[Dspace-general] dc.format

[Scott Yeadon]

Just FYI, as of DSpace 1.4.1 format.extent and format.mimetype 
information is no longer stored at the item level but remains associated 
with the individual bitstreams. Previously this information was 
duplicated at bitstream and item level, and as Ingrid pointed out it's 
pretty much useless at item level since it isn't clear which size and 
format belongs to which bitstream.

The size information can be useful for showing users the size of files 
before they download or for analysing content sizes in the repository, 
maybe also as a marginal check the original bitstream hasn't changed or 
been tampered with. Whether this needs to be stored as metadata could be 
debated. Aside from the fact that mime-types aren't all that specific, 
they could be useful for categorising content types in the repository 
(for services such as format migration, identifying formats in danger of 
becoming obsolete) or for identifying particular files when rendering an 
object via browser or other application.

I'm sure there are other cases where people are using this information 
as well.

Scott.

> Date: Mon, 16 Jul 2007 14:41:54 +1200
> From: "Ingrid Mason" <Ingrid.Mason at vuw.ac.nz>
> Subject: Re: [Dspace-general] dc.format
> To: "Beth Tillinghast" <betht at hawaii.edu>, <dspace-general at mit.edu>
> Message-ID:
> 	<75CF552F30ECFA439D9B3008906F2A37022D4A48 at STAWINCOMAILCL1.staff.vuw.ac.nz>
> 	
> Content-Type: text/plain;	charset="us-ascii"
>
> Hi Beth,
>
> My knowledge of this is scrappy, but here goes.
>
> DSpace records in the DBMS the mimetype of the files ingested, so this
> type of 'data' and metadata is already in the system.  
>
> But, one of the good reasons to record dc.format.mimetype is that you
> can search/sort and know *exactly* what mimetypes you have in the GUI,
> because if it's in the metadata you can find it easily, for whatever
> reason.  
>
> It starts to get 'interesting' when you have more than 1 object
> associated with your item record though:  
>
> e.g a thesis with a dataset:
>
> dc.format.mimetype = application/pdf 
> dc.format.mimetype = application/octet-stream 
>
> It seems obvious that the mimetypes are respectively the thesis and
> dataset, but who knows?  In some way, it might be important to know
> which is which, which means indicating this in the metadata.  Maybe add
> an XML file outlining which file is which?  Plenty of people are fine
> about just listing them and letting the searcher piece their way through
> this.  
>
> We have chosen not to implement dc.format to avoid this.  We didn't see
> that that many users would search for file format and the file extension
> indicates the mimetype in the UI.  It may come back to bite us later on,
> if for example it is really important for searchers to know what
> mimetypes are available (for compatibility with software applications).
> Or, with a view to undertaking preservation interventions/migrations.
> But, hopefully they would be done via the 'back end' (DBMS) rather than
> through the metadata anyway.   
>
> Hope this helps.  
>
> Ingrid 
>
> Ingrid Mason
> Digital Research Repository Coordinator
> Victoria University of Wellington
> New Zealand = Aotearoa
>
>
>
>
> -----Original Message-----
> From: dspace-general-bounces at mit.edu
> [mailto:dspace-general-bounces at mit.edu] On Behalf Of Beth Tillinghast
> Sent: Saturday, 14 July 2007 12:38 p.m.
> To: dspace-general at mit.edu
> Subject: [Dspace-general] dc.format
>
> Aloha,
>
> I have a question about an element's use in the metadata schema for 
> our DSpace instance. I notice many, but not all, DSpace instances use 
> the dc.format.extent and the dc.format.mimetype elements and 
> qualifiers. I am curious if there is a best-practices reason for this 
> other than a nice-to-know reason. Can this information be used to 
> generate certain reports, for example?
>
> Thank you in advance for your responses,
> Beth
>
> _______________________________________________
> Dspace-general mailing list
> Dspace-general at mit.edu
> http://mailman.mit.edu/mailman/listinfo/dspace-general
>
>
>
> ------------------------------
>
> _______________________________________________
> Dspace-general mailing list
> Dspace-general at mit.edu
> http://mailman.mit.edu/mailman/listinfo/dspace-general
>
>
> End of Dspace-general Digest, Vol 48, Issue 10
> **********************************************
>
>