Re: [PHP-GTK] Status of PHP-GTK

Bob Majdak Jr wrote:
> people actually willing to do the docs. And the coders can usually
> figure out what to do by knowing  GTK in C or by looking at the C GTK
> manual.

I have been stuck on a few problems I could eventually figure out with
the help of the C GTK manual. However, sharing the solution that seems
so obvious once you know it, is also a way of giving back to the community.

> The PHP doc system i am sure is not going anywhere, and personally I
> like comments off, it is a manual not a wbe forum.  That is one of the
> things I hate the most about the actual PHP docs, especally when a new
> guy posts a code sample that is horrid and the next 20 posts are people
> saying "oh hay i made this better". I am not being bitter. Honest.

I agree with you in that it is very annoying to scroll through tens or
hundreds of posts commenting other people's sample code. But there is
still kksou's website and it seems very successful. So if the comments
ever work, there could be a note on the "new comment" page that sample
codes don't belong here but at kksou.com instead. Besides, I don't think
there are as many newbies at work with PHP-Gtk as with PHP.

The only way I found to share my knowledge was either kksou.com or the
mailing list. When I offered to update the glade tutorial in the docs,
the resources I found and the hints on IRC helped me get to the manual
source code, but I couldn't even manage to build it. Nor get my head
around the whole structure. I do know C and DocBook but I ain't got all
day, so in my opinion, it is just way too complicated to get started
with the current docs system.

That said, there seems to be the need for a public place to share
PHP-Gtk revelations. Here are my thoughts on it:

1. IRC without logs is pretty useless to the outside world. I do
acknowledge and appreciate the quick help though. Could someone easily
provide IRC logs on the web? Wasn't there a bot regular on the channel?
It could also track the channel topic now that we can set it again.

2. The mailing list should definitely be informed of any planned changes
or discovered solutions IMHO. It is the usual place for anyone to start
looking. If 1. was implemented, a link to the IRC log date or topic
start would be a good start.

3. Updating the manual is hard to get started with. Someone with
experience could regularly add notes and hints with the conclusions of
current ML threads.

4. Making docs comments work seems to be non-trivial, I would dump them
in favor of 3.

Concerning the state of the manual: What would it take to fix the build
system? It worked at one time, but all I hear is that it needs to be
updated to DocBook V5. Whoever broke it should just fix the damn thing
so the old DocBook V3 sources can still be maintained. After all, it is
in a VCS where one can go back in time, right? I don't see a realistic
chance that anybody will step up and rewrite the whole manual just for
fun. Besides, there is still the problem with OOP support in DocBook I
presume?

I know I'm only saying what "should be done". If there was an easy way,
I'd gladly help with the documentation. Until then, I'll just try to
post whatever problems (and solutions) I come across to the ML and other
public places. Unfortunately I don't have the time to work on the big
problems, so this is my way to contribute.

All developers, keep up the nice work even if it's in the shadows to the
common subscriber :-)

Regards, Andre
-- 
Greetings...
From: Andre Colomb <[email protected]>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAknmXRYACgkQZ6c/lG/u2qzM7QCgmEZx6h3kfoRg4eQF+kcwLo1w
5JwAn2hvNtGfXyPN+N3Rfh9jLTT39JnK
=jZiP
-----END PGP SIGNATURE-----