On Mon, Dec 19, 2011 at 9:43 AM, Glenn Adams <gl...@skynav.com> wrote:
> in other words, I believe that the W3C's tasks do not necessarily have to > include normatively defining specific concrete version mappings for > dependent spec references; this can be accomplished in (2), which need not > be done by the W3C (and indeed has not been done historically, i.e., > defining the criteria for successful certification); > a clarification is perhaps in order; when I say that the W3C has not historically defined the criteria for successful certification, I mean that the W3C has not published "compliance test specifications"; nonetheless, it is true that in defining conformance criteria and levels, the W3C has specified, to a certain degree of abstraction, some (but not all) criteria that would go into (or be referenced by) a compliance test specification; the point is that there is some abstraction at the level of W3C conformance definition that must (already) be made more concrete at the compliance/certification layer; in this regard, the proposal to make use of non-specific (unversioned) references is consistent with providing a certain degree of abstraction that facilitates further concrete definition in the compliance/certification phases; G.