- From: Lynne Rosenthal <lynne.rosenthal@nist.gov>
- Date: Thu, 06 May 2004 18:50:23 -0400
- To: Alex Rousskov <rousskov@measurement-factory.com>, Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Cc: Mark Skall <mark.skall@nist.gov>, www-qa@w3.org
I think there is danger in this. As Alex points out, for outsiders, this doesn't work. It also will add confusion to a topic, I think we are finally getting clarity on. As for SpecGL - this isn't something to promote, but maybe something to provide as an example of something not recommended. --Lynne At 02:50 PM 5/6/2004 -0600, Alex Rousskov wrote: >On Thu, 6 May 2004, Jeremy Carroll wrote: > > > it allows extensions to negate the original spec but only after a > > negotiation which meets Mark's criteria of not "break[ing]" the other > > parts of the spec. > >A non-obvious side-effect of the model is that those who do not >actively participate in the negotiation may be affected (broken) by >negotiated extensions. > >Consider, for example, a tool that visualizes and analyses traces. >Such a tool has to handle post-negotiated extensions but cannot affect >the results of the negotiation. Thus, it may understand the portion of >a trace up to the extension negotiation, but may not be able to grok >the rest of the trace (beyond trivial syntax analyses). > >It is not clear to me whether this side-effect meets the intent behind >Mark's criteria or is good/desirable in a general sense. > >Alex.
Received on Thursday, 6 May 2004 18:51:04 UTC