[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: identifying DTDs, schemas

• From: Norman Walsh <ndw@n...>
• To: xml-dev@l...
• Date: Fri, 24 Nov 2000 13:43:54 -0500

/ "Simon St.Laurent" <simonstl@s...> was heard to say:
| This is a 'best practices' question.
| 
| Say I'm developing XML standards for widespread use across the Internet, in
| systems I don't control.
| 
| Would it make sense for me to:
| 
| 1) Deprecate the use of public identifiers (don't provide them, don't
| recommend their use)

Absolutely not. You should provide them for everything and document a
reasonable strategy for consumers not only to use yours but to extend
yours for their own use (as per the DocBook guidelines for
constructiong public ids for derivative DTDs).

At the very least, public identifiers give consumers with reasonable
entity management systems a chance to painlessly use your standards
over time and they certainly do no harm to those people with, ahem,
poorly designed software.

| 2) Recommend the use of URNs as SystemLiterals in the SYSTEM part of the
| DOCTYPE declaration, in addition to URNs for namespaces, XML and RDF schema
| identifiers, without comment on whether these resources make Infoset
| contributions.

I haven't quite reached this point yet because I haven't seen URN
resolution widely deployed. And unlike public identifiers, a system
identifier that's not resolvable is not harmless.

| 3) Create a registry listing those URNs.

If you use URNs, you should definitely do this. I have some catalog
resolution stuff that also does resolution over the web (as described
in one of the RFCs). I hope to get that published, um, real soon now.

| If your answer is yes, you'll like:
| http://www.ietf.org/internet-drafts/draft-mealling-iana-xmlns-registry-00.txt

Actually, you'll probably like any number of proposals for doing
registries and repositories, assuming you've got a reasonable
resolution system in place. See also ebXML and OASIS efforts along
these lines.

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@E... | A man can believe a considerable deal of
XML Technology Center     | rubbish, and yet go about his daily work in
Sun Microsystems, Inc.    | a rational and cheerful manner.--Norman
                          | Douglas

• References:
  • Re: URNs as SYSTEM IDs
    • From: "Simon St.Laurent" <simonstl@s...>
  • identifying DTDs, schemas
    • From: "Simon St.Laurent" <simonstl@s...>

• Prev by Date: [Zvon Announcement] Namespace reference
• Next by Date: Re: URNs as SYSTEM IDs
• Previous by thread: identifying DTDs, schemas
• Next by thread: Re: URNs as SYSTEM IDs
• Index(es):
  • Date
  • Thread