Re: HOW to push to internet

On Thu, Feb 28, 2008 at 11:21:46AM +0100, Cédric Villemain wrote:
> We don't exactly now all the things we want to write. But ...
>
> * we need to be able to translate webpage
>
> * 'interactive' wiki does not seems to be mandatory

Do you by this mean a wiki used for discussions? If so, I agree.

> * full static pages are just pain to change when needed ( do pgweb team
> member confirm me ? )

Yes. There are reasons why they are good for the main site, like
scalability to deal with slashdot effects or whatever, but I expect a lot
less traffic go a site like this.

> Thus, twiki or dokuwiki are *very* simple to install/maintain. They are
> like static (no DB, only flat files).
> About mediawiki, it can be very good (I like the job Greg did), it is
> tsearch2 propulsed !
> Drupal is already deployed in postgresql.org infra, and I think Selena
> just make it like a standart to pugs ? Drupal got also some modules for
> i18n.
>
> Let's vote :
>
> * dokuwiki
>
> * twiki

AFAIK, nobody in the infrastructure group has any experience with this.

> * mediawiki
>
> * drupal
> +1

AFAIK, the pugs site isn't actually maintained by the infrastructure group.
It's maintained by Selena. That only means that the infrastructure group
doesn't really know how to deal with it, not that it's a bad product.

Now, I've personally had nothing but trouble with all the drupal installs
I've had, so pen me in for a -1.

> PS: I don't find usefull to add lot of different solutions wich are not
> already well known by at least one of the postgresql.org team member.

We need to define here what we mean by this. It was originally said that we
wanted this to be maintained as part of the main infrastructure, in order
to be able to use the resources there already. In that case, of all the
services listed here, *only* mediawiki is actually in production today.
Which is why it gets my vote.

Now, we don't *have* to maintain this as part of the infrastructure group.
But we really should. That means that if we pick another tool because of
some reason that makes us really need it, it has to (IMHO of course) be
"backed" by someone willing to join the infrastructure team, and make said
tool work alongside all the other parts of the infrastructure. And document
+ inform the rest of the infra team enough that they can help out in this.

//Magnus