I too endorse the value in a mechanism to create a set of records and
still have a file that can be validated and processed as a single XML
file. The TEI2 (Text Encoding Initiative) DTD has a similar capability
and it proves very useful in dealing with collections of small documents
(such as issues of a magazine). And as I think about doing (in an XML
Schema environment) the tasks I have done with sets of records (MARC, and
using a variety of XML DTDs) when integrating collections from other
institutions into American Memory, I am confident that I would soon notice
its lack.
Caroline Arms [log in to unmask]
Office of Strategic Initiatives
Library of Congress
On Thu, 28 Mar 2002, Dick Thaxter wrote:
> On Wed, 27 Mar 2002, Houghton,Andrew wrote:
>
> >
> > A problem that I have is that MODS defines no recordset or
> > dataset element. With MARC21 for example, I can specify
> > multiple records in a single MARC file. The way MODS is
> > specified is that it only defines a record. You should
> > consider having the <mods> element as the recordset or
> > dataset element and have a separate <record> element.
> > This would allow placing multiple MODS records into one
> > XML data file. For example:
> >
> > <?xml version="1.0" ?>
> > <mods ...>
> > <record ... />
> > <record ... />
> > </mods>
> >
> > The recordset or dataset level element should allow zero
> > or more records, like the record element allows zero or
> > more elements. For a single record you now have introduced
> > an additional node level but that doesn't bother me because
> > the abstraction allows me to easily determine how many records
> > are in the dataset by looking at the number of child elements,
> > in the XMLDOM, that are attached to the document node <mods>.
> >
> > Andy.
>
> I have to strongly agree with this proposal. I've written a MARC to MODS
> conversion program and it can be very messy to create hundreds of output
> files. A <record> element would be very handy for creating sets of
> records within one file. In this case would the record identifier element
> serve as the unique id for each record, or would an XML ID attribute be
> assigned to each <record id=xxx>? I'm ready to try this out.
>
> Dick Thaxter
>
> *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*==*=*=*=*=*=*=*=*=*=*=*=*=*=*
> * Dick Thaxter [log in to unmask] 202 707-7208 *
> * Automation Specialist *
> * Motion Picture, Broadcasting & Recorded Sound Division *
> * Library of Congress *
> * The usual disclaimers apply *
> *=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
|