Bibliographic metadata and work-to-work relationships in IAO?
Adam M. Goldstein
as journals, articles, authors, books, and so on---the kind of thing
you'd find in a library catalog. Is this within the scope of the IAO?
If it is, how have people modeling scientific literature with IAO done
it in the past?
Also---and you will have to forgive me if this is something I have
overlooked when studying the IAO, I'd like to model three particular
relationships: the derivation of one work from another, one work's
listing another in its list of works cited, and one work's being cited
by another. Again maybe IAO is not the place for this, although I
would think that it is.
I am happy to start working on this, since it's something I will need
myself.
Any and all comments are appreciated---
Adam
------------------
Adam M. Goldstein PhD, MSLIS
--
agold...@iona.edu
a.m.go...@mac.com
http://www.iona.edu/faculty/agoldstein
--
(914) 637-2717
--
Dept of Philosophy
Iona College
715 North Avenue
New Rochelle NY 10801
Alan Ruttenberg
>
> I need an ontology that models bibliographic metadata for works such
> as journals, articles, authors, books, and so on---the kind of thing
> you'd find in a library catalog. Is this within the scope of the IAO?
> If it is, how have people modeling scientific literature with IAO done
> it in the past?
many "schemas" for representing bibliographic information, but little
in the way of having a framework to integrate them - no way to
organize that a certain piece of a marc record and the object of a
dc:creator relation both name a person, for example. So one thought
about what IAO's place in this is to define terms in this way. We had
a nice meeting about postal addresses, in which we agreed that we
would define information artifacts in terms of what they were about,
taking the target entities from the GEO ontology.
What I don't think we want is just another schema.
For projects needing to cite literature I've been encouraging them to
just use a URI to identify the article. We have
http://purl.org/science/article/pmid/xxxxx in Neurocommons now, but
expect to migrate to shared names (http://sharednames.org/) as soon as
that's ready. Having a simple pointer satisfies the citation required,
and further information can be retrieved at the URI.
> Also---and you will have to forgive me if this is something I have
> overlooked when studying the IAO, I'd like to model three particular
> relationships: the derivation of one work from another,
to IAO, defined as planned processes with both specified input and
output being information entities, but I suspect that's too high level
for what you want. Care to hazard a definition for "derivation" in the
sense you are using it?
> one work's
> listing another in its list of works cited, and one work's being cited
> by another.
> Again maybe IAO is not the place for this, although I would think that it is.
>
> I am happy to start working on this, since it's something I will need
> myself.
describe processes which result in entities related in the way you
intend.
Regards,
Alan
Adam M. Goldstein
> On Thu, Oct 15, 2009 at 9:25 PM, Adam M. Goldstein <agold...@iona.edu
> > wrote:
>>
>> I need an ontology that models bibliographic metadata for works such
>> as journals, articles, authors, books, and so on---the kind of thing
>> you'd find in a library catalog. Is this within the scope of the IAO?
>> If it is, how have people modeling scientific literature with IAO
>> done
>> it in the past?
>
> We've talked about this a bit. One of the issues is that there are
> many "schemas" for representing bibliographic information, but little
> in the way of having a framework to integrate them - no way to
> organize that a certain piece of a marc record and the object of a
> dc:creator relation both name a person, for example. So one thought
> about what IAO's place in this is to define terms in this way. We had
> a nice meeting about postal addresses, in which we agreed that we
> would define information artifacts in terms of what they were about,
> taking the target entities from the GEO ontology.
Hmmm . . . I see. What IAO is doing with GEO doesn't capture the
bibliographic entities themselves, for instance, if these is a journal
article about the Galapagos, and a multivolume work about the
Galapagos, if I understand what you're saying, identifies these as the
same kind of work, because they are about the same place; but the
information that one is a journal article and another is a multivolume
work is lost. I think that these would be further subdivision of the
material information bearer class, if I understand what that is
correctly, which I not confident that I do.
Then again maybe this is the wrong way to go entirely: maybe the
bibliographic entities ought to be understood in terms of products of
their publishers. IAO can be used to describe the various elements of
the content in the context of entities such as material information
bearer; and the ontology of publishing describes the practices and
entities of producing physical objects to distribute information.
There would have to be a way of describing manuscripts and other
archives, never published, as well.
>
> What I don't think we want is just another schema.
I agree with that. Eventually I am going to have to make a decision
about which schema to use, because eventually I want to use the
ontology to organize bibliographic databases by subject.
>
> For projects needing to cite literature I've been encouraging them to
> just use a URI to identify the article. We have
> http://purl.org/science/article/pmid/xxxxx in Neurocommons now, but
> expect to migrate to shared names (http://sharednames.org/) as soon as
> that's ready. Having a simple pointer satisfies the citation required,
> and further information can be retrieved at the URI.
So there will be something like a relationship CitedBy which will work
like
Cyx
y is cited by x
y: a work described in the ontology, http://. . . .
x: another work described by he ontology, http:// . . .
>
>> Also---and you will have to forgive me if this is something I have
>> overlooked when studying the IAO, I'd like to model three particular
>> relationships: the derivation of one work from another,
>
> There is "information processing" in OBI, which I think will migrate
> to IAO, defined as planned processes with both specified input and
> output being information entities, but I suspect that's too high level
> for what you want. Care to hazard a definition for "derivation" in the
> sense you are using it?
Ha ha, I would love to hazard a definition, hang on, I have one right
here . . . oh wait, I'll get back to you on that! Probably I need to
pull apart some of different ideas about what I mean by "derived." The
most interesting and important case, though, is something like the
following: In the late 19th Century, Darwin and Moritz Wagner argued
about whether speciation required geographic isolation. My sense is
that the subsequent history in some sense derives from this initial
discussion. People such as Mayr knew of Wagner, and formed their ideas
in response the difference of opinion between Wagner and Darwin. So
there is a continuous chain of readers and a continuous chain of
theories and their modification that we could, in principle, trace the
idea and its changes from the late 19th Century until today.
One problem is that the idea has changed significantly since then, for
instance, the original discussion was not framed in terms of
population genetics, which didn't exist yet.
>
>> one work's
>> listing another in its list of works cited, and one work's being
>> cited
>> by another.
>
> The easiest thing here is to relate the two by the "mentions"
> relationship.
OK, citing as a kind of mentioning.
>
>> Again maybe IAO is not the place for this, although I would think
>> that it is.
>>
>> I am happy to start working on this, since it's something I will need
>> myself.
>
> How about starting with derivation. I would attack it by trying to
> describe processes which result in entities related in the way you
> intend.
>
Good idea, I will start there.
There are some interesting philosophical questions here. For instance,
if Kuhn is right that there are paradigm changes, and that there is
incommensurability between paradigms, then there may well be no such
deriving one work from another. Well, I think he's wrong about
incommensurability. Accounting for that in the derivation relationship
will be important, though.
Thanks for the thoughtful and useful comments, Alan.
Larry Hunter
Adam,
Pretty much the only work we have done in IAO that is relevant so far
is the definition of a scientific publication, and its sections
(including references). We found that the library of congress FRBR
effort (see, e.g. http://www.loc.gov/cds/downloads/FRBR.PDF) is a
useful starting point, although the definitions generally are not
clear enough for ontological use directly. I specifically addresses
the derivation question you raise.
I, for one, would love to see this material get the careful treatment
it needs for inclusion in the IAO, and would be glad to help you work
on it. A good place to start would be proposing some terms and
definitions in the IAO tracker. We'd then schedule a discussion of
the issues for the IAO telecon (tuesdays, noon eastern time) which you
could lead.
Larry
Goldstein, Adam
I would feel privileged and honored to lead the discussion, and I am glad there is interest in this issue. How about I work up some preliminary ideas and create some tracker issues over the next week to 10 days and then we can discuss it in the next Tuesday conference after that? I will brush up on my FRBR but I have some deeper concerns as well having to do with change over time---in what sense do people now and Darwin address the same issues about species, for instance, when almost none of our beliefs (or at least, most of them) about them are the same? The contexts are quite different. In the Evolution Ontology, I want to a reader to be able to trace back the intellectual lineage of works developing ideas about the same things, even as the concepts and beliefs about those things change with new developments. Well, I do have a solution to this problem, I think. Anyhow, I will explain this in a little more detail in a tracker issue or a linked document.
Best
Adam
=================================
Adam M. Goldstein PhD MSLIS
Assistant Professor of Philosophy
Iona College
--
agold...@iona.edu
http://www.iona.edu/faculty/agoldstein/
(914) 637-2717
Iona College
Department of Philosophy
715 North Avenue
New Rochelle, NY 10801
Larry Hunter
On Oct 19, 2009, at 10:02 AM, Goldstein, Adam wrote:
> How about I work up some preliminary ideas and create some tracker
> issues over the next week to 10 days and then we can discuss it in
> the next Tuesday conference after that?
Sounds good.
> I will brush up on my FRBR but I have some deeper concerns as well
> having to do with change over time---in what sense do people now and
> Darwin address the same issues about species, for instance, when
> almost none of our beliefs (or at least, most of them) about them
> are the same?
Their idea of derived work is more like a translation or a
screenplay. What you are talking about seems closer to trying to
recognize the subject or topic of a work. Handling semantic drift
over time in what a word (like species) means is indeed a challenge.
I look forward to hearing about your approach to this. It would also
be nice to have some specific use cases to think about.
Larry