Re: www.gnome.org

[Thilo Pfennig <tp pfennigsolutions de>]

Behdad Esfahbod schrieb:
> I probably go dig the requirements discussions from two years ago and start
> playing with Joomla.  That said, no changes in the Plone plan needed.
>   
I wouldn't really start with yet another CMS. I think a lot can be done
now with patches on SVN - many small fixes. I do not think one should
trough away the Plone solution - thats not what I ever wanted to suggest
- although I prefered Drupal for many reasons. But now that some work is
done it should not be wasted. Unless the people working on it give up.

I think what should be needed is:

a) A Beta deadline - meaning a date when the site should really running.
So a finished status for testing. And what would be especially
interesting from my perspective is how much  of the www.gnome.org/* that
Plone will than be able to deliver. I did a quick count that might be
very unaccurate:
$ find ./ -name '*wml'|wc -l
167

So this are 167 pages maybe. If this is wrong please correct that count.
My guess is that the Plone Beta maybe will have only 40 pages at time of
Beta. The rest would need to be hacked in by volunteers. My estimate
would be that Beta testing plus hacking would need three months after
the Beta release date (about 42 pages/month) .


b) A date for the final release. I would say if on March 2009 the site
would be 100% complete this would be ok. April maybe also - but
everything after that would not be acceptable.

If there would be deadline somebody would need to make them happen and
also take responsibility if they are missed. Somebody would need to step
up and say that this is doable.

And if you say this is not doable one needs to find an alternative
solution that would give similar results in a shorter time.

I think the whole problem started when the CmsRequirements started to
grow. Like on localization: What if Plones localization support is
really better when we will not get a working installation? Plone is
great in many fields - but this does not mean at all that we would ever
get to the point to use it to the full extend.

The practical approach would be: WGO wasnt build with version control
and a build system for many years - every modern CMS will give some
practical advance. The option to keep the current build system was
always in place if the one alternative would not make it - or as long as
it would not make it. So this was an option in play that was never
decided but  it is what we now (still) have, because we did not count in
a failure of the plan.

Some say that we now need to go back to the requirements that were
agreed. I think that if we do we would need to come to the same result
as we have been. Another thing that was missed is that all CMSes
develop. So what GNOME needs is a sufficiently working solution at the
point of the release - it does not matter much what which CMS can do
today or could do yesterday. Also the question is if we have the
possibility to fix some stuff that is not working perfectly ourselves.
This means that some of the hard requirements will actually be not as
important as they looked from a frozen perspective back then. And this
also means that some soft requirements like usability would actually
become more important.

I think the essential point is that a CMS for WGO needs to be workable,
manageable and extendable. And that it should be actively developed by a
community one can trust. Then security sure is a high priority. Also it
should be a system which many of the developers and people who want to
help would be comfortable using it. I think we had a large portion of
users who knew and liked Drupal - I dont remember much of the same for
other CMSes. Also Drupal was already used for GUADEC. From what I have
read Joomla seems to be less secure. What I also like about Drupal is
that you do not have to update to a new version for every vulnerability
but that you get nice patches like this:
http://drupal.org/files/sa-2008-067/SA-2008-067-6.5.patch .

But again, I think the people who will actually do the most work should
have a say here. I think theoretically PHP based systems are always
worse when it comes to security. But I dont know any handy Python based
solution. And with plone I think you need 5-6 people who know this beast
very well and stand besides the users to help them get their work done
and/or understand the system.

Summary: Decide soon (in the beginning week maybe) where to go next with
a rough consensus and then go that way, wherever that path leads. The
worst situation is the open decision with no deadline.

Regards,
Thilo

-- 
Thilo Pfennig - PfennigSolutions IT-Beratung- Wiki-Systeme
Sandkrug 28 - 24143 Kiel (Germany)
http://www.pfennigsolutions.de/
XING: https://www.xing.com/profile/Thilo_Pfennig -
LinkedIn: http://www.linkedin.com/in/tpfennig