Re: Submission: deltav subset

My 0.5 US cents worth (after currency exchange rates).

I look at the subset submission from the view of what has happened
with Z39.50 (a search and retrieval protocol). The protocol has
so many options and variations, no-one implements them all.
So profiles started to appear. There are a small number of such
profiles. This allowed people to say 'we conform to this profile'.
Not everyone supports all the profiles. Some people don't support
any of the profiles. But it has advantages in that new people
coming along can see some subsets some groups of people agreed to.
It helps newbies pick a subset with a degree of confidence that
its not a silly subset - at least one other person agrees!

So I have no trouble with the idea of profiles being defined
outside the DeltaV spec. Profiles should not be defined based
on what *everyone* thinks is minimal. They should be defined
on what a group thinks is minimal (where the group could be
a single person :-). Others can either decide to support the
profile, partially support the profile, or not support the profile.
They can even choose to write their own profile.
But it makes it easier to understand what different people support.
Eg: "I support the 'Lisa' profile except for this bit." Over time
the profiles may change or profiles may merge or whatever.
Sounds fine to me - as long as they are not a part of the spec.
They are more like implementor agreements etc outside the spec.

So I personally encourage Lisa to define a profile. Lisa can
choose from other people's comments what she thinks should and
should not be included. Others can then choose to support it
or not support it. Just as long as it stays out of the official
standard. DeltaV is complex enough as it is.

Alan

Received on Tuesday, 23 October 2001 19:44:51 UTC