Comments on: Here be (XML) dragons

By: William Vambenepe — Anthology of blog posts about protocols and data formats

Fri, 12 Nov 2010 20:21:46 +0000

[...] While I am at it, I should add to the list this to this bird-eye view of all the XML obstacles that pedestrians run into (I have highlighted this entry in a previous post). [...]

By: William Vambenepe

William Vambenepe — Mon, 22 Sep 2008 22:21:27 +0000

Thanks for the pointer to the new paper Erik. I like it (and agree with its content) but I don’t see it as as major a contribution as your “XML Fever” paper. In the SOAP world, the fallacy of assuming that an XML serialization magically gives you a robust and widely re-usable API (a.k.a the “loose coupling” buzzword) has been pretty widely debunked. Especially by the Alpine paper in the case where the system model is based on the Java object model. Of course the fact that it has been debunked doesn’t mean that the lesson has actually been learned…

What I think your paper does, in this context, is illustrate and explain this phenomenon in a more general way (rather than focusing on the shortcomings of java2wsdl). This is useful and needed.

By: Erik Wilde

Erik Wilde — Mon, 22 Sep 2008 21:20:16 +0000

Thanks a lot for the positive review of our article! And since you seemed to be missing one thing in our discussion of XML, namely the fact that XML might not always be an appropriate representation for data models, we have quickly written a follow-up article, “Document Design Matters”, which will be published in the October edition of CACM.

Just kidding about the just-in-time-writing, of course, but we did write that article and it will be published in October and if you don’t want to wait, a preprint version is available at http://dret.net/netdret/docs/wilde-cacm2008-document-design-matters in the same old-school vi-edited HTML/CSS style. Hope you like this one as well!