Data Management Blog: when alpha testing doesn't go well: a terrier's view


Data Management Blog
Data management, Cloud, Transformation and anything else...

when alpha testing doesn't go well: a terrier's view

Finally got a chance to work on a tool people have asked me for a number of times. It's a free tool that mirrors the successful SchemaLightener / SchemaFlattener for which I continue to get frequent requests and positive feedback. The tool is supposed to do for WSDL files what the SchemaFlattener does for Xml Schemas. Namely, to merge modularly developed files into a single, stand alone file. (Both modular WSDL files and the schemas within them.) One would think the two are so related that reuse would be high. They both use a similar "include" construct with paths. Even further, since there is less namespace/targetNamespace issues with WSDL, that it might even be easier to do.

In fact, I did reuse templates across these tools. But certainly did not find it easier to get working. The SchemaFlattener works great. But the WSDLFlattener was quite troublesome. This is due to the fact that one WSDL file can have multiple xsd:schema elements in its wsdl:types section. And I've seen some tools have a problem with multiple xsd:schema elements with the same targetNamespace. So I was determined to have only a single xsd:schema for each targetNamespace.

When I get a problem like this, I become a terrier with a bone. No matter how many times I make a mistake or follow a circular path and repeat mistakes, I can't seem to give it up. The problem comes alive and mocks me. "I shall not be defeated", I tell myself.

Well after seriously considering starting from scratch, I managed to simplify and narrow the problem. And soon after was able to solve it. I'll write more as testing continues. Why do I fret so over a problem in a tool I plan to give away for free, which has only a niche usage?

...on to beta testing ....
© Copyright Paul Kiel.