 |
 |
For what its worth, I'm attaching how we are handling journals with METS at Brown. Attached is a copy of a journal record (rhythm_run.xml) and a linked issue record (rhythm_issue.xml). These records were jointly developed with the staff of the Modernist Journals Project (www.modjourn.org) at Brown University. patrick -----Original Message----- From: Metadata Encoding and Transmission Standard [mailto:[log in to unmask]] On Behalf Of Scott Yeadon Sent: Wednesday, May 09, 2007 2:19 AM To: [log in to unmask] Subject: Re: [METS] Packaging Journals with METS Hi Jerome, The METS packages we're working with could contain just a single issue or an entire journal, depending on size. Should we come across a large enough issue we could support individual articles, but at this stage that hasn't been necessary. But yes, splitting up larger packages is also a good idea. Thanks for the info, using individual dmdSec elements make more sense to me as well. Scott. > From: Jerome McDonough <[log in to unmask]> > Subject: Re: Packaging Journals with METS > > --Apple-Mail-2-613607687 > Content-Transfer-Encoding: 7bit > Content-Type: text/plain; > charset=US-ASCII; > delsp=yes; > format=flowed > > If it was me, I would actually try to avoid having a single METS > record contain that many objects > and subobjects. Harvard did some experimentation with using the MOA2 > standard for encoding journal > runs, and found that the time involved in processing the resulting > extremely large XML files was a serious > performance problem. The <mptr> element was added to MOA2 (and > carried over into METS) to allow > those with things like journal runs (which can be neatly decomposed > into separate, hierarchically organized > units) to break the various subunits up into separate MOA2 / METS > files to speed processing and make it > easier to extract/process a particular subcomponent from the hierarchy. > > So, you might have a METS record for the journal, with a MODS record > for the journal, with <mptr> > elements referencing separate METS records for each volume (with an > appropriate volume level MODS > record), with <mptr> elements referencing METS records for each issue > (with appropriate MODS records), > with <mptr> elements referencing METS records for articles.... You > get the idea. > > As a practical matter, though, if you're going to have multiple > descriptive metadata records in a single > METS file, it's better to put each one in its own dmdSec. If you're > going to be making links between portions > of a structural map and/or fileSec and individual descriptive > metadata records, you need to put each of > those descriptive metadata records in their own dmdSec, so they've > each got a separate ID attribute value > you can reference from DMDID attributes in the structmap/filesec. > > On May 7, 2007, at 5:59 PM, Scott Yeadon wrote: > > >> Hi, >> >> I have a question on marking up a journal issue or an entire >> journal in >> METS. If for example I was using MODS for descriptive metadata to >> describe >> the journal is it more correct to have the entire journal's MODS >> metadata >> within a single dmdSec (i.e. covering all discrete objects which >> make up the >> journal), or to have dmdSec containers for each of the discrete >> objects >> (e.g. the journal, issues, sections, articles, etc). >> >> In the situation where a single dmdSec held all MODS, the structure >> reflected in the structMap would be (redundantly) replicated within >> the MODS >> metadata (using the part and relatedItem MODS elements). It would >> also mean >> DMDIDs from within the structMap and fileSec sub-elements would be >> linked to >> IDs within the MODS metadata which seems not be very package-oriented >> (surely better to link to package-level elements rather than >> payload elements?). >> >> The description of the dmdSec element within the METS docs appears to >> support both ways (which makes sense given the docs aren't content >> sensitive), so I just thought I'd get the opinions of some experts >> before >> progressing. Thanks! >> >> Scott. >> > > Asst. Prof. Jerome McDonough > Graduate School of Library & Information Science > University of Illinois at Urbana-Champaign > 501 E. Daniel Street, MC-493 > Champaign, IL 61820-6211 > (217) 244-5916 > [log in to unmask] > > >