I would agree, technically speaking, PEAR is not a framework. But this is what the PEAR people say on the home page of their web site:

‘PEAR is a framework and distribution system for reusable PHP components. You can find help using PEAR packages in the online manual and the FAQ.’

Anyway, that is an argument of semantics and nerd-details .. not the point of my article.

I also agree that RubyGems is cool.

All that said, I think that PHP is a number one web application development language when all things are considered. Yes, Ruby is a cleaner looking language and ColdFusion is easier to learn. But again, PHP overall has the most to offer IMHO.

Want to see good OO PHP projects? Check out:

Zend Framework
Wordpress
PHP Cake

… and there are more. There are also nice procedural based PHP projects:

- Drupal
- Punbb

So yes, PHP has a legacy to contend with. But the new breed of PHP projects and programmers are good.

Looking at it like RubyGems, CPAN or Python Eggs.

So i do not understand this discussion.

My first impression about PEAR is unfortunately that it is very broken and unprofessional (well like 95% of PHP). Documentation is bad and the installation on windows is in just one word insane - and there i don’t mean faulty but from a damaged brain.

I didn’t find a compelling reason to switch to Zend Framework. Infact, as an early consumer of Zend via the PEAR channel; then having it go away all of a sudden, I was quite a bit disgruntled.

I also found the CLA (http://framework.zend.com/framework_cla_1.0.pdf) a bit of an affront - why do I need this when I have BSD style licences? Why do I have to fax it in?
I know the answers to these questions, but I don’t agree with or like them.

If you campare it to zend_form you will find that zend_form is still in the very early stage

I think that when i start using ZF i will only use the MVC functionality (Zend DB, Zend_View, Zend_Controller)
and use pear for every thing else

PEAR has a lot of momentum, however I think a lot of the weaknesses come from a lack of understanding from web hosts. I know that most PHP developers started their development from the incredible wealth of comments and samples on the php.net site for each method. I love being able to do an “I’m Feeling Lucky” search with ’substr_replace php net’ and get over 100 different samples on how to apply substring replace plus derivatives using regex and a 2 year history of practical examples.

PEAR has none of this and I think it leaves a lot of developers a little wanting. It’s interesting to look at the history of the Zend Framework and how during the RC days, it had a similar model, where you would load in the Zend.php file and it was an aggregate root into the Zend Framework. It’s really nice to see that the Zend Framework architects moved away from this model. Regardless, I’d still like to see further decoupling for things like Zend_Layout and Zend_View (I know there’s a strong Smarty community out there). While Zend_Db is miles ahead of doing straight SQL through a $db object, I still would like to see Hibernate (a Java persistence layer) fully integrated into a Zend frontend, and sufficient community / documentation to support this kind of integration.

PEAR always seemed to me to be a one-size-fits-all solution, with projects gaining steam and then fading for months at a time (PEAR’s server-side form implementation comes to mind…)

j

“Ultimately, a few extra MB’s on a hard drive is not a big deal for most developers/sys admins.”

I would agree with that entirely. Disk space is cheap …

My liking the decoupled nature of the Zend Framework (a great design decision IMHO) has nothing to do with disk space, but rather with being locked into a bunch of stuff I don’t want to use.

Other frameworks I’ve used in the past, forced you to eat the whole meal … sometimes its’ cool to be able to grab what you want (and leaving the rest) at the buffet table.

Stef