Seen this WACT-based CMF?

[tonyskyday]

I just wanted to point this out incase ya'll developers hadn't seen this yet, it's a content management framework based on WACT called LIMB.

http://limb-project.com/root

I figured you should know it's out there, and maybe see what kind of successes or challenges they are having with their framework.

I know from perusing their forums that they are looking at how Mambo does some things for inspiration, which is a compliment, I think.

-Tony

[pagerefiner]

I just wanted to point this out incase ya'll developers hadn't seen this yet, it's a content management framework based on WACT called LIMB.

http://limb-project.com/root

I figured you should know it's out there, and maybe see what kind of successes or challenges they are having with their framework.

I know from perusing their forums that they are looking at how Mambo does some things for inspiration, which is a compliment, I think.

-TonyLIMB only uses the WACT 0.1 template engine which changed considerably in 0.2. I believe that they used their own MVC design patterns for their modules. LIMB is only usable on a webserver running PHP5.

Another similar idea is the Bitflux CMS which uses popoon (a php port of Apache cocoon). This is also based on a MVC design pattern, but it uses a pipeline approach and a site map. In this case, all content is stored in XML format. The XML content is transformed on the fly using a pipeline to render the content in multiple formats (html, xhtml, svg, pdf, etc.). The current version of BitFlux is also for PHP5 only.

A third one is Kcomplete Lite from Inerakt which is a PHP4 only CMS based on the open source Krysalis library which is another port of Apache cocoon.

The first two are a bit slow at the moment, but should become significantly faster as the underlying source code libraries are improved and better caching is utilized. Komplete Lite appears to have decent caching, but I'm not too keen on the user interface, although it's fully templated and could be made Mambo like.

[pachanga]

LIMB only uses the WACT 0.1 template engine which changed considerably in 0.2. I believe that they used their own MVC design patterns for their modules. LIMB is only usable on a webserver running PHP5.


All above is true except the fact current stable version of LIMB is PHP4 fully compatible(actually it's written for PHP4 not PHP5). However we're actively developing the new version which will use WACT 0.2 and will be PHP5 only.

[pagerefiner]

All above is true except the fact current stable version of LIMB is PHP4 fully compatible(actually it's written for PHP4 not PHP5). However we're actively developing the new version which will use WACT 0.2 and will be PHP5 only.I was not aware of the PHP4 capability, and it's good to know for current and future reference. Thanks for clearing that up. It would be good to know if you plan to continue maintaining the PHP4 version until PHP4 reaches the end of its life cycle.

The other fact that was omitted and should have been stated is that LIMB is also a framework for building content management systems and other applications. There are probably a lot of third-party developers eyeing LIMB for future projects.

Nice to see you here.

[mmx]

All above is true except the fact current stable version of LIMB is PHP4 fully compatible(actually it's written for PHP4 not PHP5). However we're actively developing the new version which will use WACT 0.2 and will be PHP5 only.

pachanga... I've tested LIMB myself and it has a lot of potential. I was also surprised about the short duration of time it took you to get the first version out (lots of quality work in a very short period of time). Speed was a concern of mine, but some model and caching improvements might speed it up considerably. LIMB is still in its infancy, so I expect to see speed gains in future releases.

I would be interested to know if you foresee a speed increase after switching to selkirk's 0.2 template system. I experimented with 0.1 as a possible solution for component-level templates in Mambo some time back.

My only other concern is multisite capability. I've yet to see a CMS based on the MVC design pattern that can handle this need. Multisite support is not for every user but it is a consideration when developing corporate sites. I would be interested in hearing any feedback you might have about this.

I usually push Mambo, but often get customers who come over from the Java world, and they generally feel more at home with a MVC-based solution (struts, swing, cocoon, etc.). I don't think we have any qualms here about sending users elsewhere (other than this site) if their needs and wants can be addressed better by another CMS.

[pachanga]

I was not aware of the PHP4 capability, and it's good to know for current and future reference. Thanks for clearing that up. It would be good to know if you plan to continue maintaining the PHP4 version until PHP4 reaches the end of its life cycle.


Sure, we're going to support PHP4 for a while until the number of host providers using PHP5 increase. However maintaining several almost incompatible dev.branches is such a pain :(


The other fact that was omitted and should have been stated is that LIMB is also a framework for building content management systems and other applications. There are probably a lot of third-party developers eyeing LIMB for future projects.


You're absolutely right, we're positioning LIMB as a framework and really hope to see more third-party developers develop their own LIMB based solutions. Yet i have to admit that the current stable version of LIMB is not that flexible and modular as we want it to be....that's why we're hacking the new version ;)


Nice to see you here.

Thanx :)

[pachanga]

pachanga... I've tested LIMB myself and it has a lot of potential. I was also surprised about the short duration of time it took you to get the first version out (lots of quality work in a very short period of time). Speed was a concern of mine, but some model and caching improvements might speed it up considerably. LIMB is still in its infancy, so I expect to see speed gains in future releases.


You're right LIMB is not a speedster yet :) However there is a number of caching filters including full page cache filter which can be dynamically plugged into the LIMB based application and dramatically speed up the perfomance. By default it's switched off since it requires some intelligent tuning via config files.


I would be interested to know if you foresee a speed increase after switching to selkirk's 0.2 template system. I experimented with 0.1 as a possible solution for component-level templates in Mambo some time back.


Well Jeff's WACT 0.2 version is a LOT more complex than 0.1 and switching to 0.2 doesn't provide any speed gain. At the same time WACT 0.2 has a number of invaluable features that simply don't exit in 0.1. BTW, WACT has 2 template processing stages: compilation and execution. In 0.2 the compilation stage is much heavier than in 0.1 and requires way more cpu cycles but it's not that important since this stage is usually omitted(using cached compiled templates) on the production server.


My only other concern is multisite capability. I've yet to see a CMS based on the MVC design pattern that can handle this need. Multisite support is not for every user but it is a consideration when developing corporate sites. I would be interested in hearing any feedback you might have about this.


Currently LIMB doesn't support natively several sites at the same time yet this feature is on the TODO list. The only option is having a separate LIMB installation per site....

[mmx]

Sure, we're going to support PHP4 for a while until the number of host providers using PHP5 increase. However maintaining several almost incompatible dev.branches is such a pain :(

Hopefully, the time lag in switching from 4 to 5 will not be too great and the pain will be more of a short-term annoyance than anything else. As more PHP5-only applications like LIMB are released, more host providers will make the switch. The true benefits and recognition of switching to PHP5 and its better OO features are still off in the distance yet.

You're absolutely right, we're positioning LIMB as a framework and really hope to see more third-party developers develop their own LIMB based solutions. Yet i have to admit that the current stable version of LIMB is not that flexible and modular as we want it to be....that's why we're hacking the new version ;)
Thanx :)

Maybe a some work there,

An installer module for the framework would solve much of this. Since I knew this was possible, any temporary inflexibility was looked at as a bump in the road on my part. The use of MVC and other design patterns was the major attaction on my part. Even in its current state, some development documentation and tutorials are going to stir up more interest. Developing a Phing build tool for modules and other objects would certainly attract more developers. This will all come in time.

There is more inherent flexibility in using MVC and other design patterns than immediately meets the eye. Most PHP developers probably don't realize this yet and might not appreciate some of the benefits that LIMB offers at this point. The important thing to note about LIMB is that you are building for the future with your decision to develop a PHP5 trunk.

[pagerefiner]

Speed was a concern of mine, but some model and caching improvements might speed it up considerably. LIMB is still in its infancy, so I expect to see speed gains in future releases.
I think you're making some assumptions based on an earlier version of LIMB. Try the 2.3RC3 release or run the demo from the LIMB site. Your concerns about speed should disappear after doing this.

I would be interested to know if you foresee a speed increase after switching to selkirk's 0.2 template system. I experimented with 0.1 as a possible solution for component-level templates in Mambo some time back.

Does it really matter at this point? WACT and PHP5 are going to undergo some growing pains as work on the project proceeds. WACT 0.3 will get better, 0.4 will get even better and so on.

What you want to do on REO could just as easily be done in LIMB long before Mambo 6 arrives on the scene.

[pagerefiner]

Mmx... to clarify, you could could use Mambo 4.5.x/5.0 for PHP4 REO testing and LIMB for PHP5 testing. By the time Mambo 6 is available, you could have a PHP5 version of REO ready to implement on Mambo 6 and a LIMB implementation as well.

[pagerefiner]

You're right LIMB is not a speedster yet :) However there is a number of caching filters including full page cache filter which can be dynamically plugged into the LIMB based application and dramatically speed up the perfomance. By default it's switched off since it requires some intelligent tuning via config files.
Actually, I think it's fast in its current state even without caching. Using WACT .01 was a good choice even though it had to be modified to meet your objectives. I think the fundamental ideas behind WACT are a better approach to templating. I'm convinced of this based on some of my own experiments.

[mmx]

Mmx... to clarify, you could could use Mambo 4.5.x/5.0 for PHP4 REO testing and LIMB for PHP5 testing. By the time Mambo 6 is available, you could have a PHP5 version of REO ready to implement on Mambo 6 and a LIMB implementation as well.

Better to talk about this elsewhere.