Re: Merging PHPLIB into PEAR

> > > Let's look forward how we can add the PHPLib to PEAR to build an > > > "application framework" (in the sense as PHPLib is one) that makes PHP 4 > > > even more powerful. > > > > I don't think that this is necessarily a good idea. An application framework > > dictates a specific philosophy of application design on the people writing > > Stop. Don't build a simple CPAN copy. None of the "enterprise" users > needs lots of little, nice classes that do not adress key issues in the > webapplication development: db, session, perm, auth, xml, xslt, > template, some html widget classes. I agree with Ulf that the primary audience for PEAR will and should be web application developers. This would place a requirement on PEAR that it support the activies of web application developers. This will be difficult to do it the components are not integrated in some fashion. Once we start getting into integrating multiple components there is really no escaping the fact that you are implicitly endorsing an application framework. Importing PHPLIB into the PEAR tree will definitely promote a certain method of developing web applications. This would be a concern if the applicaiton frameword was deficient, however, I am not aware of any show stopping deficiencies other than that it hasn't supported native sessions well until recently. Also, the abstraction layer may need to be swapped for PEAR's. That said I do have my own issues with PHPLIB. I don't think dumping all application code into a single PHP directory is scalable or promotes independent development of the code base very well. I think this has been corrected in the development releases but there is no downloadable snapshot for me to verify that it is a workable partitioning. Also, the logic of some of the authentication is extremely difficult to follow so when you have problems with your authentication it is very difficult to figure out where the source of the problem is. Finally, the effort behind PHPLIB seemed to have died for awhile and I'm not sure why. I am thus wary of the offer to incorporate PHPLIB into PEARs codebase unless I can see what concrete steps the primary authors intend to take towards making this a reality. This offer could in fact impede development of an alternative application framework if we are all waiting with our hands in our pockets for PHPLIB to arrive on the scene. Regards, Paul