Looking at ISBD

It's quiet here in OCLC Research with several of my collegues at ALA, and the topic of ISBD Punctuation has come up several times recently, so I thought it would be interesting to learn more about it.  I've been aware of ISBD since it was first issued in the 1970's.  I remember the series of pamphlets describing it, but only recently realized how entrenched it is in AACR2 with the discussions about its place in RDA.  From my point of view it mostly gets in the way of processing bibliographic information, so I've never been very particularly enamored with it.

Update (January 28, 2008): Please forgive my confusion about what is and what isn't ISBD punctuation (see comments).  I suspect I'm not alone in this, but should have known better.  In fact, I think some of the principles of ISBD punctuation (such as leaving double punctuation in) leak into many headings that aren't part of ISBD. In particular, the 110 field is not covered by ISBD.

I suspect that part of my problem with ISBD has been a lack of understanding of the whole relationship between ISBD, AACR2 and MARC21, a lack I suspect many of us have.  Of course I don't actually have to catalog anything, I just try to cope with what catalogers (or would-be catalogers) generate and try to make sense of it.

Although I've never processed records in UKMARC, I've always heard that it avoided ISBD punctuation.  Looking into that I found that UKMARC does have ISBD punctuation, although it eschews it at subfield boundaries (an approach now used by MODS).  Now ISBD punctuation doesn't get in the way too badly if what you want to do with the data is a card-like metadata display.  In fact it is a benefit since the punctuation gives you clues about the meaning of the text even if you can't read the text itself, which can be helpful.  But the minute you want to do something slightly different with the data, the punctuation gets in the way.  Is that period (i.e. full-stop or point) at the end of the text there because of ISBD or because the word before it is an abbreviation?  What were our experts thinking about when they decided to prescribe ambiguous punctuation?

Well, of course they didn't really decide that, and the punctuation looks fairly unambiguous if done correctly.  The trick is to double up, as described in Preliminary Consolidated Edition of the rules:

0.3.2.7 When an element or area ends with a point and the prescribed punctuation for the element or area that follows begins with a point, in order to take into account punctuation for both abbreviations and prescribed punctuation, both points are given.

Although I also found in ISBD(ER):

0.4.7 When an element ends with a point and the prescribed punctuation for the element which follows begins with a point, only one of the two points is given.
e.g. 3rd ed. -

not 3rd ed.. -

And then ... - 4th ed.
not And then .... - 4th ed.

After a couple of readings, that started to make sense.  This doubling seems to always have been there, as I found similar language in the 1974 ISBD(M):

When other punctuation is included, the prescribed punctuation is given, even though this may result in double punctuation.

And, in fact, double punctuation is often found in bibliographic records.  Unfortunately not regularly enough so you can depend on it.

As an exercise I extracted the 9.3 million 110 (corporate author) fields from WorldCat and tried to eliminate the ISBD punctuation.

I aimed for something like 99.5% accuracy, so that only 5 headings out of a 1,000 would have either ISBD punctuation left on or non-ISBD punctuation taken off.  With a table of around 60 abbreviations to look for (found by looking at abbreviations that inappropriately got their tailing period deleted) and a few simple rules, it's not that hard to get pretty close to that for the 110's.

One of the main problems with ISBD punctuation is that it is applied inconsistently.  Combine that with a syntax that is hard to distinguish from normal text and you are guaranteed cataloging that is difficult to process.  In those 9.3 million corporate authors I found some 700,000 subfields that might have had an ISBD 'point' at their end, but my program's best guess was that it actually belonged to an abbreviation at the end of the subfield.

ISBD is certainly an accomplishment in its influence on cataloging practices around the world, and we all benefit by the increase in consistency it undoubtedly brings.  The idea of mixing standard punctuation in with coded and uncoded data though, is dubious at best, and at worst ends up with some very strange displays of bibliographic data.

For FRBR processing, what I'd really like to do is parse the parallel title information embedded in title fields with a complicated set of ISBD colons, slashes and equal signs.  A preliminary look at the 1.5 million 245 fields in WorldCat with equal signs in them is enough to scare almost anyone, but I might give it a try since it may substantially help matching titles of some classes of material.

--Th

Comments

[kentsin]

The problem of ISBD is, on the first look, it can co-exist with marc, but after some try. You find it in-compatible with MARC.

Anyone try to generate full ISBD from UNIMARC known this.

It is time to drop ISBD completely.

Or an ISBD version which can be fully generated from (at least UN)MARC, i.e., use BNF to define it.

[J. McRee Elrod]

You make two basic mistakes. You assume ISBD is punctuation. ISBD is primarily a selection and order of elements. Punctuation is secondary.
You assume 110 has ISBD punctuation. ISBD (speaking in MARC terms) begins with 245 and ends with 5XX notes and 0XX standard numbers. Main, added, and subject entries are outside ISBD.

[John Hostage]

I don't understand why you were studying the 110 field, or what that has to do with ISBD, since the ISBD says nothing about headings.