April 20, 2005

Why am I seeing double form subdivisions?

Now that authority linking has been turned back on, please be aware of a temporary issue regarding the updating of headings with form subdivisions. Specifically, when you update a record with a form subdivision in $$v and the authority still uses $$x for the same form subdivision, Aleph will double the subdivision, e.g.,

Authority: Children $$x Prayer-books and devotions.
Bibl hdg.: Children $$v Prayer-books and devotions.
Upd bibl.: Children $$x Prayer-books and devotions $$v Prayer-books and devotions.

The simplest solution is to follow the authorized form and use $$x. Eventually, these will be corrected to $$v; and in the meantime, $$v and $$x will interfile happily, so the coding difference will be invisible to users. For a more detailed explanation of the issue, read on.

When Aleph checks and finds an authority which matches a bib heading, it replaces the bib heading with the authority heading. This serves to correct minor differences in capitalization, diacritics, punctuation, etc., and is generally a good thing. The problem arises because Aleph's sensitivity to coding variations when checking headings and when updating headings is different. When matching a bib heading to an authority, Aleph normalizes out coding differences (along with minor differences in capitalization, diacritics, punctuation, etc.). When Aleph updates a bib heading, it attends to coding differences; in this case, it sees the authority $$x and the bib $$v as different, and preserves both. This preservation of different subfields is good when the difference is that the bib heading includes subdivisions that aren't in the authority, such as $$e or $$4; but in cases like the one above, it can cause problems.

When Database Management and LEO staff were working through these issues in order to get flipped headings from the V14 to V16 conversion fixed and get automatic updating from authorities turned back on, we determined that in most cases, the doubling of form subdivisions was due to the presence of an updated authority containing $$v and older bib headings containing $$x. Most of our bib form subdivisions are still in $$x. To solve the doubling in these cases, LEO ran a job which identified all the authorities containing $$v and set the UPD=N flag in them. This means that Aleph will not update a bib record which appears to match one of these authorities, and will not mistakenly create doubled form subdivisions.

The other side of this coin is the authorities which still contain form subdivisions coded $$x. LC is gradually working through its authority files creating form subdivision authority records (using tag 185) and revising heading authority records which contain a form subdivision. Our authority file contains a mix of these updated and non-updated authorities. As LC completes its work and as we resume our batch updating of authorities with newer versions from LC, these instances of authorized headings using $$x for form subdivisions will go away. For now, though, we're in an interim phase. (NB: These authorities are exceptions. LC does not routinely create authorities for headings ending in form subdivisions, but it does do so when such a heading has been used as an example or a reference on another authority.)

One good thing about the indexing in Aleph version 16 is that it is NOT sensitive to nearly as many minor differences in headings as version 14 was. In version 14, minor differences in capitalization, diacritics, punctuation, etc., and in coding, would cause separate lines (i.e., split files) to appear in the index. Version 16 is much better at recognizing that such differences still belong to the "same" heading, and is able to represent them with a single merged index entry. Therefore, the urgency of standardizing the coding of form subfields has considerably lessened with version 16

So, back to solutions: the simplest correction is to follow the coding in the authority, even when it is obsolete. Doing so will avoid the subdivision doubling, and will not cause problems in the index. The more involved solution is to revise the coding in the authority to $$v, and add the UPD=N field to avoid causing doubling of older bib headings containing $$x form subdivisions. Database Management will revise any authorities like this that we find or have reported to us during this interim phase. This will allow use of the current $$v coding in the bib record. The long term solution will be to replace these authorities with updates from LC. Our batch updating process will include a step to set such authorities to UPD=N as we load them. Eventually we may update all the older $$x bib form subdivisions to $$v, but that is not a high priority given that they are no longer causing split files.

In any case, it is always a good idea to glance over a bib record after sending it to the server. In most cases any automatic heading updates will be good changes; in some cases, the updated record may reveal a case where an incorrectly coded subdivision (e.g., a place name in $$x rather than $$z) caused the kind of doubling that's been described, and should be corrected; and in some cases, there may be a doubling of something that was correctly coded in the bib record but incorrectly coded in the authority, and therefore has doubled.

If you have any questions about this, please let me know.

Posted by s-hear at April 20, 2005 11:32 AM
Comments
Post a comment









Remember personal info?






The views and opinions expressed in this page are strictly those of the page author. The contents of this page have not been reviewed or approved by the University of Minnesota.