Re: [phplib] [RFC] Future of phplib From: Bertrand Mansion (bmansion <email protected>)
Date: 07/21/01

Hi,

I don't agree with Nathan on the point that PHPLIB should move to
SourceForge. I don't see any benefits with that. According to me, the
problem is elsewhere.

But I agree that a lot of things are happening in PEAR and on the other hand
there has not been anything new with PHPLIB for a long time. Something has
to happen soon or PHPLIB will be forgotten.

I am using both PHPLIB and PEAR at work. I can tell you that developing an
application with PHPLIB DB wrapper is much faster. Even the library is
faster in execution time (according to my own benchmarks). And at the end,
the code is easier to read, easier to maintain.

PHPLIB DB lacks many features and perhaps some consistency. PEAR DB is
trying to be the best wrapper ever. It can handle almost anything (in
emulation mode at least). The price of this is slower execution times and
more code.

What do people need ?

Both : a lightweight and fast solution for lighter applications (PHPLIB). An
heavyweight solution for big apps on big clustered servers (PEAR).

In my opinion, now PHPLIB objectives should be : do things like PEAR but
faster and easier !

Let's start with that and add some consistency. Let's define a core part
where attention should focus. I think that we should concentrate on :

- DB
- Sessions
- Auth
- Page

And forget about the rest for now. OOForms is not needed and many other
solutions are available (I use my own and it is a lot more powerful).
Templates are the same (there is Smarty, IT, PHPlib templates and so on).
Not to say about Table.inc, query_sql.inc, mail.inc, setup.inc,
tree_sql.inc, strings2.inc... These should be deleted, it is messy and
confusing.

At the same time, directories should be created to reflect what is their
purpose, just like it is done in PEAR. Attempts have been done for that but
it is still too confusing.

When the core is ready, the rest will follow naturally.

The only reason why PHPLIB should be moved to somewhere else is if there are
barriers against developers initiatives. I don't see any for now.

Let's talk together about a new and strong structure for the core of PHPLIB.

See you

Bertrand Mansion
Mamasam 

-- 
Abbestellen mit Mail an:   phplib-unsubscribe <email protected>
Kommandoliste mit Mail an: phplib-help <email protected>