 |
 |
Dear metadata lovers, Encoding journal article information in LOM is subjective and dependent on the use/application context of the metadata. There are many feasibilities in doing so as discussed in this thread, and I really would like to emphasise that there is no 'THE ONE' approach that suits all application context, but one that is pragmatic. To share our experience: 1. OpenURL in Location We did implement this function with several location elements with different OpenURL permutations (query syntax). This is a valid approach, despite several issues. One is the URLs pointing to a specific base url of a resolver. This approach also makes decoding OpenURL a precondition for extracting the citation information (c.f. vCard hence is valid). Both issues makes the metadata more platform dependent (specific OpenURL syntax and link resolver). The url may not be valid for other link resolver, targets that uses different OpenURL snytax etc. If the URL is implemented (i.e. it resolved to the actual resource location), it is fine, otherwise, the URL becomes just a proxy string for encoding the citation information, a.k.a bending the location field? A source from the IMS RLI also suggested that OpenURL and other proprietary, standard type links (e.g. JSTOR SICI, IEEE) could be inferred and constructed from the metadata itself, by the application involved dynamically, in addition to using the transmitted hard coded links. This gives flexibility for generating various query syntax suitable for different targets and link resolvers. Also some our links resolve to a webpage representing the full record + several URLs related to use of the item (and location). This use is not in the strictest sense, resolving to the actual location of the resource, but invoking an 'option representation' instead, not sure if LOM location accomodates such use. 2. Relation A journal article is part of a *specific* journal issue, and can be described in the 'relaition/isPartOf'. Andy: you are right to point out the /relation//description field is used for describing 'the other learning object', i.e. the journal title (granularity debate elsewhere please). But it is not wrong to have the citation in the description field because we are referring to a specific journal issue here: volume, issue, part, number details all give the information about the journal title to which the article is related. Hence this is perfectly valid. The tricky bit here is the page details which is related to the article, but this information is still contained within the journal title and hence can also be used to description the journal title: i.e. 'vol 1, part 2, p.3' give description *about the journal title* where the article can be found. If you felt adding the page no. is bending the rule a bit, than perhaps the page details can be encoded in /general//title, description (CanCore method 1 and 2 in Lorna message - below), or annotation fields. For us, this method is practical since the description is parsed and used as it is without any more manipulation. And I don't necessary purport this as a de-facto approach, since it also has the dependency issue similar to OpenURL, i.e. making the description specific to a particular reference system. Lorna: I am interested in the CanCore method 3, but felt that, it may be difficult to use two LOM records per article manifestation. Our application provides discovery services to repurpose library records for use in VLEs; each query can generate many (e.g. 50) articles metadata from different journal sources in a go. This approach add both complexities and runtime overheads - but I am interested in exploring it and in particular, how this can be automated. I don't feel that this approach will become a common practice for manual usage and if it is devolved to users who prone to taking the use of journal article out of context (in reading list, papers etc.) without having to worry too much about (tagging) the related journal title itself. Making tagging of the journal title a prerequisite to the use of an article metadata may likely to become a barrier, and may be best served by the proliferation of existing and easily linked 'journal title/have part' records (not common even in library practice?), and tools to automate the process, making it transparent to users. Bye for now. Boon ---------------------- Boon Low, Learning Technology Officer Edinburgh University Library http://homepages.ed.ac.uk/boon/ On Wednesday, November 19, 2003, at 03:29 PM, Lorna M. Campbell wrote: > Are you library types having fun??! You're seriously distracting me > from > the rest of my work you know .... :-} > > Just to pick up on a couple of points.... > > Paul's vision of world domination i.e. dissagregate LOM, delete bits > already covered by Dublin Core and use RDF to mix an match has already > bee suggested by dozens of frustrated LOM implementors world wide. I > think even those involved in the development of the LOM agree that if > they knew then what they know now they would not have developed such a > monolithic spec. Hopefully modular will be the way to go in the future. > > Sarah's comment about LOM being a baby spec in comparison to > MARC....next time someone complains about having to use LOM I'll quote > you! (Btw Sarah, you are an incorrigible librarian! :-) > > Paul's query about serials and collection titles... CanCore suggest the > following: > > --Series-- > "Examples of series include television shows comprising individual > episodes, individually titled books or e-texts grouped under a common > title, or individually named learning objects grouped together under a > course title. There are several possible solutions to accommodating > both > series and individual titles, the first of which is recommended by > CanCore: > > 1. Indicate the title of the individual component or series episode > followed by the series title in parentheses within the title element, > including the word "Series:" at the beginning. This should be simple to > implement, and its syntax lends itself to automated data migration. > > 2. Include series or secondary title information in > 1.5:General.Description. This suggestion does not preserve the > important > semantic distinction between title and description, and may present > difficulties for automated data migration. > > 3. Use the relation element to point from a metadata record for the > series as a whole to those that describe the individual series > components. In the metadata record for the individual episode, text, or > other learning object, enter its title but do not enter the series > title. In the relation element in the same metadata record, reference > the series using the is part of vocabulary item for 7.1:Kind. > Finally, > ensure that a separate metadata record for the series itself is > created. > In the relation element(s) for this record, use the has part > vocabulary item for 7.1:Kind and reference as many individual series > items as apply. Note that the General and Relation "Identifier" > elements > are understood as referring to a learning object itself rather than its > metadata record. Consequently, this method indicates a relationship > between metadata instances only indirectly, and may consequently > present > data integrity problems in distributed systems." > > Now I'd better get on with the rest of my job... > > Bye > Lorna > > Paul Hollands wrote: > >> Thanks Sarah for formulating my insane mumblings into a cogent >> argument. >> Even I understand what I was talkling about now :) >> >> I think the only bit that wouldn't really fit in relation is the page >> reference. >> >> Sarah Currier wrote: >> >>> Hi Andy and Paul, >>> >>> I don't think this is part of a discussion about what is a LO at all, >>> although I really like the rest of Paul's email! I think the problem >>> comes out of the complexity we are used to describing in a library >>> catalogue record according to AACR2 and MARC, and what can be done in >>> the LOM, and thus, how do we make the LOM do what MARC can do, or do >>> we do what Paul suggests and mix-and-match? While the LOM looks >>> complex to non-cataloguers' eyes, if you've ever worked with the four >>> volumes of LC guidelines for using MARC, it looks like a little baby >>> spec, barely formed. >>> >>> Anyway, the distinction I THINK Andy is making I have laid out below, >>> to be shot down if necessary: >>> >>> Andy Powell wrote: >>> >>>>> I would have thought that using the Relation element would be >>>>> another >>>>> way of doing this, but Andy's subsequent email made me think; does >>>>> this >>>>> element only allow for relations to other electronic learning >>>>> objects, >>>>> or to any resource, such as a journal? >>>>> >>>>> >>>> >>>> I guess that relation can be used to provide information about 'any >>>> other >>>> resource' (on the basis that for most things, there will be >>>> somebody, >>>> somewhere that considers it to be a 'learning object')! >>>> >>>> My point is that, if you are describing a journal article then the >>>> journal >>>> is a 'related resource' (and therefore, information about it could >>>> go into >>>> 'relation') but the citation for 'the article in the journal' is >>>> information about the article - therefore it doesn't belong in >>>> relation. >>>> >>> Took me a minute to get my head round this; now I think you are >>> differentiating between, for the following fabulous article: >>> >>> All About the Correct use of Semi-Colons by Sarah Currier published >>> in >>> the July 1561 issue of the Fully Refereed Journal of Unbelievably >>> Anal >>> Cataloguing Discussions. >>> >>> The "Relation" thing would be: >>> Title: All About the Correct use of Semi-Colons >>> IsPartOf (I'm making this up, can't be bothered looking at the LOM): >>> The Fully Refereed Journal of Unbelievably Anal Cataloguing >>> Discussions >>> >>> The citation thing (which would go in the URL according to Andy) >>> would be: >>> Title: All About the Correct use of Semi-Colons >>> Citation: Fully Refereed Journal of Unbelievably Anal Cataloguing >>> Discussions (July 1561), pp. 23-346. >>> >>> Now, I'm not sure that the latter COULDN'T just go in the relation >>> field, although I take Andy's point about the distinction. I guess >>> the >>> question is, what is the easiest and most useful way to do this for >>> the sake of both end users looking at records, and for sharing, >>> cross-searching and exposing of metadata records? >>> >>> As for series titles and collection titles: good point Paul. >>> Repeatable alternate title would be an excellent element for the LOM. >>> >>> Phew, I don't often get to have fun like this in this job. >>> >>> ;-) >>> Sarah >>> >>> >>> >>> >>> >>> >>> >>>> >>>> >>>>> If Andy's solution was used, how would the information about the >>>>> journal >>>>> be made easily accessible to the end user? Would the user interface >>>>> have >>>>> to take the journal details out of the OpenURL and present them as >>>>> a >>>>> journal citation? >>>>> >>>>> >>>> >>>> Yes. That's what computers are good at! ;-) The OPenURL is designed >>>> to >>>> be machine-parsable. >>>> >>>> Andy >>>> -- >>>> Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK >>>> http://www.ukoln.ac.uk/ukoln/staff/a.powell +44 1225 383933 >>>> Resource Discovery Network http://www.rdn.ac.uk/ >>>> >>>> >>>> >>> >>> -- >>> ******************************************* >>> Ms. Sarah Currier >>> Coordinator / Research Fellow >>> Educational Content Special Interest Group (EC-SIG) >>> CETIS (Centre for Educational Technology Interoperability Standards) >>> Rm. 2.08B, Centre for Academic Practice, University of Strathclyde >>> Graham Hills Building, 50 George Street >>> Glasgow G1 1QE, Scotland, United Kingdom >>> Tel.: +44 (0)141 548 4573 Fax: +44 (0)141 553 2053 >>> E-mail: [log in to unmask] Mob.: +44 (0)7980 855 801 >>> Web (EC-SIG): http://www.cetis.ac.uk/educational-content/ >>> Web (Dept.): http://www.strath.ac.uk/Departments/CAP/ >>> ******************************************* >>> >> >> -- >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> Paul Hollands <[log in to unmask]> >> LTSN-01 Information and Web Support Officer >> University of Newcastle, 16/17 Framlington Place >> Newcastle upon Tyne, NE2 4AB >> 0191 222 5888 >> <http://www.ltsn-01.ac.uk/> >> > > -- > Lorna M. Campbell > Assistant Director > Centre for Educational Technology Interoperability Standards (CETIS) > Centre for Academic Practice, University of Strathclyde > +44 (0)141 548 3072 > http://www.cetis.ac.uk/ >