LD4LT community group call

31 Jul 2014


See also: IRC log


DaveLewis, fsasaki, tflati, chriLi, Ali_H_Vahid, renat, rnavigli, jgacia, victor, serge, penny, Tcarrasco, martha, pennyl
Dave lewis


<daveL> Agenda: http://lists.w3.org/Archives/Public/public-ld4lt/2014Jul/0028.html

call intro

<daveL> scribe: tflati


<fsasaki> action-11?

<trackbot> action-11 -- David Lewis to Gather info on how to provide more detailed mapping from meta-share to dcat -- due 2014-07-24 -- OPEN

<trackbot> http://www.w3.org/community/ld4lt/track/actions/11

<daveL> http://lists.w3.org/Archives/Public/public-ld4lt/2014Jul/0033.html

<daveL> http://www.w3.org/TR/vocab-dcat/#vocabulary-overview

<tflati_> Dave: there are still 3 points left to discuss from the META-SHARE metamodel

<daveL> https://www.w3.org/community/ld4lt/wiki/DCAT_MetaShare_Mapping#Make_LanguageResource_a_DCAT_Dataset

<jgracia> dave: dataset corresponds closely with language resource

<tflati_> dave: some of some metashare properties could be replaced by some dcat properties

<tflati_> geographic coverage, time coverage, etc. they all have a corresponding property

<tflati_> penny: about making the language resource

<jgracia> penny: ms already have a distribution module

<tflati_> penny: for example the download URL, etc.

<tflati_> dave: how deep should the separation be made? Are you trying to encourage more than just separation?

<tflati_> penny: at the conceptual level, i think that usually there are different forms with different licenses

<jgracia> penny: frequently diferent distributions in MS are simply same resource under different licensees

<tflati_> victor: but they should both allow for distribution

<tflati_> penny: at least now there is a mix up with dublin core etc.

<tflati_> dave: we covered both the two points about DCAT

<Tcarrasco> Data access: record vs. database - URI local vs. remote (file and http schemes)

<tflati_> penny: i never used SKOS concepts, but I have it like for any classification schema

<tflati_> not sure how SKOS deals with different dimensions of classification

<tflati_> we had also a discussion with Marta about subclasses of Corpus and there is no consensus about lexicon or other things, depending several criteria

<tflati_> the linguality dimension (mono-, bi-, multi-lingual)

<tflati_> the media type of Corpus (audio, text, etc.)

<tflati_> one could end up with different taxonomies depending on the used criteria

<Tcarrasco> http://lists.w3.org/Archives/Public/public-ld4lt/2014Jul/0024.html

<tflati_> dave: there is not a good model of what a schema should be. Hard to say in a definitive way what is a subclass of what.

<tflati_> roberto: when we encoded babelnet, we had the issue to see various options. For us SKOS is okay, up to a certain extent

<tflati_> if you're interested in just a taxonomy SKOS is okay

<tflati_> otherwise if you want better relations, SKOS might be limited

<tflati_> we just decided to reduce the granularity of the relations

<tflati_> we kept the narrower and broader properties for the taxonomic aspect

<tflati_> and related for all other properties

<tflati_> jgracia: I think that META-SHARE does not need more expressive properties than those

<tflati_> Roberto: in this way we also limit variability

<tflati_> penny: SKOS should be okay

<tflati_> dave: moving this forward, do we want to separate

<tflati_> penny: text type, genre, video type would be perfect for the SKOS schema, we could add to them

<tflati_> dave: yes, i agree. it is part of the issue. There is an involving consensus for the definition of the core

<fsasaki> ACTION: penny to capture list of parameters that may be relevant for a skos classification [recorded in http://www.w3.org/2014/07/31-ld4lt-minutes.html#action01]

<trackbot> Created ACTION-12 - Capture list of parameters that may be relevant for a skos classification [on Penny Labropoulou - due 2014-08-07].

<tflati_> jgracia: the added value of using SKOS rather than OWL?

<tflati_> dave: you can use both

<tflati_> sometimes one is not sure whether something is strictly a subclass of what

<daveL> ACTION: pennyl to identify which properties and classes could move to a skos-based conceptual schema [recorded in http://www.w3.org/2014/07/31-ld4lt-minutes.html#action02]

<trackbot> Error finding 'pennyl'. You can review and register nicknames at <http://www.w3.org/community/ld4lt/track/users>.

<tflati_> jorge: narrower could be interpreted in a more loose sense, without being that strict

<jgracia> so more flexibility in SKOS not commiting to OWL semantics

meta-share discussion

<tflati_> thomas: the access of a record vs. the whole database

<tflati_> and also the aspect of the access to the schema

<tflati_> these two aspects have to be considered, also

<tflati_> dave: this should be included in the distribution module

<tflati_> because it depends on the media type you are using

<tflati_> for example if you have a TM you could use a fragment identifier to make the work easily

<tflati_> thomas: if data can be downloaded, the distribution aspect breaks up because from the application point of view everything is a URI

<tflati_> penny: this also connects to waht i was discussing with Marta: language resource taxonomy

<tflati_> the 4 classes of Corpus

<tflati_> and also a second dimension: the media of resource types

<tflati_> to MS these are not really subclasses, rather they are parts of a corpus

<tflati_> that is the reason we called text corpus

<tflati_> the idea is that you try to put information that describes better the dimension of the corpus into a text file

<tflati_> we might add another class, multilingualcorpus

<tflati_> the relation subclass should be ispartof, actually

<tflati_> should be "contains"

<tflati_> dave: in DCAT you have a Dataset which can have multiple distributions

<Tcarrasco> We should use the Media Types terminology http://www.iana.org/assignments/media-types/media-types.xhtml

<tflati_> a dataset if a collection of one or more resources

<tflati_> we might need to use "contains" or "collection"

<tflati_> of the dcterms collection

<tflati_> for container of different resources

<tflati_> marta: it was not my fault to introduce these classes

<tflati_> the parts should not be seen as collection, but just as a model of having different media types

<tflati_> not different files

<Tcarrasco> Multilingual Dataset Format (muset) - http://dragoman.org/muset

<tflati_> the corpus is something, how you physically store it is another thing

<tflati_> this has nothing to do with the implementation/format of your dataset

<tflati_> dave: terminology is important, one could easily misunderstanf

<tflati_> thomas: we should use the existing media type terminology

<tflati_> for example text would be "plain"

<tflati_> otherwise we have to reinvent it

<chriLi> I have to leave ...

<tflati_> we should build on top of IANA

<tflati_> penny: text is "text", not "plain"

<tflati_> in the upper level

<Tcarrasco> according to http://www.iana.org/assignments/media-types/media-types.xhtm is "plain text" called "plain" -

<Tcarrasco> "text" is a whole category - http://www.iana.org/assignments/media-types/media-types.xhtml#text

<tflati_> I have problems in my internet connection, could you scribe at my place for the next few moments, please?

<daveL> marta: describes strategy on mapping elements of original meta-share schema into RDF

<Tcarrasco> in http://tools.ietf.org/html/rfc5147 is called "text/plain"

<daveL> .. in response to vistors question on whther to include 'flat' license attributes in current model, or structure these in ODRL

<daveL> jorge: this is slightly different from victors questions, since ODRL can represent rules that cannot be represented from the current properties

<daveL> marta: asks difference between cc model and odrl

<tflati_> I am back, thanks dave

<daveL> victor: odrl is more expressive and extensible

<daveL> scribe: tflati

<tflati_> victor: maybe not complete, but extensible and flexible for sure

<jgracia> you have a nice example of ODRL & MS here https://www.w3.org/community/ld4lt/wiki/Licensing_information

<tflati_> victor: the same resource could have different permissions, depending on the distribution media

<tflati_> jorge: as for me it is complicated to represent this in ODRL (but only for the machine) and can be done only once

<tflati_> victor: i completely agree

<jgracia> but with content negotiation you can reach the simple readable information for humans


<fsasaki> http://lists.w3.org/Archives/Public/public-ld4lt/2014Jul/0034.html

<tflati_> dave: just one issue: thanks Tiziano and Roberto for the report on the META-SHARE through the group

<fsasaki> adjourned

Summary of Action Items

[NEW] ACTION: penny to capture list of parameters that may be relevant for a skos classification [recorded in http://www.w3.org/2014/07/31-ld4lt-minutes.html#action01]
[NEW] ACTION: pennyl to identify which properties and classes could move to a skos-based conceptual schema [recorded in http://www.w3.org/2014/07/31-ld4lt-minutes.html#action02]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014-07-31 14:08:27 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
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)

Found Scribe: tflati
Found Scribe: tflati
Present: DaveLewis fsasaki tflati chriLi Ali_H_Vahid renat rnavigli jgacia victor serge penny Tcarrasco martha pennyl
Agenda: http://lists.w3.org/Archives/Public/public-ld4lt/2014Jul/0028.html
Got date from IRC log name: 31 Jul 2014
Guessing minutes URL: http://www.w3.org/2014/07/31-ld4lt-minutes.html
People with action items: penny pennyl

[End of scribe.perl diagnostic output]