Re: [Phpgroupware-developers] phpGW NetxtGen and further developm ent path

[Dan Kuykendall]

Alejandro Borges wrote:
> Christian Boettger - Mailings wrote:
>
> > Hi,
> >
> > <snip>
> >  
> >
> > Would be even easier if the IPC stuff (ipc manager + the (in this case
> > simple) IPC interface classes for each app) goes into nextgen as well.
> >  
> Well... nextgen is XML aware all over the place, and as far as i know, a 
> standard argument set and functions at the business logic level  will be 
> enforced even for just coding an app in it. So, the IPC should be in 
> there BY DESIGN! (which is the right way to do it anyhow).

This is correct. The more I have learned about the IPC, is that its 
basicly a wrapper around what the apps already have, to make them have 
standardized functions and params. The nextgen code basicly does this 
without even thinking about it.
Of course we will need to make sure that things stay consistant, but it 
has always been the intention behind the design.

> > but first, we must have the apps in bextgen to migrate to; and those are
> > missing apart from the core apps... and not many head/.16 apps do 
> > currently
> > have IPC interface classes.  
> All 16 apps that want to migrate to head should have IPC, and really, 
> head should implement IPC as a core feature of every app.
> Here is the thing. The greatest value of the IPC is that its a standard 
> api for add/edit/delete for generic data. We need that in head, in 16 
> and in nextgen as a concept (if not the exact same code), because its 
> what allows for really extensible and reusable api and applications.

To address the IPC migration option.
I really do not think this is the way to go. The IPC is good, and 
useful, but it will be alot easier to just take the current phpGW that 
is out, for example 0.9.16/0.9.18/1.0, and write some table migration 
scripts. We will make it simple by supporting only the latest version.

If a user is a few versions back, they will have to upgrade using the 
current version first, and then do the migration.
I do not think this will be all that hard.

To address the porting of apps issue.
Not many apps will simply "port" over as in be dropped in, tweak and 
make work. The apps should all go thru a phase of being re-thought out. 
It will be a time to improve the apps, and in many cases re-write them 
using all that has been learned. We will of course be able to copy over 
a great deal of the logic involved in each app, and that will be the 
reason that it will go much faster than a true re-write.

Dan