ZikeBase and Rapid Development [0107.2000]

I feel like I'm trying to build a factory out of junkyard parts with no tools but a stick, a rock, and a little bit of string. But ugly, kludgy, rigged as it is, that factory may just start sending products out the door. I call it ZikeBase.

The idea has always been to build a RAD tool along the lines of Microsoft Access, but for building websites. Zebra was part of it all along. The name even came from ZBRE - ZikeBase Reporting Engine. Somehow I got the idea I could put the whole tool inside Zebra. Well, I can't. Turns out I still need a PHP3 framework. And guess what? I have almost every piece I need lying around already. Part of it's in the shopping cart. Part of it I built for powerreporting.com, part of it's at manifestation.com, part of it's in linkwatcher. I wrote a big chunk of what I want in ASP for abel solutions, and spent most of today rewriting (rescribbling, really-I cut and paste like the devil) it in PHP. I still need to weld most of it together.

Boy is it a mess. But it just might work as a tool for building websites.

They say "plan to throw one away, because you will anyway." Tom Peters tells managers to get a prototype out the door as soon as they can. Mozilla talks about dogfood - using the code you're working on ("eating your own dogfood"). I have to admit I like the model. I'm solving problems the messy way just to get something working.

But the messy way isn't the right way, and I'm in danger of leaving things as is. Like Zebra. I realize now it needs to use a DOM-based XML parser, not an event based one. But I don't have time to hunt down nonexistant documentation or write my own parser... so for now, I'll just do things the wrong way.

Part of the problem is that many of my "reusable" objects aren't as flexible as I'd like them to be.. Not easy to subclass and what have you. So I wind up cutting and pasting, or breaking old code, because I never seem to have time to do it right.

I need to start reading about good design patterns. I've got the coding part down. I can hack out any algorithm you throw my way, even in languages I've never heard of, if you give me documentation. But designing an application.. That's a bit harder for me. Oh, sure, I know data-driven web apps inside out.. At least the framework for em. That's how I dreamed up Zebra to begin with.. But there's a lot I don't yet understand.

I guess I'm rambling here. Well, great. First rant of the new year! :)