Re: Launching PostgreSQL KB Project Mark 2

Well excuse me for trying to organize a home for whatever experimentations I
was doing for the KB. Maybe it was better to not have been enthusiastic and
certainly not to have shown *ideas* and *concepts* which have raised all
kinds questions and irritations on people.

> -----Original Message-----
> From: pgsql-www-owner(at)postgresql(dot)org [mailto:pgsql-www-
> owner(at)postgresql(dot)org] On Behalf Of Josh Berkus
> Sent: Tuesday, December 06, 2005 6:17 PM
> To: pgsql-www(at)postgresql(dot)org
> Cc: Magnus Hagander; Marc G. Fournier
> Subject: Re: [pgsql-www] Launching PostgreSQL KB Project Mark 2
>
> Magnus,
>
> > But this is *just a techdocs replacement*, not a whizz-bang KB. It
> > fulfills the techdocs requirements for easy editing and searching, and
> > the communitys erquirements for easy maintenance. It does not fullfill
> > all the other KB requiremetns (I'm sure, though I haven't seen teh whole
> > list since it's not finished - but there are definitly several that
> > aren't). It places the bar a lot lower, making it easier to get done...
>
> Hmmm ... well, any good KB should subsume techdocs as one aspect of its
> functionality. I don't like to see duplicated effort if we can help it.
> Heck, for the much-maligned corporate KB effort step #2 will be evaluating
> existing KB projects to see if any of them meets requirements or can be
> made
> to with a minimum of effort. And even if we do something custom it will
> probably be based on GreenPlum's custom KB I'm currently building (in PHP
> &
> PL/pgSQL).
>
> On the other hand, I wouldn't blame you if you thought that the latest KB
> effort was liable to founder and die and that you needed to have a back-
> up.
> It wouldn't be the first one. But in that case I'd suggest looking at
> adapting something existing (like Bricolage, Framewerk, Drupal, etc.)
> rather
> than coding up from scratch. If you keep your requirements simple, at
> least
> one of these should suffice, and has the tremendous advantage of having
> external code maintainence, documentation, and help.
>
> > We're not against it. I think the irkness (sp? :-):-) ) comes from the
> fact
> > that the ongoing project (Gevik had been regularly demo'ing his
> > progress) was essentially ripped out from here without our knowledge and
> > reformed on a pgFoundry site,
>
> Ah, Kennisgres was put up months ago (like, August). I'd no idea that
> Gevik
> hadn't discussed it here. I can see that that would be a rude shock.
>
> > and was about to be re-specced possibly in
> > complete contradition to what we had originally thrashed out - without
> > so much as a 'by your leave'.
>
> Well, unfortunately no spec document ever came out of dicussion on this
> list.
> In fact, I was on this list for that discussion, and my recall of it is
> that
> there was a lot of shooting the breeze but no real decisions were actually
> made. At least, nothing that had a consensus behind it.
>
> One of the big issues -- in fact, THE big issue -- with increasing
> participation in WWW administration is the total and complete lack of
> documentation for any WWW decisions, infrastructure, or code. While I can
> understand lagging in documenting stuff (like, I have a draft of the
> release
> PR procedure I have yet to discuss online despite being on my HDD for a
> month), it's extremely irrational for people on this list to pitch a fit
> at
> potential contributors for not psychically understanding what WWW wants or
> not reading the WWW list back to the beginning of time.
>
> That is, it's one thing to say: "Hey, you should probably read this
> thread, we
> already discussed it here ____________", and another thing entirely to say
> "You asshole! We already decided that, why didn't you pay attention!"
> Currently, this list has an awful lot of the former.
>
> This isn't just the KB. It affects the whole web infrastructure. For
> example, we've been running on the new web site code for almost 2 years,
> and
> how many translations of the site have there been? Exactly none. Why?
> Zero documentation on how to translate the site.
>
> If our project can insist that all database code patches come with full
> documentation, I think maybe it's time that we start insisting that all
> WWW
> patches come with documentation.
>
> --
> Josh Berkus
> Aglio Database Solutions
> San Francisco
>
> ---------------------------(end of broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
> choose an index scan if your joining column's datatypes do not
> match