Re: [xml-dev] DTDs Only Smell Funny

[cbullard@hiwaay.net]

@Michael:  True.  The first job is to expand them and have a reference  
copy for human consumption next to the formal copy for formal proofs.   
OTW, not terrible but pattern facets make it possible to catch the now  
familiar common mistakes; so this time I am happy to use XSD as this  
is the first application I've worked that makes heavy use of  
attributes for critical data.  The data module codes present  
interesting advantages and challenges in that a language that is  
self-descriptive in the way it handles resources to generate  
identifiers is actually quite good at getting around the "blind"  
alleyways so often discussed on this list.  Many problems are better  
solved with superior language design and smarter monkeys.   I debated  
in the beginning of the project if the specificity of 40051 and 24784  
were superior to the generality of S1000D and I'm coming to appreciate  
the latter as more requirements for compose-on-demand show up in the  
queue.

@henry:  Thanks.  That would be nice but I'm stuck with ArborText 5.3  
Windows 7 for the duration and am informed no upgrades are  
contemplated unless they win the next phase of the contract.   So the  
challenge is to work around the lack of a fully implemented X:Include  
and the fact that AT 5.3 dialogs freeze on Win7.  I trick the FOSI  
with the occurrence attributes and wrote an editor and set of  
utilities using system.xml to do the tasks I can't get done with AT  
because of the version/opsys incompatibilities.

Sometimes it helps to be one of the dinosaurs in this trade. :)

len