[
Lists Home |
Date Index |
Thread Index
]
- From: Norman Walsh <ndw@nwalsh.com>
- To: xml-dev <xml-dev@xml.org>
- Date: 29 Jun 2000 13:31:08 -0400
/ Leigh Dodds <ldodds@ingenta.com> was heard to say:
| Personally I think the best way to resolve this is to use a PUBLIC
| identifier
| rather than a SYSTEM identifier (which is what you're using I believe).
You can also use a URN instead of a URL as your URI (try saying that
three times fast :-). But then you *must* have a resolution method
(whereas with a URL you only need a resolution method if you can't or
don't want to reach over the web to the actual location specified by
the URL). And if you have an resolution method, you can use it for
URLs as well.
I've started using URLs for system identifiers for work I'm publishing
on the web (though I may shortly go through the IETF URN registration
process and get my own official NID. Then I'll be tempted to start
using URNs in a vain effort to force the issue of resolution. It
*stuns* me that the world has survived so long without it.).
For example, I use something like:
<!DOCTYPE whatever PUBLIC "-//Norman Walsh//DTD whatever V1.0//EN"
"http://nwalsh.com/whatever/1.0/whatever.dtd">
On my local system, I have a SAX entityResolver hook that looks up the
specified system identifier and replaces it with a local declaration
if one exists. This way, I get fast access to the resource locally,
but I don't publish a bunch of system identifiers that only work on my
machine.
| PUBLIC identifiers avoid the need to hard code path names into DOCTYPE
| declarations.
| Until recently there was no easy/standard way to achieve this in XML.
Unfortunately, you must provide a SYSTEM identifier and consequently
many systems assume they can use that without indirection.
| Norman Walsh has an article [1] on this, with some accompanying classes
| which should
| help you out further.
(Thanks :-)
|
| [1].
| http://www.arbortext.com/Think_Tank/Norm_s_Column/issue_three/issue_three.ht
| ml
As a next step in the evolution of the work I've done on catalogs, I'm
considering attempting to get an OASIS TC formed to revise the TR9401
Catalog file format into an official XML vocabulary with a proper
namespace. Then, in addition to updating the classes to support an
official XML catalog specification, they could be extended to handle
RFC2483 resolution over the web. (In fact, I have a prototype of this
system already running on my laptop :-)
Be seeing you,
norm
--
Norman.Walsh@East.Sun.COM | Why in our youth does the life we still
XML Technology Center | have before us look so immeasurably long?
Sun Microsystems, Inc. | Because we have to find room for the
| boundless hopes with which we cram
| it.--Schopenhauer
***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************
|