>Paul Grosso <[log in to unmask]> writes:
>>> Date: Sun, 7 Jan 96 03:37:19 EST
>>> From: "Daniel W. Connolly" <[log in to unmask]>
>>> I'm curious about testing: I don't see any sample documents. Have you
>>> done any testing?
>>> What I'd really like to do is get a random sampling of about 5000
>>> documents from OpenText and test them against the DTD, and catalog
>>> the errors we find.
>>> Hmmm... that would be a browser-DTD test. What I want is a random
>>> sampling of documents that were produced by authoring tools, i.e.
>>> stuff that folks expect to be able to read in and edit.
>>Something in what you say about testing is gnawing at the back of
>>my mind. It's not a big issue, but I think it's worth mentioning.
>>I think there are some unstated assumptions or biases in the idea of
>>"testing [documents that were produced by authoring tools] against the
>>DTD." In the case of robust 8879-compliant authoring tools that can
>>operate with any valid DTD (such as several SGML editors that have been
>>on the market for years now), almost by definition the documents they
>>author will test as valid against the DTD by which they were authored.
Sure. But until authoring tools are 8879-compliant, it's useful
to track the deviant idioms.
>>As such, any test wouldn't be a test of either the DTD or the documents,
>>but of the tool. If, in fact, you wanted to compare how Murray's
>>modular DTD would stack up against the results of such tools, all you
>>have to do is compare the DTD these tools used to author their documents
>>against Murray's. No need for test documents.
I want to see how the HTML language is evolving. What idioms
are folks using that are not captured by this new DTD?
In message <firstname.lastname@example.org>, Murray Altheim writes:
>I agree that we really need is a test of DTD equivalence, not document
>validation equivalence. It sounds like I need to include in my draft
>proposal a modularized version of the HTML 2.0 DTD in addition to the "HTML
>2.2 DTD" currently posted.
I'd be interested to see the output of Earl Hood's DTD diff tool
on those two DTDs.