[osis-core] Functional Requirements

[Chris Little]

On Mon, 23 Jun 2003, Steven J. DeRose wrote:

> >>  Metadata.2 OSIS 1.0 MUST include mechanisms for recording library 
> >>classification data such as Library of Congress, Dewey or other 
> >>publication or cataloging data.
> >>  Metadata.7 OSIS 1.0 MUST provide a means of self-identification of 
> >>metadata based upon any standard classification schemes.
> >
> >Not specifically addressed.  Harry's documents converted from ThML 
> >use <subject subType="LCCN"> for Library of Congress Call Numbers 
> >and <subject subType="lcsh[1-9]"> for Library of Congress Subject 
> >Headings.
> 
> for us an lccn would go under identifier, and lcsh would make a fine 
> subject type (do we define it?)

Defining sounds good, but without requiring x- values for other types.  We 
should at least have best practice note in the manual.
 
> >I would recommend adopting modified versions of these as best practices:
> ><identifier type="LCCN|LCCall|Dewey|ISSN|ISBN|OCoLC|OSIS"> for 
> >LCCN/LOC Call number/Dewey/ISSN/ISBN/OCoLC/OSIS identifiers (plus 
> >more if anyone knows of them).
> ><subject type="LCSH"> for LCSH.  (The LOC's MARC records don't give 
> >any distinction between different subject headings, so I assume the 
> >number in the CCEL versions, as in lcsh1, is unnecessary.)
> 
> sounds good to me, modulo swapping in SICI for ISSN (A SICI is an 
> ISSN plus extra fields that let you specify a particular issue of the 
> periodical, and a particular contribution within the issue (like one 
> article).

Works for me.

I should point out here the important distinction between LCCN and LC Call
Numbers.  "LCCN" is really the abbreviation for Library of Congress
Control Numbers, not Call Numbers--a mistake many people (frequently
including me) make.  I don't know that "LCCall" is be best type for LC
Call Numbers--I just made that one up.

--Chris