See also: IRC log
<YvesS> Yves +present
http://lists.w3.org/Archives/Public/public-i18n-its-ig/2014Apr/0008.html
https://www.w3.org/International/its/ig/track/issues/3
The set of namespace declarations are those in scope on the element that has the attribute in which the expression occurs. This includes the implicit declaration of the prefix xml required by the XML Namespaces Recommendation; the default namespace (as declared by xmlns) is not part of this set.
http://www.w3.org/International/its/ig/its20-errata.html
<dF> just joining the goto
felix: propose to add this as a non-normative errate - comments?
https://www.w3.org/International/its/ig/track/issues/3
<YvesS> +1 for adding the item to the errata
<dF> +1
<daveL> +1
<Ankit> +!
<Ankit> +1
<scribe> ACTION: felix to send a mail about errtaum and make the update within one week [recorded in http://www.w3.org/2014/04/30-i18nits-minutes.html#action01]
<trackbot> Created ACTION-43 - Send a mail about errtaum and make the update within one week [on Felix Sasaki - due 2014-05-07].
felix describing feedback from the EU
<YvesS> Several data categories to review: would be nice to get some feedback. Felix asked: schema for using ITS in XLIFF needed? -> yes, very much so.
<YvesS> yes
XLIFF 2.0 mapping needs feedback
felix proposes to write ITS schema for using ITS in XLIFF
david: wonders who should host
the schema
... we need a specialized schema, no doubt
... need to see who is the host
... should be one of the topics of XLIFF TC meeting in May and
in June during FEISGILTT session
<YvesS> sure
david: can be handled during ITS mapping session during FEISGILTT
<scribe> ACTION: felix to follow up on ITS in XLIFF schema [recorded in http://www.w3.org/2014/04/30-i18nits-minutes.html#action02]
<trackbot> Created ACTION-44 - Follow up on its in xliff schema [on Felix Sasaki - due 2014-05-07].
<dF> XLIFF 2.1 seesion June 4 afternoon
<dF> I am
<scribe> ACTION: dF to review XLIFF 2.0 mappings [recorded in http://www.w3.org/2014/04/30-i18nits-minutes.html#action03]
<trackbot> Created ACTION-45 - Review xliff 2.0 mappings [on David Filip - due 2014-05-07].
david: yesterday submitted XLIFF candidate ticket
https://www.w3.org/International/its/wiki/ITS-RDF_mapping
<dF> COS - candidate OASIS standard
dave: work in FALCON
project
... will send URL to the list
... we put a deliverable togehter. Working primarily with xtm.
Extended work we did with XLIFF > ITS mapping into RDF. With
where we maintain ITS > NIF mapping and other ITS
parameters
... what we have done: we are pulling togehter an (RDF?) schema
that extends w3c provenacne model
... idea is: continue work we did with Ankit on re-training MT
components
... uses a lot of ITS metadata, extending what we did in
LT-Web
... now handling also @@@ packages that Interverbum uses
... e.g. how to handle terminology markup - xtm are doing this
but so far they did not use its:terminology
... are looking into how to also bring that into interop
now
[missed dF comment]
<YvesS> Will that RDF mapping available publicly (soon)?
<dF> I said that both XLIFF 1.2 and 2.0 ITS mappings are usning mrk mark up
dave: will put the RDF on github
<YvesS> Great, thanks.
dave: hopefully this week
<dF> so the reconciliation should not be difficlut, just using its attributes for what does not exist in XLIFF core
dave: there haven't been a lot of changes, there seem to be changes to NIF ontology
http://persistence.uni-leipzig.org/nlp2rdf/specifications/validate.jar
http://persistence.uni-leipzig.org/nlp2rdf/
dave: will not change a lot, just
more usage examples
... plan to have a demo for FEISGILTT
... will do similar like LT-Web: put tooling up, reference
architecture, then input files as example how to work with
this
chriLi: W3C provenance model is target towards content, not processing? so it does not allow you to track: this is component that has been used in this chain: ...
dave: yes
... we in FALCON do a component factory, which can create
components
... old ideas from Corba etc.
discussion about components, versions, ...
dave: have not considered this in detail, good points
christian: suggestion: how to
approach quest for standard direction. That also could be
answered from industry, community services
... asking them: what are the hot service component
architectures? what to apply for processing chains
... going out asking different communities might prevent us
from building and ivory tower
dave: really good point
... where would be a good forum for that?
... you won't get a lot of service people in feisgillt or MLW
WS
christian: talking to other EU funded projects?
dave: I tracked some projects
previously, will check again - OASIS may not use semantic
service descriptions
... it is an interesting point, if you have input for that
(also your colleagues) please let us know, very interesting
also e.g. for LIDER
... important are the use cases, e.g. role of MT component in a
work flow
... because we also look into re-training of MT, that needs to
be managed as well
... I'll check with Klaus
christian: we had tool information related discussion in ITS
felix: have that with annotators-ref attribute: we did not specify our own thing but just the URI so that people can do what is right for them
<YvesS> sure
http://dbpedia.org/resource/Eye
its-ta-ident-ref
<YvesS> yes more or less
<YvesS> The code is working as of few minutes ago thanks to your sparql help.
felix explains workflow: getting cross lingual info from dbpedia
ankit: we did something simliar
during LT-Web
... we manually put dbpedia link into ITS wrapper
... the MT step can then use the info
dave: the MT did the lookup of the target language term from dbpedia
<YvesS> more like annotation
dave: how would work in okapi: would that be done as part of the MT component?
<YvesS> and glossary
<YvesS> yes
dave: or would that be a tbx entry?
yves: yes
<YvesS> it seems the wikidata Q labels provide better "translations" though. but sparql can be used to get the Q id.
dave: makes sense, we had that in one of the scenarios
http://googleknowledge.github.io/qlabel/
<YvesS> that works too, but it requires one additional call per resource
dave: you can also do the same
with babelnet and rdf versions of wordnet
... bablenet will have translations but not so many named
entities
<YvesS> any links to babelnet? do thay have api?
dave: it is the same principle, sparql queries would be a bit different
yves, see here: http://babelnet.org/
<YvesS> thanks
dave: they should have a sparql endpoint
<YvesS> will try that
dave: or you can download the RDF
as well
... interesting also: babelnet also provides services that does
annotation, like enrycher
... so you can do named entity annotation funtionality too
<YvesS> good to know. for now I'm using dbpediaspotligh to get the entities
felix: having ability to do sparql queries in a tool like okapi woudl be great to easiyl adapt this to other data sources
dave: +1, and documenting the sparql queries would be great too, so that people see what queries work well
agree
<YvesS> agree
dave: in FALCON we are convervative - offering a very simple REST interface so that people don't have to worry about sparql
<scribe> ACTION: felix to set up a sparql query page for use with its-ta-ref [recorded in http://www.w3.org/2014/04/30-i18nits-minutes.html#action04]
<trackbot> Created ACTION-46 - Set up a sparql query page for use with its-ta-ref [on Felix Sasaki - due 2014-05-07].
felix: will advertise that to lider people to get their input on typical queries
<YvesS> bye
adjourne
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/during/during FEISGILTT/ Succeeded: s/DF/dF/ No ScribeNick specified. Guessing ScribeNick: fsasaki Inferring Scribes: fsasaki Present: felix DaveLewis Ankit dF yves cl Agenda: http://lists.w3.org/Archives/Public/public-i18n-its-ig/2014Apr/0008.html Got date from IRC log name: 30 Apr 2014 Guessing minutes URL: http://www.w3.org/2014/04/30-i18nits-minutes.html People with action items: df felix[End of scribe.perl diagnostic output]