 |
 |
[Sorry for cross-posting; but we agreed to focus this discussion on cetis-metadata] > From: Noble Howard [mailto:[log in to unmask]] > To: [log in to unmask] > > If there were to be a Reading List XML template it perhaps > should be here: > http://www.xml.org/xml/registry.jsp I searched for "reading" there and got nil results. >From: Andy Powell [mailto:[log in to unmask]] >I suspect that we'll have to invent something. There is some work in this >area at >http://www.rdn.ac.uk/resource-centre/documentation/ >but you should note that this is an unfinished bit of work undertaken by >Monica Bonett at UKOLN. We did this as part of some thinking behind an >RDN service we were going to call the RDN ResourceCentre - at the >current time we have no plans to take this work forward, partly while we >wait to see how commercial tools in this area pan out. I've just read the above document, and the item description in it looks very appropriate to me. Can anyone (who knows OpenURL v2 better than me) read this and confirm whether all the bib. attributes described can be expressed in it? Monica's <owners> table is unnecessary (the RL as a whole can include DC creator fields). Separating the non-bib. pedagogical application (of this item to this course/list), as Monica does in a <notes> (suggestions for a better name? <resourceLearningApplication>???) table, is a good idea. The unique IDs (for <items>, <notes> and <readinglists> can be replaced by something defined within an institutional domain/namespace. They can't just be course codes, because there may be a requirement to hold past, current, future lists for the same course/unit/module. We (DELIVER) will draft some additional attributes/elements that we think are needed within <ReadingList> and <resourceLearningApplication>. Also anything we think of that will accommodate technical interoperability (without data loss) with the vendors we're working with (but still be non-proprietary). [Andy] >We took the view that an item in a reading list is either a book or a >journal article or a Web resource and identified a set of properties for >each, based largely on the fields in the OpenURL and in ReferenceManager >export format I think. I think this limit of item types is acceptable. The mapping to RIS is very useful. We know some academics who use EndNote to compile lists, so a conversion tool (RIS -> ReadingListML) would be handy. Conversion the other way would inevitably lose our non-bib. attributes; but I guess might be useful for students to import a list. John -- John Paschoud Project Manager, DELIVER <http://www.angel.ac.uk/DELIVER>