:-)
We really seem to have got bogged down in trivialities, and are trying to
decide things in the wrong order (hence, partly, all the arguments.)
I spent last night learning about Zope (http://www.zope.org) . There is a
12-page Zope tour at http://www.zope.org/Tour/ar01.html . Please go
through at least that before you comment ;-)
Of the things we need, it provides:
- an enterprise-level management system for a large website
- low barrier to entry for contribution
- extremely fine-grained permission controls (much better than currently)
- easy to add interactivity (another big must)
- Administerable via Web, FTP or WebDAV
- Excellent (almost CVS level) change control
- Standards support (HTML, CSS etc.)
- Simple URLs which relate to the content (as everyone seems to want)
- Can embed MySQL calls for easy Bugzilla etc. integration
- Built in search engine
- You can rearrange stuff easily (which is very hard with CVS)
- The wrapper implementation would be trivial
- Easy migration path
In addition, mozilla.org has very good relations with the Zope folks and
we may well be able to borrow a Zope guru or two to help with the
implementation. Having someone who knows nothing of Mozilla working on it
will do wonders for its accessibility.
I also assert that, in Brendan's words, Zope is "the best open-source
solution to solve a well-understood problem."
I assert that we use Zope for www.mozilla.org version 2. I further assert
that, in order for us not to use it, someone will have to:
- name a problem that we have that Zope does not or could not solve - i.e.
using Zope prevents us from solving it
- name an alternative basis for the website that solves all the above
problems and that one as well
Making this decision gives us a firm basis to work from, and a new angle
on a lot of the other decisions we are making. I think it's the first firm
decision we need to make - either for or against Zope (and, if against,
then for some other system.)
Whatever your views on the Zope question, please also express views on
that last assertion.
Gerv
