Hi,
TL;DR: Did anybody consider using Wikidata items of Wikipedia templates to
store multilingual template parameters mapping?
Full explanation:
As in many other projects in the Wikimedia world, templates are one of the
biggest challenges in developing the ContentTranslation extension.
Translating a template between languages is tedious - many templates are
language-specific, many others have a corresponding template, but
incompatible parameters, and even if the parameters are compatible, there
is usually no comfortable mapping. Some work in that direction was done in
DBpedia, but AFAIK it's far from complete.
In ContentTranslation we have a simplistic mechanism for mapping between
template parameters in pairs of languages, with proof of concept for three
templates. We can enhance it with more templates, but the question is how
much can it scale.
Some templates shouldn't need such mapping at all - they should pull their
data from Wikidata. This is gradually being done for infoboxes in some
languages, and it's great.
But not all templates can be easily mapped to Wikidata data. For example -
reference templates, various IPA and language templates, quotation
formatting, and so on. For these, parameter mapping could be useful, but
doing this for a single language pair doesn't seem robust and reminds me of
the old ways in which interlanguage links were stored.
So, did anybody consider using Wikidata items of templates to store
multilingual template parameters mapping?
--
Amir Elisha Aharoni · אָמִיר אֱלִישָׁע אַהֲרוֹנִי
http://aharoni.wordpress.com
“We're living in pieces,
I want to live in peace.” – T. Moore
Hi, I got an open question about Wikidata concepts, partly related to the
idea of selecting a templates wrt. a query for placeholder articles.
One question about this idea is : what to do when several templates are
possible for an item, for example the item with no article is in the result
set of several queries associated with article stubs templates, say:
* the query "anything", that could be associated with a totally generic
templates that shows a Wikibase page like article templates that shows all
the claims about this item
* a more specific query "living organism"
* another even more specific query like "animal"
* ...
In this example each more specific query results is obviously a subset of
each more generic one. In such cases it could be useful to choose the
template of the most specific one.
In the same spirit of the "subclass of" property we can create (or reuse
it) for the queries. But as no property has in Wikibase itself a meaning,
this means the choice of the template would not be possible using raw
Wikibase concepts, which partly breaks the interests of the idea.
Any thoughts about this problem ?
Cheers, TomT0m
Having followed Freebase and the announcement about migrating to
Wikidata, I'm trying to get up to speed on the structure of Wikidata. I
read on the site that relationships such as subclass-of and instance-of
are managed by everyone. Looking at automobile (Q1420) I see that it is
both subclass-of and instance-of motor road vehicle, which I imagine is
not correct. Are there processes in place to manage the integrity of
these structural components of Wikidata?
Thanks,
James Weaver
Hey folks :)
People have been bugging me for a while now about this so I started a
page to gather input.
If you are searching for a topic on a Wikipedia for example but it
doesn't have an article about it then we can look at Wikidata and see
if there is a matching item for that topic. If we have data about it
on Wikidata then we can show some kind of placeholder. But how should
this look like and work? I'd love your input at
https://www.wikidata.org/wiki/Wikidata:Article_placeholder_input
Please keep in mind the important points at the beginning of that
page.
Also: If someone comes up with a better name for this that'd be awesome.
<3
Cheers
Lydia
--
Lydia Pintscher - http://about.me/lydia.pintscher
Product Manager for Wikidata
Wikimedia Deutschland e.V.
Tempelhofer Ufer 23-24
10963 Berlin
www.wikimedia.de
Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e. V.
Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg
unter der Nummer 23855 Nz. Als gemeinnützig anerkannt durch das
Finanzamt für Körperschaften I Berlin, Steuernummer 27/681/51985.
I would like to thank all for very informative feedback. Apologies for the term "CIDAR" in my original email as I meant "CIDOC," which has been suggested by Dov Winer (and also by a CIDOC contact - see email below). Let's say CIDOC were used, what would be required to harmonize the CIDOC structure and nomenclature with WikiData conventions? Thanks also to Daniel Mietchen and his feedback and interest in learning more about what may be required to make this happen. I am including email below that provides information on CIDOC from Martin Doerr. Jeff Thompson raised the issue of CIDOC (an ISO standard) is behind a paywall, and I do not know what issues this raises. However, it touches on something the WD4R project will need to address - how to incorporate reference to, use of, and possible access to, valuable research information that is subscription-based (eg. Nature, Science). Has there been any consideration for two-tiered access to WD4R - a free basic access and a subscription access that may include access to Science, Nature and other valuable resources? Best regards - Sam
--- The following note from Martin Doerr at ICS.forth.gr <CIDOC support> ---
From: martin [mailto:martin@ics.forth.gr]
Sent: Monday, December 29, 2014 5:46 AM
To: Sam Smith
Subject: Re: CIDOC or PAPYRUS as an Ontology for Historical Information
Dear Mr. Smith,
There are dozens or may be hundreds of CIDOC CRM extensions. Most do not come to our attention <This is in regard to the PAPYRUS history ontology project, that appears to be unused>. Many are created because people do not take their time to understand the concepts in depth. If they come in contact with us, we do everything to provide good consulting.
Here, the creators extended the ontology with what we call "terminology", i.e., classes which do not introduce new relevant relationships in order to connect things to facts. They are just for classification, such as "forest".
That does not make a "history ontology" in the proper sense. In confuses geographic classification and others with the core notion of history.
We recommend to keep a system of concepts for classification, albeit a "formal ontology", separate from the ontology that provides relationship semantics. To our understanding, the CIDOC CRM has a fairly complete coverage of history in the mechanical sense. By coverage we mean to provide generalizations that cover the phenomena in the domain of interest. Specializations may elaborate general relationships into more specific patterns of behavior. For instance, Steffen Hennicke from the Humboldt University in Berlin is working on an extension of CRM to detail into things like political activities and archival recording, which introduces two or three new classes such as "expression of will".
So, I'd recommend you just use CIDOC CRM as is, and combine via "P2 has type" with adequate vocabularies. If you like, I can subscribe you to crm-sig mailing list, then you can discuss directly with all experts. You are also kindly invited to join our next meeting in Oxford, Feb 9-12, and shortly present your project. Please do not hesitate to ask me any questions. As first reading, look at:
http://www.cidoc-crm.org/cidoc_tutorial/index.html (better take the three hours time to see this!)
and then recommendations on:
http://www.cidoc-crm.org/comprehensive_intro.html
Best season greetings,
Martin
hi,
i was wondering what the most simple way to query multiple categories
and intersections. the most simple application for e.g. GLAM is:
search museums of modern art in the city of zürich. if i do not find
enough i want to see all museums in the city of zürich. how and where
i would query that information?
another example would be barack obama:
https://en.wikipedia.org/wiki/Barack_Obama, which has the categories:
politicians from chicago, politicians from hawei, afro american
politicians, afro american christians, afro american non fiction
writers, and so on. basically all of them include:
* location
* profession
* race
* religion
* awards
in some combination. l was wondering how these categories can be
matched with what is stored in wikidata: www.wikidata.org/wiki/Q76,
or, e.g. query wikidata for afro-american politicians from hawei.
happy new year btw :)
rupert
Hey folks :)
If you're at 31C3 in Hamburg at the moment: There will be a Wikidata
meetup tomorrow. Say hi to some other cool Wikidata folks.
http://events.ccc.de/congress/2014/wiki/Session:Wikidata:_The_free_knowledg…
Cheers
Lydia, who would love to be there now
--
Lydia Pintscher - http://about.me/lydia.pintscher
Product Manager for Wikidata
Wikimedia Deutschland e.V.
Tempelhofer Ufer 23-24
10963 Berlin
www.wikimedia.de
Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e. V.
Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg
unter der Nummer 23855 Nz. Als gemeinnützig anerkannt durch das
Finanzamt für Körperschaften I Berlin, Steuernummer 27/681/51985.