On Thu, Jan 31, 2013 at 08:57:06PM +0000, Yanan Zhao wrote:
> Bibliographic metadata - What do you think?
> 
> The major shortcoming with the dc.bibliographic.citation (or 
> dcterms:bibliographicCitation) format of capturing bibliographic citation 
> information is that it's a large string of text in which can be in many 
> different formats, so parsing it out for reuse in Open URL resolvers, for 
> harvesting by citation and referencing software, to create RDF, XML, and 
> other structured data is difficult. (This is not impossible as crosswalking 
> and creating special parsers is achievable.)

Yes, it is usually a mistake to pack multiple concepts into a single
field.  On the othe hand, there are a wide variety of established,
well-known formats for citation.  It might be good enough just to
insist that submitted citation data conform to one that DSpace
supports, and to mark these values according to the format in which it
is expressed.

Is it fair to say that the problem lies in ensuring that the citation
will be decipherable by mechanical means?  Such tools need to know
what format they should parse, and to receive data which are expressed
in that format.

> Would it be better to store the parts of a dc.bibliographic.citation 
> separately and then employ programmatic way to render these as one field? 
> This would be easy and consistency assured. The separate parts would still be 
> available for other functions.

I'd need to study the code to see what is already available, but I
think the idea of a "virtual" metadata namespace whose field values
are the result of crosswalking from other namespaces might be quite useful.

> There is a slight issue in that DC does not have fields for all the parts of 
> a dc.bibliographic.citation; e.g., volume, issue, pages etc however PRISM 
> does, http://www.idealliance.org/specifications/prism/ so this is what we 
> have implemented in our DSpace https://researchspace.auckland.ac.nz/; its a 
> Prism like schema. However, instead of creating local schema, this issue is 
> probably common enough to be better included into standard DSpace schema.

[Deliberately misunderstanding you, to make a point]
There is no standard DSpace schema.  We probably ought to have one,
and move into it a few fields which are in our "DC" schema but aren't in
DC.  There is a task on the list to make DSpace's description of DC
align with the current standard, and a few things that DSpace depends
on would have no place when this happens, if we did not define a new
namespace to hold them.

Meanwhile, better to describe PRISM in a way that can be loaded into
DSpace and distribute that alongside the "DC" description.  I hope
that we can all benefit from your experience here.

-- 
Mark H. Wood, Lead System Programmer   [email protected]
There's an app for that:  your browser

Attachment: pgp297c4Jvtc0.pgp
Description: PGP signature

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_jan
_______________________________________________
Dspace-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-general

Reply via email to