
>update published. E.g. the fedora.us build system is used and won't be
>happy about missing build dependencies (Red Hat's build system is
>different). Instead, much time has been spent on discussing poor press
>releases or web site layout stuff. However, if you check the fedora
legacy
>list, a few people are trying to deal with the recent flood of security
>issues. But the general tendency is, if no one shows interest in a
>particular package and a bit of work on it, there won't be updates for
>it. A recent suggestion has been to track such packages in bugzilla as
>"unmaintained".
>
>As a side-note, there's also the risk that as soon as someone offers a
fix
>for a package at some place without submitting it as an official Fedora
>Legacy update, subscribers of fedora-legacy-list might find that
>satisfactory and no one might do the additional work of getting the
>package approved and published.

I think that this shows that producing and supporting a Linux
distribution is not a simple project on its own for a small
number of non-paid people.  The OSs are pre-existing of course os
that part is done.  Many people I've seen in IRC and/or email
lists mention that they would volunteer to work on a legacy
support project for EOL releases, however I haven't personally 
seen very many people actually step forward to do real work, or 
contribute to the necessary infrastructural and organizational 
efforts that would be required in order for such a Fedora Legacy 
project to truely be successful.

Unpaid volunteers are hard to come by.  I cerainly wish any 
Fedora Legacy project, and its members much success in trying to 
support EOL products, but it is a huge undertaking, and will 
require not one or two, but a great number of dedicated 
volunteers putting in a large number of man hours of work to put 
together all of the infrastructure and organizational bits, and 
to do the work of backporting or creating fixes for legacy rpms 
and building and testing them on all platforms that the project 
wishes to support.  It's not a small task any way you look at it.

To get started, the organizational bits need to be in place 
first, and some mechanism for quality control needs to be 
developed and manned.  Then people need to volunteer for specific 
packages, and be willing to also be assigned to do something.  
That's the only way I can see such a huge effort succeeding, and 
that's enough work right there just to support a single OS 
release alone, rather than all EOL products.

Security bugs are the biggest problem, as they hit you when you
least expect it, and have tonnes of existing work on your plate, 
be it job, school, other projects, etc. yet you need to drop what 
you're doing, and fix some possibly obscure complex security 
issue of which you may have a patch fed to you, or may need to 
debug and figure it out yourself, and make sure you're not 
breaking anything while you're trying to whip up the fixes 
quickly and get them tested (by someone else ALWAYS).

So, it is a very huge project to undertake for anyone at all 
IMHO.  Since it's all volunteer work, and all volunteer work has 
non-monetary motivators by definition, perhaps that is a good 
place to start looking to build the project.  What are the 
motivations for someone to join and/or contribute?  Come up with 
some realistic motivations, and post them to the site, and 
advertise getting involved.  Get the infrastructure together, and 
have a leader who has the time to spare to be the leader.

That's the best advice I can try to provide for now anyway.  I'll 
do my part by keeping XFree86 4.3.0 building on RHL 8.0 and 
newer, and 4.1.0 building on 7.[12].  4.2.x is free for grabs to 
anyone who wants it though.  ;o)

Take care,
TTYL

P.S. Feel free to forward between mailing lists if you think 
others may benefit from my words above to aide the Fedora Legacy 
project.



-- 
Mike A. Harris     ftp://people.redhat.com/mharris
OS Systems Engineer - XFree86 maintainer - Red Hat


"There is a lot of speculation and I guess there is going to continue to
be a lot of speculation until the speculation ends." - George W. Bush on
October 18, 1998






More information about the fedora-legacy-list mailing list