[phpGroupWare-developers] Planning ahead - was: Goodbye

[Sigurd Nes]

Maât wrote:
> Sigurd Nes wrote:
>>> From: Dan Kuykendall address@hidden
>>> Sent: 2009-10-06 08:22:41 CEST
>>> To: address@hidden
>>> Cc: phpGroupWare Developers address@hidden, phpGroupWare Users 
>>> address@hidden
>>> Subject: Re: [phpGroupWare coordinators] Goodbye
>>>
>>> Hi all,
>>>
>>> My preference at this point would be to shut down the project. It seems
>>> like the codebase is just too outdated. Maybe at somepoint it would be
>>> worth picking up and starting from scratch with all thats been learned
>>> both in phpGW and across all the various projects in the open source
>>> community these last 9+ years.
>>>
>>> Anyways, if a clear vision can be had and a developer to continue it is
>>> around, then please contact me... otherwise a leaderless project is a
>>> dead one.
>>>
>>> Dan
>>>     
>> Hi all,
>>
>> I have plans for it and want to keep it alive.
>>
>> I think the best approach will be to focus on the system as a general 
>> application development framework - starting with the API and some core 
>> modules (admin, setup, preferences).
>>   
> Every single framework does that... and many better than we do : Zend
> first of them but also Symfony, drupal, eZ...
> 
> If we plan to gon on their scope we are dead...
> 
>> I also think the majority of the existing applications without a minimum 
>> level of maintenance has to be put in a historical archive for future 
>> reference and a possible source of inspiration only.
>>   
> That makes sense but the svn system as i re-organized it just need to
> change the list of externals to let these module aside
> 
Great :)

>> Important features are:
>>  * user-handling
>>   
> Yes and there we'll need to put hard work
> 
> And part of this work will involve thinking about template system :
> coders are often very poor interface designers
> 
> And good interface designers have often very poor php and xml skills.
> Dreamweaver (sorry guys for the ugly word) and css and htmls are their
> worlds.
> 
> Relying on current xsltemplate even if it's sexy on the paper  will
> ensure  that zero descent web designer will be able to get in and help
> us make nice  looking user interfaces.
> 
> As far as web design is concerned the previous phplib based  template
> system was loads better.
> 

A lot of work has been put into making use of yui.
There is a (bit old) demo at http://beta.resight.no/login.php

>>  * integration capabilities (xmlrpc/soap/ldap...)
>>   
> Agreed
> 
>>  * building blocks for ui as super-objects prepared to utilize common 
>> elements (as tables, lists, calendars)
>>   
> Not agreed
> 
>>  * mechanism for internal integration across modules
>>   
> Agreed a million times
> 
>> I will fix the API (and core) to a usable state (running php 5.3) - and 
>> update to the latest 3-party libraries.
>>   
> Ok on the goal but if we go on we'll have to discuss the method : i
> don't wand to see a giant commit changing things everywhere whitout more
> thant "Merge from my working company tree"
> 
> Suvbersion is all about keeping track of changes and helping bug hunting
> by the means of changelog and commit date analysis
> 
> A million times okay for code fixing... but a million times not okay for
> giant commits impossible to check
> 
> You probably did not even consider such commit... in this case please
> accept my apologies for this part of my mail :)
> 

The work has been done over a long period of time - so it is not very easy
differentiate the fixes - but some splits should be doable.

One commit per app - some more for the API

- Think of it as a fix of the latest mega-commit.

>> I think that once the system is in a shape that makes is possible to install 
>> and operate - it will attract developers and users.
>>   
> indeed having something that installs and works would be a nice idea :)
> 
> but if you want to attract devs ans users that will not be enough
> 
> we'll need documentation (people willing to write it) and user + dev
> support (people willing to help people getting in... explaining things
> again and again)
> 
> And till we have enough manpower to make separate teams for support and
> developpement and doc writing and betatesting the remaining people will
> have to be everywhere
> 

Well - we have to start somewhere :)


Regards

Sigurd