[rak-list] mehrbändig

[Gunilla Jonsson]

Dear German colleagues,

I try to add, little by little, a fuller background to our policy decision regarding
cataloguing of multi volume publications. Firstly, a few more words about the one
record solution (with 505) -- even if the Americans do not always use it, or use it
inconsistently, it would, of course, be possible to establish a policy for it. Two
factors, mainly, are behind our decision against it. One is the requirements of the
national bibliography, where you want to have an annual recording of the national
publications. It seems much simpler in that environment to have a proper record for
each part, than to derive something from a single record with a 505 note. The other
factor is that it is not unusual, especially in scientific publishing, to have
different parts issued in more than one series, or different parts issued in different
series. Honestly, I don't know how you would record such relationships within a single
record and keep it clear and easy to interpret (for machine as well as people).

There is another possibility to treat multi volume publications, which would meet Mr
Eversberg's wish to have an "eigenen Datensatz für das Hauptwerk". You might treat it
as a serial publication and make a main entry for the "serial", das Hauptwerk, and
separate records for the parts. Our principal objection to this method is that we
would have to treat an awful lot of publications as serials, which are not true
serials. We looked rather carefully at the consequences of this method, and we didn't
like them. The records for the parts would have the title from 245 p in our example in
245 a. The title from 245 a in our example would end up in 440 as "series title" (or
in 490, in case there is an author involved). You would also provide a main record for
the "series", which would serve as "eigenen Datensatz für das Hauptwerk", and so far
it is good. The records for the parts, however, become very awkward. We looked at many
examples, and the volume titles, with very few exceptions, were not sufficiently
independent. If we look back at the German example I quoted, you would get a record
with 245 a "Die Orgeln und ihre Restaurierung ...", welche Orgeln, muss man fragen,
nicht wahr? Even a cataloguer would startle a little at the first look. The
cataloguer, of course, would soon detect the 440 (or 490) further down and understand
the picture. For the public user, it must be very difficult. These difficulties could
be overcome, I suppose, by a smart display interface. That, however, is not a standard
feature in turn key systems, and would have to be built in house. On the whole, our
chosen policy seems simpler and safer to us.

All this said, I must add that we also decided that the serial treatment can be
applied in cases were the parts really are independent -- works published in someone's
"Gesammelte Werke", e.g., might qualify for this treatment. Even the one record model
can be used, when there are no specific volume titles (apart from volume number), and
continuous pagination, e.g.

The difficulty we might face is that we might be forced to extensive editing of
records imported from external sources  used for secondary cataloguing. We will watch
this conditions carefully, to see wether we will need to revise the policy to some
extent.

Finally, I would like to add that the policy is approved by the national cataloguing
committee, not just the national bibliographic agency. The representatives from the
public library sector, which does not change to MARC21, were very much in favour of
this policy.

Best regards,

Gunilla Jonsson

Bernhard Eversberg wrote:

> On 6 Feb 02, at 9:23, Monika Muennich wrote:
>
> > Liebe Kolleginnen und Kollegen, hier Frau Jonssons Antwort:
> >
> > Personally I wouldn't care that you cannot make
> > the article in p non sorting -- I'm mainly
> > concerned with the function in the database, and besides, I happen to be
> > of the opinion that we might just as
> > well sort on articles, it works beautifully
> Bad idea because you do not always know, when searching, if there IS an article
> on the title page or not. So, you sometimes have to look in two places to be sure.
> AND: titles beginning with the same word are torn apart because the one has an
> article and the other not. This is not beautiful at all, it destroys the value
> of the title index for browsing. This index is useful ONLY if you know the EXACT
> wording of the title including the article AND you are not interested in any
> other titles. Even two editions can differ in nothing but the article. Then, the
> resultind index violates the stated purpose of the catalog to collocate editions.
> >
>
> > I don't know
> > how well 505 works as a basis for indexing.
> Not well. You get too many results and completely wrong results (like for
> "find mozart and sonata" you find something that has a Beethoven sonata and a
> Mozart trio both in the 505.
>
> > There is a subfield t which
> > can be used for title, but the Americans
> > do not always use that.
> They rarely ever use it.
>
> > In many cases the Americans do not supply any 505, or supply very little
> > information in it. How they find things then, I
> > don't know. In fact, I have been told by colleagues that there are lots
> > of duplicate records for multivol. publ.
> > in OCLC, e.g., so perhaps they do not find them.
> >
> In other words, what we get for multipart items is an awful lot of garbage.
>
> Regards, B.E.
>
> Bernhard Eversberg
> Universitaetsbibliothek, Postf. 3329,
> D-38023 Braunschweig, Germany
> Tel.  +49 531 391-5026 , -5011 , FAX  -5836
> e-mail  B.Eversberg at tu-bs.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gunilla.jonsson.vcf
Type: text/x-vcard
Size: 470 bytes
Desc: Card for Gunilla Jonsson
Url : http://lists.ddb.de/pipermail/rak-list/attachments/20020207/33f11fc5/gunilla.jonsson.bin