OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: Comparing PUBLIC identifiers for equality.

[ Lists Home | Date Index | Thread Index ]
  • From: John Cowan <jcowan@reutershealth.com>
  • To: Alan Kennedy <alank@xhaus.com>, "xml-dev@xml.org" <xml-dev@xml.org>
  • Date: Tue, 28 Nov 2000 13:08:16 -0500

Alan Kennedy wrote:

>  o Are PUBLIC identifers guaranteed to be unique? For example,
> if the PUBLIC identifier "-//W3C//DTD HTML 4.0 Transitional//EN"
> is used, can different people legally use that identifier to refer
> to different things?

Public identifiers can be anything at all provided no illegal characters
are used, so nothing can be said about them.

However, *formal* public identifiers (FPIs) are defined by an ISO
standard, and always begin with "+//", "ISO" or "-//".
In the first two cases, they can only be defined by the owner (as given
by a registry in the first case, or an ISO standard in the second
case).  The "-//" FPIs are technically private use, but in practice
organizations do not step on one another's FPIs.

The reason that "-//" is so common is that the registry for "+//" is
de facto broken.

>  o Is there a standard structure/grammar which I can use to
> decompose PUBLIC identifiers?

Formal public identifiers have the form "-//orgname//type title//lang"
where orgname is a plain-text organization name, type is "DTD" or various
other things, "title" is a plain-text title, and lang is the ISO 639 code
for the natural language in which the document described by the FPI
is written.  In the case of registered organizations, "-" is replaced by "+".
For FPIs defined directly by ISO standards, it's "ISO nnnnn:yyyy whatever"
or "ISO/IEC nnnnn:yyyy whatever".

> If PUBLIC identifiers are globally unique, and always refer to the
> same thing, then obviously (?) I can just do a simple string
> comparison to see if they are equal.

There is no guarantee that a FPI always refers to the same thing, if
by "same" you mean "byte-by-byte identical".
There is / one art                   || John Cowan <jcowan@reutershealth.com>
no more / no less                    || http://www.reutershealth.com
to do / all things                   || http://www.ccil.org/~cowan
with art- / lessness                 \\ -- Piet Hein


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS