Martha wrote:
> I thought it might be useful to some
> catalog users to have the following information readily available:
>
> 1. The first publisher to ever publish this work.
> 2. The first publisher to ever publish this expression.
>
> ...in addition to the information that we would collect
> routinely at the
> manifestation level:
>
> 3. The publisher of each particular manifestation.
>
> This raises a few questions. 1) Would this be useful to a significant
> number of users? 2) Would this have to be collected by
> catalogers, or could
> it be generated automatically by computers? I suppose the concomitant
> questions would be: 1) Would our hypothetical virtual
> catalog usually have
> all editions ever published of a particular work? 2) If not, would the
> information be available to catalogers in a significant
> number of cases?
I would argue that by definition that a 'work' cannot be 'published' -
the work entity is an abstract representation of an intellectual
creation, and can't have something as concrete as a publisher. I hope
this is not controversial. From past discussion, the issue of whether an
expression can have a publisher is clearly more debatable.
I would suggest that the question might be framed as:
Might it be useful ... to be able to tell the publisher of the earliest
manifestation of any particular expression?
The publisher of the earliest manifestation of the earliest expression
is a specific case which is the question you have expressed in 1. above
I can see in some scenarios someone might want to know this information,
but not in a finding sense - so, a scholar looking at the history of
publication may well be interested, but it seems unlikely (although
admittedly not impossible) that a user would be looking for the
'earliest published version'.
In terms of whether this question could be automatically generated by a
computer, given the right information, the answer has to be yes - given
a collection of manifestations with publication dates, linked to
expressions, linked to works, it should be easy enough to answer the
question, which is the earliest manifestation linked to expressions that
are linked to this work.
In a practical sense it is unlikely that you would add manifestation
information to your catalogue unless you actually had an item that
related to that particular manifestation. So the ability to answer this
question would imply that you could not only draw on information in your
local catalogue, but also on information from across a wide range of
catalogues (making it more likely that one of them has the earliest
manifestation catalogued). This in turn suggests that it would need to
be possible to say 'this work represented in my catalogue by these
expression(s) and manifestation(s) is the same as that work in your
catalogue represented by those expression(s) and manifestation(s)'. This
could be difficult if we all represent things in our catalogue locally
as now, but easier if we start to use the opportunity we have here to
cross reference with other catalogues - from my understanding this
linking is one of the things that adoption of RDF will help with, as .
So rather than creating a new work record in my local catalogue, I would
first establish if that work record existed somewhere else, and if so,
reference it in my local catalogue. The same would go for expressions
and manifestations, with only items necessarily having a local prescence
(I think - need to give this more thought perhaps).
I'm not sure the above is expressed that clearly - I know what I mean,
but in real life I would accompany with hand waving and doodles that
would probably not help you, but would make me feel like I was making
more sense. I'm going to follow up with a post that tries to give
concrete examples, and I'd really appreciate others feeding back if I'm
understanding the idea of a 'semantic web' of bib data.
Owen
Received on Tue Dec 11 2007 - 05:23:02 EST