[kdewebdev-site] Proposed PHP infrastructure

[Chris Hornbaker]

On Thu March 4 2004 17:41, jacob coby wrote:
> Most of my questions were already asked by Bill, but I
> still have a few:
> 
> - Is it too complex?  OSS doesn't work like commercial
> development; you can't rely on one person to stick
> around for months or years.  How long will it take to
> bring a person up to snuff with this system?  Is it so
> complex that it will put off people who want to help? 

Nope, not complex at all. Simple to follow and understand.

> - There was mention of emailing the translation files
> to the translators.  Is there a way to see what's in
> the queue to be translated, so anyone can come along
> and do the work?  We might need something simple
> translated into a (relatively) obscure language, and
> the translator might not be 100% reliable.  I may just
> need to have "Click Here" translated to Czech.  It may
> be faster to just use babelfish and do it myself
> instead of waiting a week for the guy in cz to make it
> to the internet cafe to translate the string.

The translators email the files to you. Just place a link to
the .pot files on the web site somewhere for them to download.

> - Someone mentioned that the main KDE site uses some
> system similar to this now?  Why are we re-inventing
> the wheel here?  The KDE site backend is sure to get
> more testing and more fixes/patches than whatever we
> do will.  The KDE backend is known to work.  Do we
> want to develop a website, or do we want to develop a
> whole development/deployment environment along with a
> website?  

This is not a re-invention of the wheel. It's an extension. It
does a lot more than simple translation. Also, I wanted to
make this Public Domain, so couldn't use their code. :-)

> - How well will this work with Quanta's VPL?  I don't
> think I understand how these XML files get converted
> to PHP without running some sort of script? 

It'd work OK as for as XHTML stuff. You'd have to hand
edit some stuff. The XML is converted with an XSLT and
perl script.

> 
> - In the example, why the need to define whether a
> link is a file or a dir?  If it ends in "/", it's a
> dir (per the URI RFC).  Is 'true' a dir or a file? 
> Things like that add complexity, and make it harder
> for a new dev to come up to speed.  Is it indictive of
> other things that need to be addressed?

Yes, I don't much like that either. In fact, I think I'll remove
it. It made sense at some point in time, though. :-D

> It does have some things that I like:
> 
> - Forces correct XHTML.  It's easy to fall back on old
> habits and include display with content.  Even in the
> example, he uses the <b> and <i> tags, which is a
> no-no.  With this system, you write content, and then
> add CSS later. 

I over used them too! I was testing the scripts and needed
a decent extreme example.

> - It seems pretty clean.  <i18n>string</i18n> is
> easier to understand and parse than
> <?=gettext("string")?>, and especially <?php print
> gettext("string"); ?>.

Yeah, plus you don't need to know PHP. :-)

> - I like that i18n support is built-in.
> 

If you like that, then you'll love what it does for usability
and accessibility! Not to mention the control you can have
over what you send the browser.

-- 
Christopher Hornbaker           
Jabber ID: Jilks at jabber.org     Email: chrishornbaker at earthlink.net
Join the Free State Project!    http://www.freestateproject.org
   "Liberty in Our Lifetime"
Was I helpful? Let me know! http://svcs.affero.net/rm.php?r=Jilks