On Thu, Sep 06, 2007 at 12:56:39PM -0400, simo wrote:
 
> To be honest, I think PHP would be a horrible choice. It is barely ok

Here's the (nearly) 100% reason:

http://www.egroupware.org/
http://www.phpgroupware.org/
http://www.tutos.org/
http://www.open-xchange.com/
http://www.phprojekt.com/
http://info.tikiwiki.org/

     .... and many more. Many of them have tried hard to implement
Outlook integration. There is almost certain to be a lot of activity
from groupware developers exercising libmapi if we release php bindings.
And this is what libmapi needs most of all over the next few months. The
same isn't nearly so obviously true for Python.

I dislike PHP, but that isn't quite the point here. The question I want
to answer is: where can we get the largest number of developers
experimenting with libmapi as fast as possible?

> for Websites, but in term of using it for console or GUI apps, python is
> 100 times more used and generally better. (Btw PHP imo is more horrible
> then VB even, heck, it is the only language I know that is vulnerable at
> the language level (include/import/...)).

If really bad implementation stopped people using software, then you and
I would have met in different forums :-)

> In any case I think the OpenChange project should really concentrate on
> the core functionality, which is the client library and the server side.

Ah good, we are in strong agreement then. libmapi isn't being used
nearly enough yet. And I can't think of any way to increase that usage
other than to increase priority of PHP bindings over Python bindings.

> Don't get sidetracked on other stuff, if people is really interested in
> using libmapi, bindings will easily pop up, and "choosing" a scripting

You might be right, we will see. We're trying some developer relations
work right now (Asterisk, Bacula) where we are demonstrating the point
with shelling to commandline and hoping they will take on the job of
integrating with libmapi properly. Unfortunately the most important
users are the groupware application people so just going around talking
to all the other (*many*) applications that interact with OC in various
ways won't have as quick an effect.

We also have a licensing issue currently which limits the kinds of
projects that can use libmapi, which again favours the PHP groupware
developers.

> language would cut out all the others, which, at this point, I think it
> is a dumb move, in any case PHP is not really used for anything but the
> web, choosing it as the OC scripting language would gain you basically
> nothing but headaches.

Well now isn't the time to choose such a thing, but it is the 

> Also IIRC the PHP license is incompatible with the GPL so using PHP
> would be difficult and troublesome anyway.

I don't think so. SWiK is GPL, MySQL is GPL. I haven't gone into it
recently but I don't think this is a problem. 

> Let interested parties maintain bindings and applications. Samba history
> shows that embedding web stuff is not worth it unless there is strong
> motivation in maintaining it, otherwise it will just bit-rot.

Supplying bindings is not the same as supplying web apps :-)

There is a question of a simple language for non-programmer users to
specify client and server-side rules in, which I think could well be
PHP. This is very core functionality. 

-- 
Dan Shearer
[EMAIL PROTECTED]
_______________________________________________
devel mailing list
[email protected]
http://mailman.openchange.org/listinfo/devel

Reply via email to