[OpenAFS] New OpenAFS website proposal.

Russ Allbery rra@stanford.edu
Fri, 26 Nov 2010 13:32:05 -0800

Jakub Witkowski <jpw@jabster.pl> writes:
> 2010/11/26 Simon Wilkinson <sxw@inf.ed.ac.uk>:

>> We recently switched from TWiki to ikiwiki for the AFSLore
>> content. Whilst the content was all moved over, little has been done in
>> terms of skinning that content. I suspect that it is this triumph of
>> content over presentation that makes you feel that it looks broken. The
>> huge advantage of ikiwiki is that we can maintain it using the same
>> tools as we use for our source and documentation repositories.

> Out of curiosity, what happens to local changes?

ikiwiki uses the Git repository as its database.  So all changes, however
they're made, go to the same place.

>> We have a very small number of people looking after our infrastructure,
>> and their time is pulled in many different directions. This means that
>> simplicity, and reuse of existing tools, are both huge advantages for
>> us.

> This sounds like an incentive to integrate all of the tools and content
> into one system.

On its surface, yeah, but what happens in practice is that few of us get a
lot of dedicated time to work on OpenAFS, so what's the easiest is to
reuse things we're already familiar with.

I think any of the various content management systems would be great, but
simultaneously I think we're all worried about having to maintain them
going forward.  The biggest problem is lack of resources and skills in
deploying something fancier.  One advantage to having everything in
something simplistic (like the current static pages) is that at least you
know you'll always have people around with the knowledge of how to do
something with it.

It's not an obstacle that's impossible to overcome, but I think it's the
biggest worry about a content management system.

> Agreed. I believe we should start with a slightly different question,
> though: whether we want the site to be a read-only broadcast medium or
> do we want user interaction.

Personally (and this is just a personal opinion, not a deciding vote or
anything), I think user content additions on the web site are a huge drain
of resources from the project.  I'm not horribly happy even with the
fairly open wiki editing policy because many of the edits end up being
spam.  We definitely don't have people with time and energy required to
police user additions.

Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>