indentifier

On Feb 22, 2007, at 9:03 AM, Brian Suda wrote:
> KEYS for things like BibTeX are not globally unique,
but unique
> only to the bibtext file. This can be generated by any
transforming
> application and are not needed in the mark-up. But KEYS
are not the
> same as IDENTIFIERS. Identifiers should be
human-readable because
> these are used to uniquely identify for find the object
in a physical
> world. Things like key-value-pairs are machines use
only and shouldn't
> be part of hCite.

We're agreeing, just coming at it from different directions.
Perhaps we
should change the description of identifier in the straw man
to:

"An identifier used to used to uniquely identify for
find the resource.
The identifier SHOULD be a URN, but may also be the
identifier assigned
by a specific resource, such as pubmed ID number or Library
of Congress
call number. Identifiers that are not URN's should use the
TYPE
subproperty to disambiguate."

I'm not sure we want to get into the morass of defining
values for TYPE,
but if an identifier is for a specific resource, we should
allow a way
to tell which resource it's from.

Examples:

     <a class="identifier url title"
href="http://www.free- 
culture.cc/">Free Culture</a>

     ISBN: <abbr class="identifier"
title="urn:isbn: 
1594200068">1594200068</abbr>

     <span class="identifier">
       <span class="type">Library of
Congress</span>:
       <span class="value">KF 2979 .L47
2004</span>
     </span>

-- 
Ryan

http://RyanCannon.com




_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Ryan Cannon wrote:
> I'm not sure we want to get into the morass of defining
values for TYPE,
> but if an identifier is for a specific resource, we
should allow a way
> to tell which resource it's from.
I have been thinking of identifiers in the context of
another standard. 
I think this could be a separate recommendation from the
hCite, which 
could be reused in other contexts.

The conclusion of my analysis are:

Every identifier is a collection of predicates of the
subjects that 
together uniquely identify the entity (ex. book title, book
author, and 
book edition year and type - if that works).

Using lexical identifiers 'shortcuts' to these collection of
predicates 
can be useful.

I call these lexical identifiers "issued
identifiers" and sometimes, 
some of these predicates happen to be issued identifiers
themselves.

An issued identifier is issued by another entity.

This other entity may have also been issued an identifier
(typically a 
domain name).

At the end, all these identifiers can be tracked back to a
single root 
identifier issuer.

In addition, an issued identifiers may be permanent, or may
expire.

My 2 cents

Guillaume


_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss