From: "Mike Taylor" <[log in to unmask]>
> Badger me backwards, it's true. We must all have been out of our tiny
> minds when we decided to start the "diagnostics relating to records"
> sequence at 50.
Actually, the plan was to have no gaps at all. We left a temporary gap for
convenience in case there were to be additional query diagnostics *before*
the 1.1 release, and we were going to renumber in any case, before the 1.1
release, leaving no gaps. I think we were pretty much in agreement on the
principal that we didn't want to create the illusion of finite subspaces.
But we forgot to renumber. (That's where we "dropped the ball" in Mike
words.)
As to the current two diagnostics, I don't really care whether they're 47,
48 or 201, 202. An since Jan seems to have the most passion on this, we'll
go with 47, 48.
> > Beginning now with a new range is like what we did with Z39.50
> > diagnostics which worked well, allowing us to easily associate a
> > diagnostic with a particular era.
>
> What is the benefit to that?
As Jan noted, no benefit to implementors. It is useful to those of us who
maintain and document the standard. It has helped me on occasion recollect
why a certain diagnostic was defined.
Jan notes also, the Z39.50 diagnostics are a mess. That may be, but I don't
think it's due to the numbering.
....Anyway...... my plan is to keep them grouped by category, not
consectively numbered, and to maintain an auxiliary page listing them by
number. Further, I'm working on a new page that gives examples and
elaborations, as several of you have suggested, and I'll have a draft ready
soon.
--Ray
|