dita — archive

[Eliot Kimber]

The question in this thread is whether anyone objects to Jeff's general review comment that the term "key" should be replaced by "key name" in all discussion of key-based addressing, as in "a key reference can use a key name together with an ID" rather than "a key reference can use a key together with an ID". My initial reaction was that "key name" isn't any better than "key" and that keys are not in fact names (because "name" implies non-uniqueness but keys are unique, meaning they are a form of identifier, not a form of name). But that is clearly pedantic hair splitting. Jeff's intent, if I understand it correctly, is to more clearly distinguish references to keys from "keyref", which in most contexts is the combination of the reference to a key [name] and, optionally, the ID of an element within the topic ultimately addressed by the key. Thus, there could be some confusion between the phrase "key reference" (by which I mean a reference to a key) and "keyref" (the attribute), which involves both a key reference and an (optional) ID reference. As indicated in the discussion copied below, while I don't think Jeff's terminology change is needed I don't see it as hurting either. But I didn't want to unilaterally accept or reject Jeff's suggestion without consideration by the TC as a whole. If there's no serious objection I will use "key name" in place of "key" as I work through comments on the linking and addressing parts of the spec. Cheers, Eliot On 11/3/09 9:24 AM, "Ogden, Jeff"