Open timathom opened 10 years ago
Are catalogers consciously making this kind of semantic distinction when they create enhanced 505 notes? Yes, I believe it's track titles, chapter headings, etc volumes, each of which are works in their own right. When contents are not broken out nicely, we don't have enough info to build a work, so it's a literal text field, hence the contentsNote. "Also, in the current conversion, the first subfield $g is getting omitted, which throws off the numbering . " I'll fix this. Your last point about the 020's is interesting. I did the work to break out the instances, but I don't think I finished the linkage between 505s and 020s. [note to self: look for "experimental 505a matching to isbn".] In your case, there's not an exact match in the text of the 020 and the 505 to make a link, although a human can figure out that "9788579020018 (vol. 4)" matches "4." from 505$g . I think I made it work in some cases but it resulted in too much crud in others. I'll revisit that code, and maybe have a toggle for "attempt to link 020 and 505", so that systems can opt in if they know they have clean data.
Field 505 00 has two purposes. One is provide better access (than in a 505 0_) to author and title information within a contents note. This is accomplished in BF by creating separate Works for each entry.
A second purpose is to provide a human-readable display of contents information for inclusion in a bib description. This is not available now, and it would be hard to reassemble work entries back into a contents note. To remedy the problem, we suggest also copying the 505 00 into bf:contentsNote (without subfields delimiters). In this way, there would be a parallel outcome in terms of display, whether a cataloger chose to encode a contents note in basic or enhanced format.
The parsing of $g can be off: it does not always appear before $t. Although it is messy, the true parsing is to put all of the subfields between double dashes together in a single work. Here is an example where it goes wrong.
505 00 |t The triumph of time |g (29:38) -- |t Ritual fragment |g (11:23) -- |t Gawain's journey |g (24:37)
<http://example.org/99161770439001452work25> a bf:Work ;
bf:title "The triumph of time" .
<http://example.org/99161770439001452work26> a bf:Work ;
bf:note "(29:38)" ;
bf:title "Ritual fragment" .
<http://example.org/99161770439001452work27> a bf:Work ;
bf:note "(11:23)" ;
bf:title "Gawain's journey" .
Durations are associated with the wrong works.
(OCLC # 878532352)
Well, I had it with $t first, then it was found that $g came first: https://github.com/lcnetdev/marc2bibframe/issues/89, now you have some with $t first again.
Double dashes at the end of bf:title in works can be stripped.
505 00 |t Blew -- |t Floyd the barber -- |t About a girl -- |t School -- |t Love buzz -- |t Paper cuts -- |t Negative creep -- |t Scoff -- |t Swap meet -- |t Mr. Moustache -- |t Sifting -- |t Big cheese -- |t Downer -- |g Live, Feb. 9, 1990, Pine Street Theatre, Portland, Ore. |t Intro ; |t School ; |t Floyd the barber ; |t Dive ; |t Love buzz ; |t Spank thru ; |t Molly's lips ; |t Sappy ; |t Scoff ; |t About a girl ; |t Been a son ; |t Blew.
<http://example.org/99137969100001452work16> a bf:Work ;
bf:title "Blew --" .
<http://example.org/99137969100001452work17> a bf:Work ;
bf:title "Floyd the barber --" .
<http://example.org/99137969100001452work18> a bf:Work ;
bf:title "About a girl --" .
(OCLC # 429048800)
Currently, titles from an enhanced 505 contents note (subfield $t) are being mapped to
bf:contains/bf:Work
. Subfield $g is being mapped tobf:note
.I have a record for a book set that has six volumes:
When I run the conversion, I get a top-level
bf:Work
for the set, and then sixbf:contains/bf:Work
elements for the individual volumes. In this case, it seems to make sense to model the individual volumes asbf:Work
, but will that always be the case? Are catalogers consciously making this kind of semantic distinction when they create enhanced 505 notes? Non-formatted 505 notes are not being modeled asbf:Work
, but are mapped tobf:contentsNote
.Also, in the current conversion, the first subfield $g is getting omitted, which throws off the numbering (i.e., the first
bf:contains/bf:Work
below should have<bf:note>1.</bf:note>
, not<bf:note>2.</bf:note>
):Finally, this record contains an 020 field with an ISBN for each volume. Each 020 causes a new
bf:hasInstance/bf:Instance
element to get generated. However, there is no correlation between these instances and thebf:Work
elements generated from the titles in 505 $t. It seems incorrect to assert that each individual volume is an instance of the top-levelbf:Work
. Shouldn't the instances in this case be mapped to thebf:contains/bf:Work
elements? There may not be a reliable way to do this, but with this record, at least, the volume numbers following the ISBNs could be matched to the designators in subfield $g.