Print

Print


I would like to mention something.
I agree with Bernhard, but I would like to emphasize that a HasPart and the
IspartOf are mutually exclusive. Some items do have parts, if that's the way we
want to conceive of them. In that case, we would not also make a separate record
with IsPartOf.

So, a record with a HasPart assumes that there is no separate record for the
part.
On the other hand, when the parts have separate records, what Bernhard says
fully applies.
As an example, there are many issues of "Time magazine". Each issue does not
have a separate title, other than the issue number. In US libraries, these are
all cataloged as parts under a single record. Thus, the "HasPart" could be
equivalent to the holdings (e.g. a URL for v.120, pt.2).
But the "Oxford series in ecology and evolution" has several items published,
one of which has the title "Anolis lizards of the Caribbean" by Johnathan
Roughgarden.

Obviously, people want the title of the individual item, but they need to know
(and find) all the pieces of the Oxford series in ecology and evolution. The
IsPartOf works in this way.

The *real* problem arises when someone wants the "Oxfordian ecological and
evolutionary series", but that is a different matter.
    Jim Weinheimer
    Princeton University
    [log in to unmask]

"B. Eversberg" wrote:

> Stephen Buckley:
> >
> > Can someone suggest a solution/ compromise to the following?
> >
> > Is there a recommended way of preventing the "HasPart" qualifier of
> > something like the top level of a discussion group (which could consist
> > of an unlimited number of parts) becoming unwieldy?
> >
>
> In a database, a link between two records doesn't need to be coded in
> both but only in the records at the n-side of a 1:n relationship.
> Thus, "IsPartOf" is the only thing you need, given appropriate
> indexing.
>
> We have the same situation in library catalog data: you don't want
> (and you can't) include a list of "HasPart" links into the record for
> a series (of potentially 1000s of publications) but what you do is
> put a link of "IsPartOf" into every record of a publication belonging
> to the series. The indexing has to make sure that from the record
> for the series (the main record) you can then find the various and
> scattered records belonging to it.
> This way, you also never need to touch the main record again whenever
> a new part of the series arrives. Just make sure you never change the
> key that serves for the linking.
> The same applies to authority records, like for names and
> subject terms.
>
> In terms of database management, the HasPart qualifier is
> redundant if you use "IsPartOf" consistently. For exchange purposes,
> it may have its uses.
>
> regards B.E.
>
> Bernhard Eversberg
> Universitaetsbibliothek, Postf. 3329,
> D-38023 Braunschweig, Germany
> Tel.  +49 531 391-5026 , -5011 , FAX  -5836
> e-mail  [log in to unmask]



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%