Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View

I am planning in my company to move to php.MVC framework - to create  
several big web applications.
What are your opinions of this framework ?
Is it the best MVC framework for PHP ?
What are the pros and cons ?

Best regards,

Re: php.MVC

Quoted text here. Click to load it

I would really like to listen what people say on that as I am having
the same question.
Well I want to write applications based on MVC, because that enables
me to manage them easier and leave the presentation to the design

Re: php.MVC

avlee schrieb:
Quoted text here. Click to load it

When I researched frameworks I read about php.MVC too. To mee it looks  
like the framework is not very actively maintained and/or it has a very  
small community. Compare this to the "big players" like the Zend  
Framework or Symfony. Large communities are a big pro when it comes to  
getting problems fixed.

 From the MVC point of view I like Zend's RewriteRouter component  
because its multi-route concept makes the Controller component very  
flexible. Also, the Zend Framework does not force you to use anything  
more of it than you really want. It "scales" from being an ordinary  
class library to a framework, and it's not yet a final release!


Re: php.MVC


Quoted text here. Click to load it

I like Zend too very much - but is it stable ?
Is it worth moving to Zend now ?
Could it be used in big, professional solutions ?


Re: php.MVC

avlee schrieb:

Quoted text here. Click to load it
The preview releases are _stable_, as opposed to some nightly builds, I  
suppose. The problem might be that some of the APIs aren't _fixed_ yet.  
This means that you might have to adapt your code when upgrading to a  
newer version. For example, there was a methode named noRouteAction that  
was removed in v0.6. Some users came up with workarounds for v0.6 and  
these stopped working in v0.7. The changes you might have to make to  
your code are somewhat a little annoying but they only help you in  
getting to know the framework ;-)

Quoted text here. Click to load it
I'd say taking a look wouldn't hurt. I tend to give frameworks that pass  
my "bah, that's some ugly code" filter some hours of testing and Zend  
came out well.

Quoted text here. Click to load it
IBM is developing a good share of the code and they're using it for they  
pretty big QEDWiki project. So, yeah, should be possible. *g*


Re: php.MVC

Hi Michal,

IMHO the Zend framework is a lot of code for the little function it  
provides. It makes you depend on apache, who knows what server your next  
customer will be using?

With respect to MVC and flexibility phpPeanuts does the same (decouple  
the requesthandler from the request) and much more (provide flexible  
(reflective) default handlers for most basic application functions)  
(same concept as Ruby on Rails*).  It lets you override many details  
with the freedom of expression of real php code without modifying the  
framework code itself. It separates layout from user interface logic. It  
separates the user interface from business logic. It provides framework  
based persistency with polymorphism and navigational queries. And all  
that in only 808 KB (including the default user interface layout).

OK, learning how to handle so much function comes with a steep learning  
curve. But there are tutorials that show how you step by step extend and  
refine an application. Learning how to use it may eventually enable you  
to develop real life web appplications at a staggering speed of 17.9  
function points per hour** (most of the competition is still talking  
about hours per function point). But why be negative about learning? In  
the end the knowledge you get from it will prove a lot more valuable  
then any framework can ever be (at least for your fee as a software  

Is it stable? Some people may say its aproach is experimental, but in  
fact it is the third generation of successive frameworks i have  
developed since 1992 (no, the first generation was not in php ;-)). It  
is currently in a hybrid state between php4 and php5 (it runs on both  
with only a single 8Kb file that is platform specific). But software  
like people: When it's really stable it's probably dead. Of course there  
will be changes when it is migrated to pure php5. But i don't think  
migrating a substantial (but well-written) app will take more then a few  
days. But OK, if you are not willing to learn your code won't be much  
good, so it could take a lot longer. In that case forget about reuse,  
using third party code will not solve anything if the problems are in  
your head.


Henk Verhoeven,

* Both use the principle of 'default reasoning' that is founded in  
Artificial Intenlligence research. In the predecessor of phpPeanuts it  
was introduced in 1996 when the user interfacing layer was developed.

** Measured in the hours it would have taken to develop the same  
end-user function using OOP, an IDE, an object-to-relational mapping  
library, a visual GUI editor and implementing all events that need to be  
handled in Smalltalk code. (Coding itself is in Smalltalk slightly more  
productive then in php, i guess it's about the same as in Ruby)

Oliver Grätz wrote:
Quoted text here. Click to load it

Re: php.MVC

If your applications involve a lot of content - for example, websites,
wikis, e-commerce sites, intranets, discussion forums, event
registration, etc - I can recommend SilverStripe.

SilverStripe is a CMS as well as an MVC framework: we built a
framework that quite often runs in more a MV architecture than MVC -
the modell can act as a 'default controller' - and every database
object has its own URL.  This is why it's well suited to "content
rich" applications.

For a demo of the CMS, which can be pretty rapidly customised to
create administration systems for other kinds of applications, see

Site Timeline