Zend Framework

[Luke Redpath]

Also have yet to see any big sites use it, apart from I presume 37signals suite of web applications, but...

http://www.sitepoint.com/forums/showthread.php?t=339329

Depends on how you define big though. Its just as capable as PHP of scaling to larger sites because it scales in exactly the same way, using share-nothing architecture. There is a large mortagage processing application written in RubyOnRails but the name escapes me (its in the Agile Rails book). We've also got some fairly large apps planned for a huge retail employment company in the UK.

Either way, I'm less concerned about how many "big sites" are using it, and more concerned about how productive and enjoyable development can be for me.

I have to agree on the Windows front though. I wouldn't ever recommend it as a platform for deploying Rails apps (though I wouldn't recommend it for deploying PHP apps either, not when you can deploy to a *nix platform) and its not the nicest development environment either - its not overly difficult to get up and running for development purposes though. Installing Ruby and RubyGems is a sinch using the one-click installer then installing Rails using gems is as easy as it is on *nix. I don't enjoy developing in Windows as much though, thats why I'm getting myself a MacBook Pro soon

[Luke Redpath]

Looked at the code thou, and the frontcontroller/routing appears (from my rudimentary knowledge of ruby) to first generate code for the routing and then executes it. So there is double "compliation" happening? much like using eval() or create_function() in PHP. Which doesn't seem good.

Not exactly what you mean, can you be more specific (perhaps start a thread in the Ruby forum before this becomes too off-topic)?

[Ezku]

http://www.37signals.com/?=PHPE9568F...9-00AA001ACF42

And: http://www.rubyonrails.org/?=PHPE956...9-00AA001ACF42

I don't see that as a problem.

[Luke Redpath]

And: http://www.rubyonrails.org/?=PHPE956...9-00AA001ACF42

I don't see that as a problem.

Am I missing something?

[Ren]

In rails-1.0.0\rails\vendor\rails\actionpack\lib\action_controller\routing.rb there alot of strings containing ruby being assembled.

I was only curious as to how rails did it, as I was/is routing class for PHP based on ruby documentation ( http://www.sitepoint.com/forums/show...0&postcount=27 ) and was convinced rails must've used the same method.. compiling to regexps, as it was fairly simple to implement.

[Ren]

And: http://www.rubyonrails.org/?=PHPE956...9-00AA001ACF42

I don't see that as a problem.

I wasn't being entirely serious posting it. I realise they might have valid reasons, other than RoR can't cope, for PHP installed on their webserver.

[Luke Redpath]

I wasn't being entirely serious posting it. I realise they might have valid reasons, other than RoR can't cope, for PHP installed on their webserver.

Well, being serious for a moment, their site is hardly anything more than a fairly static website for which Rails would be overkill. The wiki uses Rails though. And hey, they use trac which is written in Python. And the 37signals blog probably uses off the shelf blogging software (and lets not forget the blog predates Rails).

[McGruff]

I think this topic is going off the rails...

[33degrees]

Looked at the code thou, and the frontcontroller/routing appears (from my rudimentary knowledge of ruby) to first generate code for the routing and then executes it. So there is double "compliation" happening? much like using eval() or create_function() in PHP. Which doesn't seem good.

Sorry to contribute to the derailing of the thread (heh), but in "production" mode, rails keeps objects in memory between request, so all the dynamic code creation only happens the first time and stays in memory, rather than being rerun for each request. In the case of the routing, it was rewritten to use code generation because it turned out to be much much faster in use than the previous implementation.

Going back on topic, I'm going to refrain from commenting on the code itself until I get a better chance to go through it, but I will say I applaud zend's decision to release what they had sooner rather than later.

[Clenard]

I wish they would have made it a little more well known when they were releasing it... It's about time though!

[shea]

I mentioned above that we let our framework speak for itself, here is an example.

USWeb Blog entry

Love the irony .

[DerelictMan]

Love the irony .

Yeah, no offense intended to anyone involved, but I chuckled at that too.

[thr]

Yeah, no offense intended to anyone involved, but I chuckled at that too.

[darsh]

I think this framework is sorely needed, and it seems to be very well thought-out. I look forward to digging around and having a decent look over the next few days.

I am quite disappointed that this has become a Rails thread. Much like evangelist Christians, Rails fanboys just cannot seem to enjoy their framework without trying to recruit everyone else. And I find the fact that both 37signals.com and rubyonrails.org are running PHP is classic. It's not unlike php.net running on IIS/ASP. Enough of that though...

I would like to particularly commend Chris Shiflett for being so vocal and open with the community up to this point. I certainly appreciate all the hard work you guys are putting into this. When can we expect the next release Chris?

[spinmaster]

PHP-Architect just published a tutorial on using the ZF...

[Luke Redpath]

And I find the fact that both 37signals.com and rubyonrails.org are running PHP is classic.

Yes, the fact that they use the appropriate tools for the job is "classic".

[mx2k]

I think this framework is sorely needed, and it seems to be very well thought-out. I look forward to digging around and having a decent look over the next few days.

I am quite disappointed that this has become a Rails thread. Much like evangelist Christians, Rails fanboys just cannot seem to enjoy their framework without trying to recruit everyone else. And I find the fact that both 37signals.com and rubyonrails.org are running PHP is classic. It's not unlike php.net running on IIS/ASP. Enough of that though...

I would like to particularly commend Chris Shiflett for being so vocal and open with the community up to this point. I certainly appreciate all the hard work you guys are putting into this. When can we expect the next release Chris?

hey nix the religious allegories(comparisons), thats just stirs up more trouble than its worth.

what do you all think of the data adapter and HttpClient classes?

[thr]

[...]Rails fanboys just cannot seem to enjoy their framework without trying to recruit everyone else.[...]

Could not have said it better myself, most of them seem to think RoR is the best thing since sliced bread.

[skinny monkey]

I would like to particularly commend Chris Shiflett for being so vocal and open with the community up to this point. I certainly appreciate all the hard work you guys are putting into this. When can we expect the next release Chris?

i second that. Nice to see genuine input from the community reaching the source, and it is a great indication of the direction this framework could take.

[Ren]

http://framework.zend.com/

0.1.2 has just been released..

[siteartwork]

According to the project maintainers, the ZendFramework will be based upon PHP5.0* as the earliest version to run with. A bad decision in my eyes, since the SPL came up with nice features beginning from 5.1* on, and the "fixed bug"-list from PHP5.04 to 5.1 is quite long.
Heck, I can't even think of a big enough userbase running PHP5.0* and using ZF making this decision legitimate.

[Ren]

According to the project maintainers, the ZendFramework will be based upon PHP5.0* as the earliest version to run with. A bad decision in my eyes, since the SPL came up with nice features beginning from 5.1* on, and the "fixed bug"-list from PHP5.04 to 5.1 is quite long.
Heck, I can't even think of a big enough userbase running PHP5.0* and using ZF making this decision legitimate.

I thought 5.0 was is still being maintained with respect to bug fixes.

Don't see the 5.0 and 5.1 decision as being a major problem, if want to replace some part of ZF with something 5.1 specific, then you are free too.

I have noticed in using ZSearch that they do seem to allow bending of some rules, there are optional dependancies on pecl/bitset extension (thou looking at the pecl page for it, its maintained by Alexander Veremyev of Zend so), presumably to increase performance, and decrease memory usage.

[Dr Livingston]

> Don't see the 5.0 and 5.1 decision as being a major problem...

I think it's sensible that we support pre 5.1 but remember that at some point there may well be a BC issue, much like the BC wars with PHP4.x and PHP5.x;

There were many fatalities...

[ironikart]

Wow, the example in the php architect article is already out of date.. these guys work fast. There's an altered naming convention for the methods inheriting Zend_Controller_Action. Had a play with it in version 0.1.1, created some actions. Upgraded to 0.1.2 and bang, none of my actions work. Can't wait to see this settle into a stable release, I'm loving it so far.

[pixelsoul]

Looks interesting

Can anybody explain what is the advantage with:

// $select is now a Zend_Db_Select_PdoMysql object

// SELECT *
//     FROM round_table
//     WHERE noble_title = "Sir"
//     ORDER BY first_name
//     LIMIT 10 OFFSET 20
//

// you can use an iterative style...
$select->from('round_table', '*');
$select->where('noble_title = ?', 'Sir');
$select->order('first_name');
$select->limit(10,20);

// ...or a "fluent" style:
$select->from('round_table', '*')
       ->where('noble_title = ?', 'Sir')
       ->order('first_name')
       ->limit(10,20); 

I can't see any benefits of using a query like this?