Thursday, March 11, 2010

RE: [RDA-L] expressions and manifestations

Posting to RDA-L
Karen Coyle wrote:
<snip>
Quoting Laurence Creider

> Is their a technical reason for your statement MARC is "not up to providing" the appropriate subfields? MARC21 certainly allows for indication of the thesaurus from which subject terms are taken, and presumably that could be extended to other fields as well.

There are a number of reasons. Here are a few:

1) there are only 36 possible subfields in every field. In many fields, there are none or at most one left to use
</snip>


This assumes we are stuck forever with ISO2709 records transferred using Z39.50. The moment we change to almost any other format, we have an infinite number of fields and subfields. For example, here is part of a MARCXML record (totally made up):
<datafield tag="700" ind1="1" ind2=" ">
<subfield code="a">Jones, John</subfield>

<subfield code="t">The tree frogs of Texas</subfield>


We can add a subfield:
<subfield code="relation">b</subfield>
(b is a code defined as: "Has part" or "earlier version" or "based on" or whatever you want. If we want natural language text, we can do that too.)
<subfield code="relation">HasPart</subfield>
</datafield>


We can't do this in our current MARC format since we are stuck with single digit subfield codes because of the limitations of ISO2709:

700 1\ $aJones, John$tThe tree frogs of Texas$relationHasPart

[theoretically, today we could add the entire UNICODE character set, but I doubt if a lot of people would want to add a subfield lambda λ or shin ש! In any case, there is little sense to expand an obsolete format]

In fact, once we move beyond ISO2709, we could even do things that can interoperate with other formats, e.g. Dublin Core (for an analytic):
<DC.Relation.hasPart>
<datafield tag="100" ind1="1" ind2=" ">

<subfield code="a">Jones, John</subfield>

</datafield>

<datafield tag="245" ind1="1" ind2="4">

<subfield code="a">The tree frogs of Texas</subfield>

<subfield code="c">John Jones</subfield>

</datafield>

<datafield tag="300" ind1=" " ind2=" ">

<subfield code="a">p. 34-85</subfield>

<subfield code="b">ill.</subfield>

</datafield>

...

</DC.Relation.hasPart>


This is just as easy with RDF or almost any other modern format. The number of codes and relationships will be endless and we can gain a lot of freedom once we dump that outmoded, obsolete ISO2709 format, which has fulfilled its function but is now holding us back. This does *not* mean that we must abandon MARC. Each bibliographic agency can add on its own sets of fields and subfields, so long as the XML is correctly defined.

Whether we need an endless number of codes, fields and subfields I do not want to discuss here. But I think people can understand why non-librarians see that ISO2709 is a kind of straight-jacket in today's world. A lot of those same non-librarians also conclude that MARC format is just as obsolete, but I disagree and believe that MARC can survive so long as we rethink it.

No comments:

Post a Comment