[
Lists Home |
Date Index |
Thread Index
]
[Mark Feblowitz]
> Regardless of where you start from, versioning spans both the syntactic
and
> the semantic. I could have two syntactically identical schemas, and could
> choose to assign them different versions (and, if included in the
namespace
> name, different namespaces) if the intended semantic interpretation were
to
> change.
Yes, but then you would have two schemas that accidentally happen to have to
same syntax, like homonyms in the domain of words. There is no way a schema
processor could or should tell that you are using one of them in the wrong
context. This situation is very much related, I think, to naming elements
or program identifiers with human-useful names. They are invaluable, but
not yet to computers.
>The trouble with having a richer language for describing legal
> structures and types is that Schema blurs the distinction between
syntactic
> and semantic, even if it cannot capture/enforce all of semantics.
>
I don't know ...does it really lend itself more readily than DTDs to people
inadvertently slipping semantic expectations in? That is not obvious to me.
So far I stand by my claim that XML Schema is only about structure, even
though it is true that people use use it to apply names that are useful to
remind them or guide them about the semantics. If you want to give a schema
a new version to hint that new semantics should be applied, that is an
overlay that has nothing to do with the schema.
Anyway, if version 1 has semantics A, and version 2 is supposed to have
semantics B, then what should I expect version 3 to be about? Logical
extensions oo some completely new semantics? Confusing expectations, so it
probably would not be a good practice.
Cheers,
Tom P
|
