Re: Proposal for building knowledgebase website. - Mailing list pgsql-www

Hi all,

 > You are still the only actual volunteer at the moment :-)
> Of course, you are free to bail out if you don't agree with the
> direction things are taking, but we'd obviously prefer it if you didn't!

Indeed, things have changed direction. This is of course good. It means we
also are evolving around the KB idea. For me personally, it goes a bit fast
because I am not familiar with the current framework set aside a new CMS and
integrating it into the current system. Therefore it isn't clear to me how I
can be of help.

I think it would be wise if we just installed this thing first and play with
it for a couple of weeks. Create accounts in it and just use it as if it was
in production. Like a proof of concept as suggested in latest posts. Just
then we could really say if the CMS is working for us.

To my experience, CMS systems make you design and develop your information
system in a pre-decided and pre-dictated fashion and most of the time you
have to hack your way around to get it look and feel and work the way you
really want, resulting something far off from the original requirement.
This is where I like to remind we should be aware this out-of-the-box CMS
isn't overkill.

If you think about the real functionality we require, you end up with the
following.

The kb:

- Must integrate into the current system. I don't think we want to change
the current system drastically to fit any CMS. After all from what I saw in
the sources, there is been a lots and lots of work done to get here.

- Easy to navigate: you must be able to see the picture without digging page
to page to get any information. Have you ever checked this:

http://w4teclipse.com:8765/help/index.jsp?topic=/com.w4t.eclipse.developer.h
elp/html/reference/api/com/w4t/dhtml/package-summary.html

or http://msdn.microsoft.com/library/
or http://www.cs.washington.edu/dm/vfml/

- Easy markup. We could plug in something like "kupu"
(http://kupu.oscom.org/screenshots/) and filter the edited page on some
unwanted tags.

- People must be able to easily contribute: we can adopt the
www.codeproject.com idea. People create account, they contribute, if the
contribution ranks high or it was accepted (it was good to keep), then the
contribution gets a stable status and it is accepted in the permanent kb.

- must be manageable in the future. And not too much hassle regarding
administration, mirroring and upgrades.

I mean think about this. You can never know what a person knows about
PostgreSQL. (except for the ones that we know by communicating during past
years) Is he a database user, a pg programmer, a db administrator of just
someone who doesn't like pg and try to fill the kb with crappy information.
How are we going to address these issues with of without an out-of-the-box
CMS.  And of course you wouldn't want our dear future contributors write
wonderful articles about how you do a SELECT * FROM.

I just hate to see this becoming an uncontrollable monster eventually
disinterest people using it. I am sure the current techdocs that is almost
dead, was setup with good intentions.

Am I too far off the subject now? Or this is something catching your
attention also.

Regards,
Gevik.