XML Schemas vs. DTDs and Other Issues

John Gotze in Gotzeblogged writes:    

The most interesting document on this I've read for a long while (and I've been reading a lot) is Phil Windley's Enabling Web Services. I must follow up more in details, because there are lots of good points, but also a few places where I disagree: DTDs? No, use XML Schemas, I'd say. WSIL? Hmmm. Maybe, but we (government) need to engage in UDDI too.   

I think using XML Schema instead of DTDs is probably the right choice.  I'll update the paper.  In particular the XML Schema language gives you the power of a context sensative grammar rather than a context free grammar (at least for types) with little increase in complexity.  They also probably have a brighter future.  The main point is, however: document what you create and keep it up to date.

I think UDDI is premature except inside the orgranization, so I stick by my recommendation to use WSIL.  WSIL can be easily integrated with UDDI later when (if?) it takes off.  He then says:

In policy-making terms, however, Phil and other RESTians have a particular and peculiar problem: How do you explain what it's all about in political, non-technical words? I'm a techie, and I hardly understand it. My collegues (and bosses) are political scientists or whatever, and simply don't get it at all.

Boy, isn't that the truth!  I worked very hard on the paper because I knew this was an issue going through multiple major revisions.  Still, I'm aware that it is pretty condensed and is really more readable by a technical audience than a general audience.    On thought is to expand each principle into its own mini-paper with more room to explain, more examples, etc.  I'm open to suggestions.

Please leave comments using the Hypothes.is sidebar.