Re: mapping Bibtex in Bibo

63 views
Skip to first unread message

Jakob Voss

unread,
Dec 10, 2012, 11:59:45 AM12/10/12
to bibliographic-ontolog...@googlegroups.com
Hi,

I'd be interested in such a mapping too! Unfortunately BibTeX is much
less a simple format than it seems, so I'd start with bibutils to
convert BibTeX to MODS first:
http://sourceforge.net/p/bibutils/home/Bibutils/

Jakob


--
Verbundzentrale des GBV (VZG)
Digitale Bibliothek - Jakob Vo�
Platz der Goettinger Sieben 1
37073 Goettingen - Germany
+49 (0)551 39-10242
http://www.gbv.de
jakob...@gbv.de

Iman Gharib

unread,
Dec 11, 2012, 7:14:24 AM12/11/12
to bibliographic-ontolog...@googlegroups.com
thank you very much. I have started a project converting bibtex in Bibo and will put in internet as soon as it finishes (hopefully:)

Am Montag, 10. Dezember 2012 17:59:45 UTC+1 schrieb jakob...@gbv.de:
Hi,

I'd be interested in such a mapping too! Unfortunately BibTeX is much
less a simple format than it seems, so I'd start with bibutils to
convert BibTeX to MODS first:
http://sourceforge.net/p/bibutils/home/Bibutils/

Jakob


--
Verbundzentrale des GBV (VZG)
Digitale Bibliothek - Jakob Vo�

Bruce D'Arcus

unread,
Dec 11, 2012, 7:58:08 AM12/11/12
to Bibliographic Ontology Specification Group

This (say an XSLT stylesheet?) To convert MODS to BIBO would have the advantage that, in conjunction with bibutils, it would also allow conversion from all the other input formats bibutils supports.

In my experience, the tricky part of a good conversion is avoiding gratuitous blank nodes, and having good URIs.

On Dec 10, 2012 11:59 AM, "Jakob Voss" <Jakob...@gbv.de> wrote:
Hi,

I'd be interested in such a mapping too! Unfortunately BibTeX is much
less a simple format than it seems, so I'd start with bibutils to
convert BibTeX to MODS first:
http://sourceforge.net/p/bibutils/home/Bibutils/

Jakob


--
Verbundzentrale des GBV (VZG)
Digitale Bibliothek - Jakob Voß

Platz der Goettinger Sieben 1
37073 Goettingen - Germany
+49 (0)551 39-10242
http://www.gbv.de
jakob...@gbv.de

--
You received this message because you are subscribed to the Google Groups "Bibliographic Ontology Specification Group" group.
To post to this group, send email to bibliographic-ontolog...@googlegroups.com.
To unsubscribe from this group, send email to bibliographic-ontology-spec...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/bibliographic-ontology-specification-group?hl=en.

Christopher Gutteridge

unread,
Dec 11, 2012, 9:33:20 AM12/11/12
to bibliographic-ontolog...@googlegroups.com, Iman Gharib
That's always nice to hear!

A word of warning from the days of my youth, BibTeX files often don't follow the standard. For example, certain characters should be escaped but generally are not.

Don't be afraid to drop a bit of information rather than map it incorrectly. You can always created an additional field in the RDF data which summarises anything that couldn't be mapped.
--
You received this message because you are subscribed to the Google Groups "Bibliographic Ontology Specification Group" group.

To post to this group, send email to bibliographic-ontolog...@googlegroups.com.
To unsubscribe from this group, send email to bibliographic-ontology-spec...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/bibliographic-ontology-specification-group?hl=en.

-- 
Christopher Gutteridge -- http://id.ecs.soton.ac.uk/person/1248 

/ Lead Developer, EPrints Project, http://eprints.org/
/ Web Projects Manager, ECS, University of Southampton, http://www.ecs.soton.ac.uk/
/ Webmaster, Web Science Trust, http://www.webscience.org/


Christopher Gutteridge

unread,
Dec 11, 2012, 9:33:51 AM12/11/12
to bibliographic-ontolog...@googlegroups.com, Bruce D'Arcus
+1 for the avoiding blank nodes.

A colleague of mine runs co-referencing services which try to identify authors of works by context... it makes educated guesses based on what it already knows, but to link the data the authors on a paper need a URI.

In EPrints, for unknown authors we use the person name plus record ID and hash it. That way if you reorder the authors it keeps the URI for them.

Jakob Voss

unread,
Dec 11, 2012, 10:09:36 AM12/11/12
to bibliographic-ontolog...@googlegroups.com
Bruce D'Arcus wrote:

> In my experience, the tricky part of a good conversion is avoiding
> gratuitous blank nodes, and having good URIs.

I cannot imagine how to get good URIs without making use of authority
files [*]. By this a X-to-Bibo-Mapping is not a simple mapping anymore
but a heuristic that needs to query other data sources. I'd better
separate the task of mapping URIs and the task of mapping a Bibtex (or
MODS) record structure to a RDF graph structure. The original question
was primarily about the first.

I'd be happy about a clean solution that simply maps BibTeX to real RDF
in Bibo instead of yet another ad-hoc-created private ontology and I
prefer blank nodes for authors, publishers, series etc. instead of yet
another ad-hoc-created set of private URIs.

Jakob

[*] Some general authority files that would make sense to match with:

authors: VIAF
journals: Zeitschriftendatenbank (which already uses Bibo by the way)
publishers: DBPedia at least for the most known
series: ?



--
Verbundzentrale des GBV (VZG)
Digitale Bibliothek - Jakob Vo�

Christopher Gutteridge

unread,
Dec 11, 2012, 10:38:24 AM12/11/12
to bibliographic-ontolog...@googlegroups.com, Jakob Voss, Hugh Glaser
If you don't assign URIs to elements, they can't be used as part of
linked data. This is something I've had users of my own data feed back.

We currently consider it best practice to assign a "local" URI to all
elements of a bibliographic record, if an a more authoritative URI is
not available.

If you can't have a "good" URI, then an unresolvable one on your own
namespace is better than nothing as it can still be used to say "this
author on this paper is the same as this paper on this other system".

On 11/12/12 15:09, Jakob Voss wrote:
> Bruce D'Arcus wrote:
>
>> In my experience, the tricky part of a good conversion is avoiding
>> gratuitous blank nodes, and having good URIs.
> I cannot imagine how to get good URIs without making use of authority
> files [*]. By this a X-to-Bibo-Mapping is not a simple mapping anymore
> but a heuristic that needs to query other data sources. I'd better
> separate the task of mapping URIs and the task of mapping a Bibtex (or
> MODS) record structure to a RDF graph structure. The original question
> was primarily about the first.
>
> I'd be happy about a clean solution that simply maps BibTeX to real RDF
> in Bibo instead of yet another ad-hoc-created private ontology and I
> prefer blank nodes for authors, publishers, series etc. instead of yet
> another ad-hoc-created set of private URIs.
>
> Jakob
>
> [*] Some general authority files that would make sense to match with:
>
> authors: VIAF
> journals: Zeitschriftendatenbank (which already uses Bibo by the way)
> publishers: DBPedia at least for the most known
> series: ?
>
>
>

--
Reply all
Reply to author
Forward
0 new messages