DB ORMs

Hi guys,

After looking at wendy’s post I did a search in google and came across this link:

http://wiki.cc/php/Object-relational_mapping

It’s a list of DB ORM solutions that might be worth looking at.

I for one like EZPDO EZOQL (Object Query Language)

You know... here’s a thought we might consider.

Why don’t we put the full blown Propel Abstracted stuff over to the Tattoo site, and really focus it’s capabilities on more advanced usage and really take advantage of some of PHP 5’s capabilities. Sorta a high-end developer’s developer solution with a more "enterprise" angle. I think we’ll be in beta for at least 6 months, anyway.

We can continue the MODx track as is for PHP4, as I really feel there’s a lot of life left in it, and we can port things over as appropriate both ways.

On the MODx side, I’d like to see though is a pretty quick addition of the recursive parser after the first of the year, along with a slightly cleaned up manager (strip out many tables and the webfx stuff with consolidated stylesheets), cleaned up API and maybe even a "proper" event handling model that Victor has hinted at repeatedly. This would essentially continue to be the main distribution until such time as the Propel-powered version is faster and PHP5 becomes pervasive.

Rough thoughts, but what do you guys think?

Easy PDO sure does sound interesting as well due to the fact that it supports out of the box table prefixes, and sounds like it might be easier to get up and running with... but that’s about the extent of my ability to evaluate it! lol

From what I’m seeing is that most of the ORM guys are moving into PHP5. Jason has said that he will continue to port Propel’s php5 improvements to the php4 version.

I for one would like to say that ORM is something that we will need to move to in the future but I’m not sure it’s the right time as yet. More is being done everyday to improve on these systems but from what I’ve seen of EZPDO I think it’s also a cool ORM framework.

Jason is currently working on getting Propel into MODx ASAP but I’m not sure how long it will take the rest of us to get up to speed with the OO db stuff.

I think I would have to agree with Ryan on this. ORM seems to be the way to go for PHP5 versions of MODx. It’s much faster to get our heads wrapped around the current API along with the other changes that needs to be made to the system rather than having to restructure our minds for OO db.

Running both MODx MySQL (Public release) and MODx Propel (Private release) systems side-by-side will enable us to gradually move to Propel or some other ORM framework.

I don’t know a lot about the different versions types of ORM solutions, from what I’ve seen PDO is pretty nice but Propel offers a lot more (like an API) but I’d imagine that PDO is always going to faster than Creole. But it sounds like Propel may start to use PDO anyway.

Will Propel use PDO when that becomes available in PHP 5.1?

Perhaps. It really depends on how comprehensive PDO’s database support is, how much PDO helps performance, and how much of Creole functionality will have to move back to Propel or continue to exist next to PDO (e.g. I expect the metadata support to be an issue with PDO).

On the matter of branching a PHP5 version I will throw my two cents in. I love the idea of taking advantage of the new features in PHP5 and I think that is the way we should be moving. My one hesitation is that in maintaining two separate projects we are inevitably going to slow down our development pace. I don’t really mind that since I think it needs to happen. I’d imagine that by the end of this next year you won’t see too many people who don’t at least have the option of using PHP5.

But if we start work on a PHP 5 version I think that should be our focus and that the PHP4 releases should be only bugfixes and easily backportable funcionality from the PHP5 version. I’d hate to see us cut our progress in half by focusing equally on two projects. Internet techonology is moving way too quickly these days for that.

Adam,

currently I think there are more persons using php4 than there are persons using only php5. The PHP5 version’s main focus should be to just simply provide alternate DB abstraction via the ORM framework. There might be a few other tweaks here and there but that should not be the focus until we can say that we only require at least 5.1. Remember that the php5 version will not be public.

Here’s an atricle that compares three ORMs (Propel,EZPDO and DB DATA Object)
http://phparch.com/sample.php?mid=59

Here’s EZPDO’s response:
http://www.ezpdo.net/blog/?p=859.

I don’t see any reason not to have at least a semi-public PHP5 release. It would quite probably draw some more really good devs to the project and demonstrate better what the future has in store... It might be similar to the Charter Members program at Simple Machines in fact ... hmmmm....

That response sure does seem much more in line with the Propel figures, and now that it’s totally self-contained seems even more intriguiging. Jason, have you had a look at it by chance?

Quote from: xwisdom at Dec 22, 2005, 08:27 AM

currently I think there are more persons using php4 than there are persons using only php5.

True. Unfortunately I wasn’t able to get any good statistics on how many servers are running PHP5 but PHP5 has been recieved very well and seems to have proven itself stable so I’d imagine that almost all hosting companies will be offering it before the year is out. Unfortunately most people who are running their own servers don’t upgrade them until something forces them too but I think most of those people will be very willing to upgrade their version of PHP in order to run a better CMS.

Having a PHP5 version is great for PR too. So many CMSes are stuck in syncronous HTML, procedural-code, table-based layout, HTML 4, etc. That I think we really need to push that MODx is a cutting-edge CMS that utiziles OO code, XHTML, CSS, AJAX, etc. I really want us to be known as the cutting-edge Web 2.0 CMS.

BTW, PHP 5.1.x which is out now is pretty much between 40-100% faster than the PHP 4.x or PHP 5.0.x, too.