>None of the projects I've been have decided to keep a third party framework because of the tremendous overhead. One thing you'll always hear from clients is that they do things just a little bit differently from the rest of the civilized world. So in my case, I must either be pretty good at finding jobs with the exceptions or the exception seems to be the rule.


I guess I don't understand that. There is almost always a common set of features everyone wants, regardless of how specialized/different they do things from the rest of the work. Of course they do things differently, that's why they're paying you to write custom software. But there is still those core features a framework can give you so that you can spend more time on those specialized functions.

I'm sure it's safe to say most customers want/need: Security, Field Validation/Must key fields, flexible backend changes (even if they don't realized it yet), searching capabilities, and Error control & logging among others. That's what most frameworks give you; a stable starting point to build your application on top of. For a couple of hundred dollars, I've saved myself 6+ mos. of application development time. It's like adding another full time programmer to our staff who's only job is to improve the framework. That's a pretty good payback.

The framework you pick is based on what starting point you think you'd like and/or need, and your style of programming. Some frameworks are more than that, they also include a lot of RAD wizards/builders. The question of overhead comes down to which features your think your customer will need (and ultimately which framework you pick). Most of the available frameworks out there are pretty decent, and their authors have spent time trying to optimize them as much as possible. My question is this, if you actually need most of the functionality a framework will give you (and I would say yes, you do) then what makes you think your code will have any less overhead? That same functionality is needed, it still needs to be coded, and will require the same overhead as the frameworks. Why waste the time when someone has already done all the grunt work for you?

One of the other advantages to a framework is also hidden in that overhead: future flexibility. Customers change their mind about what they want or need. I think another important aspect of using a 3rd party framework is that they have thought about creating a very flexible base on which to build. I'm sure everyone here has written an application and then, knowing what they know now, could write the same application even better. A 3rd party framework can give you some of that insight, even if you don't realize it at the time. They have built-in hooks in which to place your own code to modify the application's behavior. Your own code may not have a good place where those hooks can exist, just because you didn't realize you needed them at the time.

Which framework you use is really dependent on your programming style. I personally feel a great majority of developers would be better off using a 3rd party product as opposed to trying to build their own. Yes, if you truly have the time to spend building your own, you may be better off. But honestly, how many of us really have that kind of time? Not to mention, have the time to continue to improve on the framework.