- From: Andrew Thackrah <andrew@opengroup.org>
- Date: Wed, 30 Apr 2003 00:26:40 +0100 (BST)
- To: <www-qa@w3.org>
On Tue, 29 Apr 2003, Mark Skall wrote: > > You're giving me a headache. My responses are in-line. > > > Correct. But it can (through a conforming extension) talk about it > specifically, rather than in general terms. Also, I'm not sure why someone > would add this new DOV. This goes back to my earlier reply to David. The comment worries me in a "640 KB should be enough for anyone" kind of way :-) I appreciate that there are subtleties of concern over DoVs being regarded as a kind of extension and extensions being regarded as a kind of DoV but I think that is going off the track a bit from my main argument which is that in the case where a spec author has given due consideration to SpecGL but decided that in fact they will use a new kind of DoV thankyouverymuch then SpecGL can either 1. be silent, since we can't second guess them or 2. Attempt to limit the potential damage of their rash decision ...by suggesting that they a) (Explicitly) identify the new DoV in the spec b) explain its relation to other DoV c) explain conformance consequences Obviously we can't go into any more detail in the checkpoints like we do for p/m/l but the question is - is it worth asking them to do (2)? (E.g. Is it so unlikely to happen that to bulk out SpecGL with more checkpoints is being over cautious, or is there likely to be enough valid deviance from the canonical DoVs in future specs to make it worth while?) I think that since we are asking them to do it for predefined DoV then we could at least ask them to think about it for their own hare-brained schemes. And if they do decide to treat their new DoV as an extension then I think we get this for free anyway, but if they don't...? The point is to give guidance about when and why > to use P, M and L's. I think we should actually discourage other ways to > do very similar things. Sure, if the things are *very similar* then we should steer them towards accepted practice. The argument above is for the case where either the new DoV is not very similar and in the case where it is and the author wants to do it another way. -Andrew
Received on Tuesday, 29 April 2003 19:27:32 UTC