FW: [RDA-L] Linked data

Posting to RDA-L concerning works with more than a single author

Hal Cain wrote:

<snip>
I think we could devise efficient ways to encode the necessary data in MARC 21, and in a way that will enable older systems (not designed for such extended provisions) to use the data no worse than they do now (supposing the data is actually there). Some may be better carried in  the authority data, perhaps.

And we record the data with all the essential information in sufficient granularity and consistently, surely it should be possible to extract it in other forms for other applications.
</snip>

I agree that doing this (in my idea, this means having more than a single “main entry” or in other words, multiple 1xx fields) could be done so that the records work as well as they ever did. Still, I have considered this, and I don’t know if it could work with the current MARC structure, or, if it could, is it worthwhile to change it? It can certainly be done in MARCXML, although major changes would have to be made, especially the prerequisite of “roundtripability” with ISO2709 format.

It is only a moment’s work to make the 1xx field repeatable, thus making it into the equivalent of dc:creator, while making the 7xx equivalent of dc:contributor, but the problem comes with analytics and name/title headings. So, a 600 for Gilbert and Sullivan must somehow allow for Sullivan as well. Plus, the 7xx and 8xx fields. I don’t know how this could be done in the current MARC format, although in a variant format, such as:

<subjects>
<subject>
<authors>
<author>
<100a>Gilbert, W. S.</100a>
<100q>(William Schwenck)</100q>
<100d>1836-1911</100d>
</author>
<author>
<100>Sullivan, Arthur</100>
<100d>1842-1900</100d>
</author>
</authors>
<titles>
<title>
<245a>H.M.S. Pinafore</245a>
</title>
</titles> [would we need this wrapper?]
<subdivisions>
<standard>
Discography
</standard>
</subdivisions>
</subject>
</subjects>

(You see I’m improvising here)

Delineating each author seems to be necessary to ensure that Gilbert does not get the dates 1842-1900. In any case, this is what we would have to solve if we were to make the 1xx field repeatable and thereby get rid of the problems of a single main entry. In an FRBR world, with a “work” record, a lot of this could be imported, but the ultimate structure would remain similar.

Can this be achieved in our current MARC structures? Perhaps. But since there are much more powerful and flexible formats out there, maybe it’s time to salvage what we can and just move on.

Or maybe there is a MARC solution (non-ISO2709!). I don’t know.

-446

Share