PHP framework: do we need it?

[dakke]

This question is mainly to get advice from those who have experience using a framework. Any thoughts are more than welcome.

 

We are a small dev team, and wonder whether we need a PHP framework in the first place? Lots of things to be found on the net, and quite a few opinions on it.

Not that I'm after a say detailed analysis of our needs and the usefulness of a framework for our particular case. More of a general opinion on what you could say are 'standards' in the industry.

 

So in short: we're an experimental dev team working on a open source project. It's experimental in the sense that's it's going to take quite a while before we get a good output. It's the beginning of a scientific tool, where a proof of concept is the initial step. Proof of concept is more important than a full featured tool.

Once the proof of concept is achieved, we will extend the tool further. It will have more features, become increasingly complex. The 'final' version will be, beyond any doubt, very complex.

This also means that once the proof of concept is achieved and we've developed some useful addons, we will be looking at integrating this with other experimental workbenches. Meaning: after a initial proof of concept, some addons, integration is the next step.

And this is where my question comes in:

What is the best strategy given the above (very general) roadmap: "Would you recommend a framework for this sort of project or not?". We currently do not use any of them, so a yes would mean rewriting them using cakephp or zend or Symfony...

What is most likely to happen (I guess):

Once we have the proof of concept and added some nice features, the integration with others will most likely mean to rewrite the code anyhow. So whether you use your sort of idiosyncratic method initially does not really matter. Cause being experimental probably means you're making errors in the first place. And wouldn't an integration essentially mean to redo the work loaded with the experience of the past few months?

 

Or rewriting the questions:

1. Does a proof of concept require (is better off with) a framework or not?

2. Do addons require a framework?

3. Does an integration require a framework?

 

[trq]

1: IMO Yes. Well, at least it would benefit from the use of a framework. Why? because a framework (as well as providing allot of functionality) provides standards. eg; If I write an application using Zend, anyone with experience with Zend should be able to easily understand my application.

 

This could be of huge benefit working within a team.

 

2: See above.

 

3: See above.

[dakke]

Thanks for the reply.

 

But what about when another framework becomes the next best thing? And doesn't using a framework also requires others to know the framework and as such is selective towards other developers? Using Zend filters out those non-Zend people, would it not?

[trq]

But what about when another framework becomes the next best thing?

 

Theres no such thing. A framework is good because it keeps to standard. Swapping and changing frameworks defeats the purpose.

 

Using Zend filters out those non-Zend people, would it not?

 

Not really. While there is some learning curve from one framework to another, its not huge. The benefits of sticking to one standard or framework far outway the costs of having all your developers simply doing there own thing.