[
Lists Home |
Date Index |
Thread Index
]
Title: Why make namespaces so complicated?
Sorry
about that last one. Key combinations are deadly.
1. Yes, they are both just strings. urn:publicId is also a
URN.
2. PublicIDs can contain rubbish, but so can a URI.
Just strings.
3. PublicIDs are as meaningful as the registration system makes
them. That is a choice. Conserve options.
4. Owner, Class, Version and Language don't change it being
just a
name. They make it a name with structured parts. Used
as
designed, that can be useful information if one cares to use it.
Again,
choice. Explore options.
5. I understand the notion of immediate self-resolution. But
that
also confuses people. I can work with it. Yes, a PublicID
forces
some thought, but it also opens up the potential to
use
the public registries and be very clear who owns a definition,
and
what class and version of that definition is asserted to be
in
effect for this content.
ROAs
have value in contract processes. Assertions in the
content about the contract under which it operates have
value.
6. Because a PublicID can be a namespace name value, we
have
to include it in our discussions. Otherwise, the decisions
made
by the TAG based on XML-Dev can deprive us of meaningful
choice
where what is meaningful is chosen locally. Conserve
options. It is there. Why?
In
other words, for those who say a namespace name should
just
be a name, the public ID is fairly clear. (it won't get turned
blue
and become clickable AFAIK). Given an OASIS catalog,
useful
things can be done with it beyond resolution.
Namespaces are better than packing the schema in the element.
Weirdly,
we do
a lot of gymnastics to get around shipping a schema
or DTD
and preserve well-formedness.
len
Understand all these things about public IDs, the
point I was trying to make is that at the end of the day both a public ID and
a URI are Just A String. URIs can be resolved by indirection just as well as
public IDs and public IDs can contain essentially meaningless rubbish. Yes,
they can carry some very useful registered information, but (correct me if I'm
wrong) that's only really meaningful if you have an ISO registration code of
some kind. I am no great authority, but by carrying things like Owner, Class,
Version and Language, suddenly they are not really Just A Name anymore, so I
can see the "information" conveyed by a public id also going stale. What do I
do when I get a document with a public ID whose owner has gone out of
business? I can't see it being any more useful to me than a dead URL
link.
The
advantage of URLs, as a particular subset of URIs, is that they have the
flexibility to be "self-resolving" as a default behaviour, which is surely
quite handy in some circumstances. They don't have to be used on the
internet, either.
The
big advantage of public IDs is that you simply can't do much with them without
a resolving mechanism (implicit or explicit), which forces people to think
about it a little more. This is surely not a Bad Thing.
John
|