I’d certainly look at CodeIgniter or Symfony if you are rolling your own from BUT limits of rolling your own complete solution include issues with scalability (RoRs scalability issues are vastly overblown). The huge advantage of a FOSS app is that you can benefit from a network of support that you don’t have with your own solution.

My experience with many of the PHP Frameworks is that 80%+ members of the community are beginners just in it to solve their own problem and you can be left taking the advice of novices.

I’m a Marketing Team member of the modx CMS project and sure I’d suggest looking at modx–especially Revolution (currently in public alpha) but no matter what solution you choose you are going to need to weigh your workflow, extensibilty and importantly end user experience.

Thanks for the post.

Cheers,

Jay Gilmore

http://modxcms.com/

Cheers
/k

http://redotheweb.com/2008/09/19/designing-a-cms-architecture/

The problem is you can code up a simple cms, but it’s time consuming and not cheap (if you can’t do it yourself). Or you can use a Wordpress or Drupal and find yourself restricted by a steep learning curve and more functionality than you might need (both are great for content rich sites).

I created Clover Content to be a simple cms that you can plug in to any site, anywhere. It’s really easy to use too. It makes editable areas on a site “widget-like”.

Twitter has a very specific kind of engineering problem which others have gone into detail on elsewhere. It’s debatable how much of the problem can be laid at Rails itself, and how much is a result of poor design choices.

It’s unlikely you or I are going to hit a similar wall unless we’re lucky enough to have the next Twitter on our hands.

This is a very hard decision to make and it’s something I’ve flip-flopped on many times. Using an off the shelf package you get the benefit of being able to hit the ground running and take advantage of improvements as they’re added in by the community.

The big downside for me is that if, as you say, you want to heavily customise the package you’re usually in for a world of hurt.

I tend to favour the roll your own approach because of this. There’s an initial lag in terms of getting your own mini platform together, but you get the benefit of knowing it inside out and building just enough to solve your problems.

My experience is similar in that I do know a lot of PHP, but my roots are in design. I’ve often considered firing up CodeIgniter and building my own super-extensible, yet lightweight, CMS that I can build onto over time. However, I have yet to really walk down that path because ExpressionEngine does most of what I want each time I pull it out for a project.

My thought is “why re-write a CMS when I can just use what EE offers me and build a module for the rest?”. That said, eventually I’ll probably write my own someday, for the experience and for the customizability. In the meantime, I find it more profitable to do the stuff I’m really good at, and use EE to help me along my way.

@Kevin & @Dave
I’d definitely classify myself as a middle guy. I think whoever does the CSS/XHTML is very well-suited to do the templates in whatever CMS is used. And then I often hire a developer for any big customization or module development that’s beyond my reach, or that someone else can just do faster than me.