Thanks Diane: Just a few comments below. (I've preceded your paragraphs
with (DH:) and mine with (CM:).)
Cliff
[log in to unmask] (Diane I. Hillmann) on 21/07/99 19:55:04
Please respond to [log in to unmask] (Diane I. Hillmann)
To: [log in to unmask]
cc: (bcc: Cliff Morgan/Chichester/Wiley)
Subject: Clean Relations
(DH:) I understand your concerns with chronology, but isn't it true that
describing chronology is harder from the point of view of a resource
provider (which is your "take," I think) than from the point of someone
describing something in front of them that was provided by someone else?
>(CM:) Actually, I don't think I said that it was difficult for the
resource provider to indicate chronology - in fact, it's very easy. I know
that the journal issue that I am describing is, say, the March 1999 issue
or a combined March/April issue, or the Spring issue, or an Easter
supplementary special, or whatever. As the resource provider, after all, I
have given the issue whatever chronological label it's got. And this label
is also easy for anyone with the object in hand to find. OK - someone
trying to find Volume 14 No. 2 doesn't know whether this is the February
issue (if it is a monthly), or the April issue (if it's a bimonthly), or
the May issue (if it's a quarterly), or the April-June issue, or the Summer
issue, or the 2nd Quarter issue, or the 14 January issue (if it's a
weekly), or some other date if it publishes, say, 18 times a year. I agree,
though, that all of these issue atributes may well be something that the
knowledgeable searcher would like to search on, i.e. if you know it's a
monthly, you might well want the option to search on the February issue
rather than on Issue No. 2.
>(CM:) So where am I getting to with this? Well, I accept that chronology
(in the sense of issue attribute) is something that is relevant to resource
discovery, even in a core set, since I accept that people may well want to
search on the March issue or the Spring issue, or whatever. I don't think
this has got much to do with *citations* (references) to articles since, in
the STM journal disciplines at least, chronology is almost *never*
indicated in a reference: volume and page(s) is as much as you usually get,
and as much as you usually need. (Exceptions are journals that start each
issue with page 1, where you therefore need the issue number as well, or of
course magazines/newsletters/newspapers, where a date is the identifier. In
our WG, we were concentrating on academic journals.) So, fine, let there be
chronology in the metadata. But I've still got a bee in my bonnet about
chronology not being a true date since it doesn't necessarily tell you when
something actually happened (as I said in earlier e-mails, the February
issue may have been published in January, or March, or April, or later - it
is called the February issue *whenever* it comes out, so is "February"
really a date in the dc.date sense?). I guess I'd got hung up on the values
of the elements following stricter rules on their representation than is
true, so I was thinking of dates in their YYYYMMDD form. (How would you
express Spring?) But if any value can be a text string, I guess (despite my
philosophical objections!), you could have dc.date = February 1999 or
dc.date = Spring 1999 or dc.date = 1st Quarter 1999, if the consensus is
that that's where chronology should go. I'd rather it be there than in
dc.identifier.
(DH:) The arguments sound very similar to those concerning the Date
element, in
some respects. A provider of information may have need for many dates, not
all of which would be relevant to someone seeking to describe or access
that information.
(CM:) Sure. Our metadata as the resource provider is indeed likely to have
all sorts of dates (all indistinguishable in DC Simple, of course) that are
relevant to the resource. Whether they're also relevant to seekers of that
resource is a moot point, but with DCQ, someone might well want to find out
about all articles submitted between two dates, or accepted, or published.
(DH:) But chronology is really, after all, primarily an aid to
identification and
verification, I think. If a user has a citation from another source which
includes a chronologic designation, and it doesn't appear on the metadata,
will this be a problem? In many cases, whether it's a problem or not may
depend on how much other data is there to match to the citation in
hand--not always predictable or definable. I think we can't just dismiss
chronology as not necessary.
(DH:) As for including the pagination in the Relation element--the more I
think
about it the less inclined I am to back off my objections. I think someone
>(CM:) Yes, this was me I'm afraid.
mentioned that whether we expressly allow or recommend using pagination,
that people will do it anyway. This has always been the case (we can't
protect people from their own poor judgment), but there's still a strong
argument for keeping the official DC recommendation "clean" in that
respect. There are alternatives, if there is a need to associate pagination
with other parts of the citation, for the purposes of ordering records,
maintenance, or whatever. But I strongly object to this portion of the
recommendation--I think it compromises our responsibility to recommend only
the optimal practices, no matter what we think will be done by implementors
in the name of expediency.
>(CM:) OK, I really can't argue that the page extent of an article
logically belongs in any IsPartOf relation. If dc.identifier is, by
consensus, the most appropriate place for page (and, by extension, other
locational information such as volume and issue number), does this mean
that we'd just stick to the issue information in dc.relation.
(CM:) For example, would we have:
dc.identifier = 0885-6230(1999)14:2<86:WDSPAP>2.0.TX;2-X
dc.identifier = 10.1002/(SICI)0885-6230(1999)14:2<86:WDSPAP>2.0.TX;2-X
dc.identifier = http://journals.wiley.com/0885-6230/v14n2p86.html
dc.identifier = International Journal of Geriatric Psychiatry, Volume 14,
Issue 2, pp. 86-96
.
.
.
dc.relation = IsPartOf "0885-6230(1999)14:2<>1.0.TX;2-X"
dc.relation = IsPartOf "10.1002/(SICI)0885-6230(1999)14:2<>1.0.TX;2-X"
dc.relation = IsPartOf "http://journals.wiley.com/0885-6230/v14n2.html"
(CM:) Would you also go for a text string here, as:
dc.relation = IsPartOf "International Journal of Geriatric Psychiatry,
Volume 14, Issue 2"?
Regards
Cliff
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|