Tim Bray has written a good post about XML and how it is not a great idea to begin writing your own XML standard in order to store and transfer your data. He makes a good point by referring to Metcalfe's law of network value and applying it to the realm of XML standards:
"The value of a markup language is proportional approximately to the square of the number of different software implementations that can process it. I could argue this from theory but would prefer to do so by example: HTML. RSS. PDF."
I think what he is saying is right, its very easy to write an XML standard that suits your needs but really, really difficult to get others to adopt it or develop special software to take advantage of it. The New Zealand government has a massive XML effort underway that seems to be bogged down in all the technical and practical issues outlined in the article. Tim suggest's reusing existing, well established standards ('the Big Five') such as XHTML, DocBook, ODF, UBL and Atom if at all possible. Given the flexibilty and power of all five standards this strategy is very wise. Its probably a lot more initial work to learn an existing standard and apply it to your use scenario but in the long term such a strategy pays off as your work gains industry acceptance simply because it is founded on something already in popular use.
As a follow-up to his first post Tim goes on to describe some very good points to consider if creating your own XML standard. The most interesting point he raises in this article is the importance at stopping at version 1.0 of a standard and not press on with an imcompatible version 2.0. Pushing development forward in such a way breaks backwards compatibility and is a sure-fire way of alienating any third-party development base you may have (or could potentially have). I think the fact that IFC's are now onto version 2x (which is in effect 3.0) and adoption whilst still growing at a very limited pace is a sign that third-party developers have been slow or hesistant to pick up the standard and work with it for fear of the standard changing.