de tomKronieken

Peter talks about his RRD impressions.

I've had a look at RRD too, a couple of months ago. My conclusion (and that of my collegues too) was that it was indeed very nice for small projects with not-too-experienced developers. But I don't believe at all that it will scale. If the business logic becomes too complex, the RRD tool won't do the job for you.

However, this is not a fault of RRD alone. Every RAD tool that I've seen so far falls in this pitfall: they start making assumptions about the code you're going to write with it, and start building fences to prevent you from jumping into the code to build things they didn't think of. While this helps enormously for the beginning programmer, it gets in the way once your project starts to mature.

One way I would consider to use RAD tools, is to build some rough chicken wire framework, on which you can then add the real logic after throwing the tool away. But this would reduce it to an (advanced) GUI builder...

But I'm not able to evaluate these things. I think Peter, as a relatively inexperienced developer (AFAIK), is in a much better position to do this. So please, Peter: get your hands dirty with the thing (IBM will give you an evaluation version if you act interested enough), and let us know what you feel.

Later: Oeps, it turns out I underestimated yet another person. Peter has 8 years experience on C/C++ development, but has always been active as an analyst as long as I know him. So let's change the offending sentence to ... as a relatively inexperienced Java developer.... I'm glad I put AFAIK there.