Keep it Simple (Score 1)

mmm... IBM's Transcoder looks to me like a 'we do everything' pitch by using modules written by others, rather like Informix Datablades. Problem is - in the Open Source world things just get done more quickly these days and this is where most of the interesting XML development is happening, with the notable exception of companies such as Webmethods.

In the darwinian environment of the web, the real growth in XML has been using very simple schemas. Why conform to a huge spec. when a ten line DTD will generate an output that I can pull into my database with a five line Perl LWP script. Where is all the ICE syndicated traffic or XML-based EDI? It simply hasn't happened because it has been too complicated.

I've been following XML-EDI for a couple of years now since working on EDML and the tendancy to re-invent the wheel has slowed development enormously, and yet XML-EDI really should be the killer app. We could all forget having to fill in horrible paper forms without having to spend huge ammounts on Legacy EDI systems.

90% of all paper transactions are Invoices and Purchase Orders, wouldn't it be great if someone concentrated on this and developed the killer XML based system that allowed us to issue these electronically.