On 4/20/15 8:29 AM, Mark Baker wrote:
> On Fri, Apr 17, 2015 at 5:30 PM, Karen Coyle <[log in to unmask]> wrote:
>> Sometimes the 240 is the title of a collection ("Collection works"
>> "Selections") but more often it's a chosen title for a work whose title
>> varies. In a large number of cases this links the title of a translation to
>> an original title. I'm not nearly as good at finding examples as those who
>> catalog will be, but here are a few examples, included the 245 so it makes
>> more sense:
>>
>> 240    10     |a Zauberberg.  |l English
>> 245    14     |a The magic mountain :  |b a novel /  |c Thomas Mann ;
>> translated from the German by John E. Woods.
>>
>> 240    10     |a Hamlet.  |l Italian
>> 245    10     |a Amleto,  |c principe di Danimarca.
>>
>> 240    10     |a Hamlet
>> 245    14     |a The tragedie of Hamlet, Prince of Denmarke :  |b folio
>> rawshakespeare edition /  |c William Shakespeare ... [et al.].
>>
>> 240    10     |a Harry Potter and the philosopher's stone
>> 245    10     |a Harry Potter and the sorcerer's stone /  |c by J.K. Rowling
>> ; illustrations by Mary GrandPré.
> Thanks, that's very interesting. I have no background (or future!) in
> cataloguing, but what you see in marcpatterns.py is the current state
> of a still very active iterative process by cataloguers. A quick look
> through some of the public library data we've received shows (to my
> untrained eyes) an approximate split between alternate/superior titles
> and collection names so it would appear that some further tweaking may
> be required. I'll pass that on to the team. Thanks again.

Public libraries are probably the lesser users of uniform titles, for a 
variety of reasons. Decisions about UTs for public libraries will not 
translate into decisions about UTs for research libraries. Also, music 
is its own special case. There's not a single answer based on MARC, 
there are multiple answers based on the users of MARC.

>
>> It's hard to read code because each line of code is just ... a line of code,
>> and you don't know from that line of code what the intention of the entire
>> program is.
> marcpatterns isn't "code" in the usual sense of the word. Think of it
> as a document containing declarative rules in a domain specific
> micro-language for mapping between MARC records and BIBFRAME
> resources. It was purpose-designed to be usable by "moderately
> tech-savvy librarians", and we use it that way internally; I don't
> touch it, except to make changes to the micro-language itself.
> Instead, our in-house SMEs do all of the work. Some of the cataloguers
> who've gone through our training have also written rules in that
> language.

OK, now I have to get on my high horse. In my 40-some years of 
developing and working on library systems I've both created and coded 
more tables that I could count. And I can definitely say that tables are 
not documentation in the sense that I mean documentation. And they ARE 
code - coded values to be used by programs. A table does not explain 
what your philosophy is, what you are aiming at, what cataloging rules 
you are attempting to emulate, whether you are following the letter of 
FRBR (the text in the final report) or the code of FRBR (FRBRer) or 
AACR2 or RDA or something you made up.

What I find, and many others find, frustrating about BIBFRAME is 
precisely this lack of a thoughtful discussion on the philosophical 
basis for decisions. I don't mind the work/instance split, but I'd like 
to know how it was decided and what functionality it was developed for.

This question about work titles is (and was intended as) the tip of the 
iceberg. The true iceberg is all of the meaning that is hidden in MARC 
records and that is hard to tease out because that data format also is 
code, and a mess of a bunch of code at that. A subfield like 240 $a has 
at least a handful of different semantics, but it's there because all of 
those different data perform a particular function (collocation in a 
linear, alphabetically organized catalog), not because they mean the 
same thing as FRBR work. If whatever we are moving to after MARC just 
carries forward that semantic mess, we will have made no progress. So 
unless BIBFRAME is based on a sophisticated analysis of the semantics of 
catalog data, it's gonna be pretty much useless.

Here I'd give a </rant> but in fact, it never ends.

kc

>
> Hopefully the documentation at the top of that file will help in
> understanding the rules and the language. We feel that marcpatterns
> offers an enormous amount of value.
>
>> Perhaps something else is done with the 240 in some other lines
>> of code, in some other module.
> Except for linking (880, $6), marcpatterns contains all of
> pybibframe's resource-oriented transformation rules, though you have
> the option of adding your own via a configuration file. Another module
> handles the static mapping of the 006/007/008 fixed length control
> fields to BIBFRAME properties, but those aren't nearly as interesting
> as marcpatterns :)
>
> BTW, if it wasn't clear, I'm with Zepheira. I was unable to subscribe
> to this list using my work email address.

-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
m: +1-510-435-8234
skype: kcoylenet/+1-510-984-3600