General developer forum

It sounds like you want to re-invent the wheel for your hovercraft.

You can do MVC in Moodle today. We have views, but we call them renderers.

And you should already be structuring your code around classes (models) to do all the real work. Until now you would typically have defined the classes in locallib.php, but in 2.6 you will use class auto-loading and put them in the classes folder in your plugin.

That just leaves the controllers, which in the Moodle world would be the scripts like view.php. That script should process the incoming data, make the appropriate calls to instantiate and possibly update the relevant model classes, and then call an appropriate renderer method (view) to generate the output.

The student-visible parts of the quiz already do a reasonably good job of following this pattern.

So, to use MVC in Moodle, you don't need some hypothetical future framework. You just need to write code that does not suck today.

Personally I think the whole MVC thing is overrated - but almost certainly better than some sort of random fumbling. But let's put that aside. 

In it's own way - Moodle is already a framework for developing plugins albeit not (quite) in the general purpose way  that the likes of Zend Framework and Symfony are. In some respects I think it's better - or at least more pragmatic. 

I guess my point - if I have a point - that I was very taken aback that someone would go to all the trouble of plugging in an entire MVC framework into a local plugin just to develop a single block. A little perverse perhaps?

Although, having said that, I doubt it was the original motivation. 

Your point that Moodle plugins are written in a "non standard" way is a little depressing to me. I assume that you are asserting that MVC is some sort of "standard". Well, I suppose it is but not in a way I find terribly useful. Moodle plugins are now very standard - within Moodle - and they work. 

Yes - that's what I expected. I came across it in the context of being asked to run one fairly simple block. In that context it starts to looks a bit bizarre, but I really didn't think that was the original idea.

The McGraw Hill plugin if you are interested. The upshot is, I suspect, that we won't be able to use it due to the resource required to evaluate the MVC framework that underpins it. 

re-inventing the wheel for each plugin makes no sense and also using a framework cuts down on maintenance and standardizes the plugin so another developer can contribute more easily

I have a very similar experience, each plugin contains a lot of of boilerplate code that could be extracted and managed under some kind of common framework. I would create a framework like this but I gave up on the idea of doing it by myself - I think it's pointless if not agreed by wider community, especially Moodle HQ. This is the reason that I've instead implemented generating templates of the code in  moosh.

If I were to do it all over again today, I would try to make use of Symfony2 components.

I would do the same, Drupal 8 made similar decision.

cheers,
Tomek