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

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   RE: [xml-dev] How did "public identifier" get its name...

[ Lists Home | Date Index | Thread Index ]

Per the SGML Handbook (Goldfarb, Clarendon Press Oxford, 1990):

pg 180:

"a public identifier is ... a unique non-system specific identifier 
that can be converted to an actual storage identifier only by a table 
lookup.  The universe in which a public identifier is unique depends 
upon whether it is a formal public identifier and upon its owner 
identifier.

4.239 public identifier: a minimum literal that identifies public 
text"

pg 378: "A public identifier is a name that is intended to be 
meaningful across systems and different user environments.  Typically 
it will be a name that has a registered owner associated with it, 
so that public identifiers will be guaranteed unique and no two 
entities will have the same public identifier.  This uniqueness can 
only be achieved when the public identifiers are "formal public 
identifiers" which is an optional feature that can be specified 
in the SGML Declaration."

The next round of quibbling will be over whether or not a non-URI 
identifier is of value, whether or not non-resolvable identifiers 
are of value, and how to blame SGML.  I'll save you some trouble:

1.  No identifier is actually independent of a system of some 
kind because an identifier without a system of identification 
is just a syntax; in other words, the only provable property 
is the spelling and structure, same as XML.

2. All a public identifier gets you is an identifier that is 
mapped to the local storage system identifier (the system 
identifier), but note, it is a mapping from one system of 
identification to another.

3.  URIs get around that by definition, though they in fact, 
are the same thing.  This is at the heart of the most non-resolvable 
permathread on the web:  is identity independent of identification 
or what do the names of resources name.   Note this thread nevers 
impacts actual web operations; just their formal definitions in 
systems that define these incompatibly (non-isomorphic operations: 
systems are defined as data plus operations; an XML application 
language is just syntax plus documentation without an accompanying 
object model for implementation.)

But the nut of it is system independence:  if you want your 
names to be independent of the URI system (say the web), then 
you need a means (a system) to map and a means (a system) 
of registration to enable uniqueness.  Think ISBN or VINs.

Yes, there are implementations that map PUBLIC identifiers to 
web identifiers.  For them that needs 'em, use 'em.  OTW, not. 
If you are using the web as your mass storage device, you 
don't need public identifiers for much if anything although 
they are still used in DOCTYPE declarations.

len 


From: collin@seu.edu.cn [mailto:collin@seu.edu.cn]

...and what does the "public" indicate?

Thanks!

Collin




 

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

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