 |
 |
It is clear that there are areas of redundancy between PREMIS and METS. The underlying principle in defining elements in PREMIS was that these are what a repository needs to know for preservation purposes. How it knows is not in scope. So, for instance, the structural map in METS defines relationships between files. PREMIS has a bunch of elements about relationships. If using METS and your structural map defines the relationships you would simply not include the PREMIS elements that detail relationships (at least those that are considered structural relationships). For those few elements that are mandatory according to the PREMIS schemas (e.g. objectIdentifier) and there is a corresponding element in METS (objID), if using METS you would have to record that redundantly if they would be the same value. We need to develop best practices for use with various implementations and document them. This can only come by trying to use the PREMIS schemas in different environments. If people are starting to have examples of how they're using PREMIS, it would be nice of people in this group would share them. We have a Swiki for our use to facilitate this. More on that in a later message. Rebecca On Tue, 18 Oct 2005, Steve Bordwell wrote: > Olaf, > > I would agree that it is essential that there is only one PREMIS > metadata schema, and that multiple flavours to meet the needs of a > specific application or implementation defeats its purpose as a > standard. > > It should not be assumed that all those who use PREMIS will do so in > conjunction with METS. For a number of reasons we will not initially be > implementing METS. We still want to use PREMIS and are designing a > database using PREMIS. From the viewpoint of a non-METS user, PREMIS > therefore needs to remain a complete and coherent schema in itself. > > Perhaps I have missed the point, but I am concerned that if there is > redundancy between PREMIS and METS, that PREMIS semantic elements will > be removed in favour of elements of METS. > > Steve > > -----Original Message----- > From: PREMIS Implementors Group Forum [mailto:[log in to unmask]] On Behalf Of > Olaf Brandt > Sent: 13 October 2005 17:53 > To: [log in to unmask] > Subject: [PIG] mandatory elements > > Hi, > > I would like to argue, that we should provide a standard, which is clear > on the one hand but which is also easily usable. Usability in the > context of the real world means, that we consider that many institutions > might use METS. > > One main result of point 5.2., metadata usage, in the PREMIS > surveyreport (page 45, > http://www.oclc.org/research/projects/pmwg/surveyreport.pdf ) > is that > "METS was by far the most commonly used scheme. Survey results indicate > adoption in all three types of institution, although to varying degrees: > 64% of libraries, 42% of archives, and 35% of other institutions used or > planned to use METS." > > The PREMIS group recommended the usage of METS: > The following appear to be trends in practice that may ultimately emerge > as best practices: > > ...Use the METS format for structural metadata and as a container for > descriptive and administrative metadata; use Z39.87/MIX for technical > metadata for still images. (Trends and conclusions, page 7) > > I my opinion, we should at least have a closer look, how the PREMIS > Metadata schemas could be integrated in METS without too many > redundancies. > > One of my main concerns is the split up of the PREMIS schemas in too > many different flavours. That could mean a lack of common tools. If > there is a way to integrate the PRMEMIS schmeas easily in METS (but > being explicit enough!), I would be glad. > > Cheers > > Olaf > > ________________________________________________________________________ > This e-mail has been scanned for all viruses by Star. The > service is powered by MessageLabs. For more information on a proactive > anti-virus service working around the clock, around the globe, visit: > http://www.star.net.uk > ________________________________________________________________________ > > ________________________________________________________________________ > This e-mail has been scanned for all viruses by Star. The > service is powered by MessageLabs. For more information on a proactive > anti-virus service working around the clock, around the globe, visit: > http://www.star.net.uk > ________________________________________________________________________ >