[Dspace-general] New METS SIP Profile

[Robert Wolfe]

Hi Scott,

I'm sorry it has taken me some time to get back to you.  I was at the 
DLF Forum last week and am just now catching up with my mail.

Here are my responses to your questions.  Thanks for giving the 
profile the once over.

The reference to the DSpace Dublin Core Metadata Application Profile 
is inaccurate.  We have chosen MODS as the transfer schema for 
descriptive metadata.  The profile documentation is correct, the 
introductory material will be changed.

The links to the governing docs for the MODS profile and the 
Technical Metadata Element Set will not work because those docs have 
not yet been written.  The links are intended to be placeholders for 
future work related to the METS SIP profile.  My word processing 
software made those links active and I will go back and disable the 
links and add more explicit language that the documents are "To Be 
Written" to avoid confusion.

I will add language to the effect that while optional, if the 
CHECKSUM attribute is used it must be accompanied by the CHECKSUMTYPE 
attribute.

We considered the use of the <mptr> element to allow for the 
submission of multiple DSpace Items in one METS document.  In the end 
we decided against transfering anything but ONE DSpace Item per METS 
document.  We couldn't envision a scenario where it would be vitally 
important to submit an entire collection or community (or even a part 
of a collection or community larger than one Item) in one METS 
document.  We have encountered the scenario where we would like to 
archive an Item in DSpace that contains other candidate items within 
itself.  We are planning to archive whole course websites in DSpace, 
but have not yet prepared a mechanism whereby the learning objects 
within the course can be identified and given Item status in DSpace.

At 3:32 PM +1100 11/8/05, Scott Yeadon wrote:
>Hi Robert,
>
>dmdSec - what is the DSpace Dublin Core Metadata Application Profile 
>(where's the schema)? NOTE: in the profile metsiprule05 states that 
>"at least one descriptive metadata record for the DSpace Item that 
>conforms to the DSpace MODS Application Profile..." so one of these 
>statements may be out of sync? I got an error page when I tried to 
>access the MODS pdf link in the profile.
>
>Is this a necessary restriction given that in an unpcoming version 
>DSpace will be able to support several multiple metadata schemas? If 
>the metadata is indicated by schema, is it necessary to have a 
>restriction on schema in this instance? If so would it be better to 
>support a range of which one must be present (DC I guess being the 
>most obvious default)?
>
>file - CHECKSUMTYPE should be mandatory in cases where CHECKSUM is used
>
>Are there any thoughts on exchange of more complex objects such as 
>entire communities, selected collections, items somehow related 
>(e.g. comprise a learning object)? Currently these aren't included 
>but will probably need to be supported in some form in the future 
>(probably moreso in the DIP I guess).
>
>Scott.


-- 
Rob Wolfe
Metadata Specialist
MIT Libraries/OpenCourseWare