Donate Share

JabRef

Tracker: Feature Requests

5 XMP Metadata mapping is too limited. - ID: 1646777
Last Update: Comment added ( ccodere )

The mapping to dublincore is too limited, since most tools do not know of
the bibtex namespace. On the top of the bibtex schema, the dublincore
schema should always be created and populated by default.

Suggested mapping is as follows:

author -> dc:creator
title -> dc:title
date -> dc:date
keywords -> dc:subject
abstract -> dc:description


If its too complicated to path this yourself, i can always send you a
patch!

Carl






Carl Eric Codere ( ccodere ) - 2007-01-29 05:01

5

Open

None

Nobody/Anonymous

None

None

Public


Comments ( 4 )




Date: 2007-02-07 05:46
Sender: ccodereAccepting Donations


I have attached the modified source code with correct identifier addition
in the dublincore schema... It is sad that the XMP Standard only supports
one dc:identifier, it should normally be a Bag, not a Text field... It
should be clarified with Adobe, since this is not consistent with
Dublincore (I will contact them).

I have made the following changes:
- Identifier doi is now encoded using DOI: uri prefix (I imagine that in
the GUI, the prefix is never present, otherwise my code is wrong)...
- I have put in comments also (untested) ISBN identifer to urn mapping for
xmp, you just need to uncomment if you need to support this feature (either
doi or isbn should be placed in dc schema, not both...)

Have done the following tests:
- Export to Embedded-XMP PDF
- Import into JabRef the Embedded-XMP PDF
- Verify that the encoding is correct by using one of my tools to read the
XMP packet and display it.

Thanks for the great project! Have fun ;)
p.s: I saw the Jempbox project i might add support for my own schema in
there it would be nice... (My dex schema supports the origin URI which is
equivalent to your URI in BibTex)...



Date: 2007-02-07 05:37
Sender: ccodereAccepting Donations


File Added: XMPUtil.java


Date: 2007-02-07 05:00
Sender: ccodereAccepting Donations


I have checked the source code, used the latest version and checked if i
could read the XMP data with my own personally developed tools... And it
works perfectly, I really don't think you need my help, it seems to be
perfect.

OTOH, small suggestion: Please consult the following pages on how to use
the dc:identifier field, even though not official, i suggest you also use
this encoding:

http://www.ukoln.ac.uk/metadata/dcmi/dc-identifiers/2002-01-07/

In summary:
DOI Number -> dc:identifier is URI Non-registered (IANA) should be encoded
as: doi:DOI Number
ISBN Number -> dc:identifier is URN Registered (IANA) should be encoded as
URN:ISBN Number
ISSN Number -> dc:identifier is URN Registered (IANA) should be encoded as
URN:ISSN Number

I will try to send you a patched version of XMLUtils which fixes
implements these... Please validate them though, as i will only test them
rapidly ...!



Date: 2007-02-06 22:27
Sender: coezbek


Hi Carl,
I have added export to dublin-core in JabRef-2.2. Please give it a spin
and suggest further improvements.

If you have some experience with DublinCore, it would be absolutely
awesome if you could give the XMPUtil class a closer look. You can find it
in

src/java/net/sf/jabref/utils/

Any help with making XMP support in JabRef first-grade would be highly
appreciated.

Greetings,
Christopher


Log in to comment.




Attached File ( 1 )

Filename Description Download
XMPUtil.java Patched XMP manager... Download

Changes ( 5 )

Field Old Value Date By
File Added 214834: XMPUtil.java 2007-02-07 05:37 ccodere
status_id Pending 2007-02-07 05:00 ccodere
close_date 2007-02-06 22:27 2007-02-07 05:00 ccodere
status_id Open 2007-02-06 22:27 coezbek
close_date - 2007-02-06 22:27 coezbek