Re: Testable assertion tagging for W3C specifications

We already do some things to ensure consistency and correctness in our 
documents.

We use xmlspec, which gives us consistency in markup and formatting 
conventions. We use conventions for our markup that allow us to do queries 
that extract the queries from the use cases or the XQuery spec, so that we 
can run them through a parser or test them against an implementation (this 
does require some manual intervention). We generate a parser from the same 
markup that generates the productions for the grammar in our documents. We 
have a Formal Semantics document. All of these things result in real 
effort, but we have decided that the quality gains associated with them 
justify the effort.

Can someone do a mockup of how automatic tagging of assertions would be 
used, the level of correctness that could be guaranteed from this approach, 
the extent to which this constrains editors in describing features, etc? So 
far, I have not heard anything concrete enough to let me assess the 
cost/benefit. I don't know the cost, and I don't know the benefit.

Jonathan

Received on Monday, 6 May 2002 18:03:06 UTC