RDF vocabularies for historic place-names and relations between them?
Jo Walsh
I've been lurking on this list awhile, since being pointed at it by Kate
Byrne and Ant Beck. Now I have a question about RDF ontologies for
historic place-name relations, and this seems like a good place to start
asking Linked Data headed archaeologists and deep historians.
We're starting work on a cool project - CHALICE - text-mining digitised
volumes from the English Place-Name Survey to produce a Linked Data
historic place-name gazetteer for some parts of the UK, with links
through to contemporary place-name sources (geonames, OS Research).
http://chalice.blogs.edina.ac.uk/2010/06/10/chalice-aims-objectives-outcomes/
A diagram of the map of relations we'll at a minimum extract from text:
http://chaliced.tumblr.com/post/698818105/will-this-diagram-help-14-06-10-this-is-to
My question - we'll need an RDF vocabulary to describe relations between
place-names that change over time.
Place-name <was previously known as> place-name
Place-name <was previously at location> location
or
[Place <was known as> Place] between at least 1257 until at 1673
Forgive the sketchiness of the modelling here - only just starting to
think about the best way to model the structures in RDF.
So we're looking for prior art - if there are ontologies out there that
express in part or whole, what we're looking for, then would much rather
re-use than mint our own.
If anyone here has a suggestion - or could point me in the direction of
someone who would be very likely to have a suggestion - I would really
appreciate that.
be well,
jo
--
Jo Walsh
Unlock places - http://unlock.edina.ac.uk/
phone: +44 (0)131 650 2973
skype: metazool
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
Leif Isaksen
Exciting project :-) You should certainly get in touch with Tom
Elliott and Sean Gillies (both of whom are on this list - poke poke)
who are behind the Pleiades Project:
It deals with multiple toponyms for classical locations (amongst many
other things). I don't know if they have exactly what you're looking
for they've done an awful lot of thinking in this space.
Best
Leif
Sean Gillies
We have a model for multiple toponyms and their associated time
periods, but – sorry to disappoint – it's an object model using
composition
http://sgillies.net/blog/1032/modeling-historical-places-for-pleiades/
not an RDF model. In the HTML representations of places, such as
http://pleiades.stoa.org/places/142847211/
there is a listing of names and their time periods, but no particular
semantics yet. I'm curious about whether our names can be evolved into
"episodes of appellation" (for lack of better term), which would let
us exploit CIDOC-CRM or HEML, but haven't taken any real steps in this
direction yet. I'm pretty sure that neither of those ontologies will
let you express
Place-name <was previously known as> place-name
so simply, though I am not at all a CRM expert and may be wrong.
Cheers,
--
Sean Gillies
Programmer
Institute for the Study of the Ancient World
New York University
Vladimir Ivanov
As far as I know there is a class "E44.Place Appellation" in the CIDOC CRM
and there is no temporal entity which describe any type of a process
or event of
creating/changing place appellations (attached to the E53.Place).
> In the HTML representations of places, such as
>
> http://pleiades.stoa.org/places/142847211/
>
> there is a listing of names and their time periods, but no particular
> semantics yet.
> We have a model for multiple toponyms and their associated time
> periods,
I think, there is a significant difference between toponyms and places.
Sean, could you explain what do you mean when associating toponyms and
time periods?
>>> Place-name <was previously known as> place-name
One might also have problems with formal definition of the phrase "known as".
--
Best regards,
Vladimir Ivanov
Kazan Federal University,
Russia
2010/7/27 Sean Gillies <sean.g...@gmail.com>:
--
С уважением,
Владимир Иванов
==============================================
Лаборатория математической и компьютерной лингвистики
НИИММ им. Н.Г. Чеботарева - cll.niimm.ksu.ru
Казанский (Приволжский) федеральный университет
--------------------------------------------------------------------------------------------
Центр оцифровки культурного наследия, ЦИТ РТ
+7 (843) 264-72-54
+7 (917) 270-03-12
e-mail: v...@tatar.ru
--------------------------------------------------------------------------------------------
Jo Walsh
can meet in the middle, just good to see some solid prior art.
One issue we have with the EPNS text is a fuzziness in some of the
relations - where we can't infer containment or identity - but a
relation of proximity will definitely have re-use value (such as, for
later disambiguation when we want to figure out which of N modern
Waltons, a historic Walton was). "Special case or fuzzy relations",
that's the ticket.
So we can work on coming up with an RDF representation that will work
for Pleiades as well as CHALICE, that would just rock. Thanks.
On 27/07/2010 13:55, Sean Gillies wrote:
> We have a model for multiple toponyms and their associated time
> periods, but � sorry to disappoint � it's an object model using
> composition
>
> http://sgillies.net/blog/1032/modeling-historical-places-for-pleiades/
>
> not an RDF model. In the HTML representations of places, such as
>
> http://pleiades.stoa.org/places/142847211/
>
> there is a listing of names and their time periods, but no particular
> semantics yet. I'm curious about whether our names can be evolved into
> "episodes of appellation" (for lack of better term), which would let
> us exploit CIDOC-CRM or HEML, but haven't taken any real steps in this
> direction yet. I'm pretty sure that neither of those ontologies will
> let you express
--
CARLISLE, Philip
It might be worthwhile talking to the Getty to see if they have any rdf plans for the Thesaurus of Geographic Names as that includes both foreign language and former names for places and this seems to be an excellent opportunity for reuse of their existing data.
Phil
Phil Carlisle
Data Standards Supervisor
English Heritage
National Monuments Record Centre
Kemble Drive
Swindon
SN2 2GZ
+44 (0)1793 414824
http://thesaurus.english-heritage.org.uk/
The information contained within this e-mail is confidential and may be privileged. It is intended for the addressee only. If you have received the e-mail in error, please inform the sender and delete it from your system. The contents of this e-mail must not be disclosed to anyone else or copied without the sender's consent.
Any views and opinions expressed in this message are those of the author and do not necessarily reflect those of English Heritage. English Heritage will not take any responsibility for the views of the author.
P Please do not print this e-mail unless you really need to
-----Original Message-----
From: caa-sema...@googlegroups.com [mailto:caa-sema...@googlegroups.com] On Behalf Of Sean Gillies
Sent: 27 July 2010 13:55
To: caa-sema...@googlegroups.com
Subject: Re: {CAA Semantic SIG} RDF vocabularies for historic place-names and relations between them?
Hi Jo, Leif,
We have a model for multiple toponyms and their associated time periods, but - sorry to disappoint - it's an object model using composition
Jo Walsh
> It might be worthwhile talking to the Getty to see if they have any
> rdf plans for the Thesaurus of Geographic Names as that includes both
> foreign language and former names for places and this seems to be an
> excellent opportunity for reuse of their existing data.
Thanks Philip, that would be interesting to see and to
(semi)automatically make the links to geonames and thence out to the
rest of the Linked Data web.
I (re)learned today that Google, via Leif, are just going to do it all
for us anyway - text mining Google Books to generate an ancient-historic
gazetteer -
http://www.alphagalileo.org/ViewItem.aspx?ItemId=81689&CultureCode=en
As long as the resulting data is open data and well linked, i suppose i
don't mind, but i feel chagrin, is there anything we can do that Google
can't just buy? Over time, probably not, if they choose to do.
Leif Isaksen
creating the gazetteer ourselves - we'll be using Pleiades, Perseus
and GeoNames (and possibly FireEagle and the TGN) as our name lists.
The first two both provide URIs for each location so we're hoping that
the semantic aspects will come for free thanks to other people's hard
work! The purpose of GAP is really to enhance discovery - of places
within a text, and of texts related to a place.
Best
L.