W3C home > Mailing lists > Public > public-media-annotation@w3.org > September 2011

WG: XSLT-Stylesheet for conversion XMP > MAWG

From: Höffernig, Martin <Martin.Hoeffernig@joanneum.at>
Date: Mon, 19 Sep 2011 14:00:03 +0200
To: "public-media-annotation@w3.org" <public-media-annotation@w3.org>, Felix Sasaki <felix.sasaki@dfki.de>
Message-ID: <CD9846F872C7874BB4E0FDF2A61EF09FCA36ADCCF1@RZJC1EX.jr1.local>
Dear Felix,

Since time-related datatypes  have to be used for date properties [1], I would choose option 1 or option 2.
[1]: http://www.w3.org/2008/WebVideo/Annotations/drafts/ontology10/PR/#ont-owl

Best,
Martin

Von: felix.sasaki@googlemail.com [mailto:felix.sasaki@googlemail.com] Im Auftrag von Felix Sasaki
Gesendet: Donnerstag, 15. September 2011 16:42
An: Höffernig, Martin
Cc: public-media-annotation@w3.org
Betreff: Re: XSLT-Stylesheet for conversion XMP > MAWG

Hi Martin,

this is on purpose. The values are not part of the lexical space of time-related datatypes in XML Schema. At
http://fabday.fh-potsdam.de/~sasaki/mawg/xmp/xslt/xml-canonical-to-mawg.xsl
I have created a function my:checkDateDataType, to deal with some of these values that occurred in the XMP example files. But the fallback is always that if no adequate datatype is found, I go to a string datatype.

I see two options:

1) Change the input data so that it contains only XML Schema dateTime compatible values
2) Extend my:checkDateDataType so that it deals with the two problematic values and converts them to dateTime values
3) Leave everything as is

My preference would be 2) or 3). 3) Would reflect reality most, at least given the XMP examples I have. Any thoughts?

Felix

2011/9/15 Höffernig, Martin <Martin.Hoeffernig@joanneum.at<mailto:Martin.Hoeffernig@joanneum.at>>
Dear Felix,

your example looks very good now.
However, there is one thing left I would like to point your attention to:
There are two ma:date values with datatype xsd:string ("2010-12-13T12:59:59.12345<tel:59.12345>+02:00"^^xsd:string and "2010-12-31T11:59Z"^^xsd:string). A time-related datatype would be better at this point.

Best,
Martin


Von: felix.sasaki@googlemail.com<mailto:felix.sasaki@googlemail.com> [mailto:felix.sasaki@googlemail.com<mailto:felix.sasaki@googlemail.com>] Im Auftrag von Felix Sasaki
Gesendet: Dienstag, 13. September 2011 15<tel:2011%2015>:11
An: Thierry Michel; Höffernig, Martin; public-media-annotation@w3.org<mailto:public-media-annotation@w3.org>

Betreff: Re: XSLT-Stylesheet for conversion XMP > MAWG

Thierry, Martin, all,

I implemented Martin's comments:
- input file with only one creation date value http://fabday.fh-potsdam.de/~sasaki/mawg/xmp/xslt/MAWG_test_canonical.xml
- XSLT transformation that solves the issues with multiple labels for ma:hasLanguage and ma:hasPolicy http://fabday.fh-potsdam.de/~sasaki/mawg/xmp/xslt/xml-canonical-to-mawg.xsl
- output in RDF/XML http://fabday.fh-potsdam.de/~sasaki/mawg/xmp/xslt/MAWG_test_canonical_mapped-to-rdf-by-xslt.xml
- output in TTL http://fabday.fh-potsdam.de/~sasaki/mawg/xmp/xslt/MAWG_test_canonical_mapped-to-rdf-by-xslt.ttl

Regards,

Felix
2011/9/12 Felix Sasaki <felix.sasaki@dfki.de<mailto:felix.sasaki@dfki.de>>
Sorry, Thierry, I don't remember the discussion with Werner  in detail. There is no label and no bag in

> ma:date "2009"^^xsd:gYear,
> "2010-11-18T11:57:24+01:00"^^xsd:dateTime,
> "2010-12<tel:2010-12>"^^xsd:gYearMonth,
> "2010-12-13T12:59:59.12345<tel:59.12345>+02:00"^^xsd:string,
> "2010-12-23T14:42:31"^^xsd:dateTime,
> "2010-12-31<tel:2010-12-31>"^^xsd:date,
> "2010-12-31T11:59:00"^^xsd:dateTime,
> "2010-12-31T11:59Z"^^xsd:string;

these are plain RDF statements. Could you clarify how to deal with the above?

Thanks,

Felix

2011/9/12 Thierry Michel <tmichel@w3.org<mailto:tmichel@w3.org>>
Felix,


It is the same issue for all;
This was already discussed with Werner.
you should use  multiple time the same labels or avoid bags in your example.


Thierry
> Dear Martin, all,
>
> in addition to your observations and suggestions - which all make sense -,
> let me mention that there are other cases of multiple examples of the same
> kind of information:
> * ma:date:
> ma:date "2009"^^xsd:gYear,
> "2010-11-18T11:57:24+01:00"^^xsd:dateTime,
> "2010-12<tel:2010-12>"^^xsd:gYearMonth,
> "2010-12-13T12:59:59.12345<tel:59.12345>+02:00"^^xsd:string,
> "2010-12-23T14:42:31"^^xsd:dateTime,
> "2010-12-31<tel:2010-12-31>"^^xsd:date,
> "2010-12-31T11:59:00"^^xsd:dateTime,
> "2010-12-31T11:59Z"^^xsd:string;
> * ma:title:
> ma:title "dc:title (de-DE)"^^xsd:string,
> "dc:title (en-US)"^^xsd:string,
> "dc:title (x-default)"^^xsd:string,
> "xmpDM:album"^^xsd:string.
> In XMP it is clear where the information is coming from, e.g. from dublin
> core within XM Por the xmpDM namespace. The mawg values of course don't
> express provenance. So an application does not know whether it should
> choose
> e.g.  2010-11-18T11:57:24+01:00 or 2009 is the date value.
>
> I don't know how to solve this problem but just wanted you to be aware of
> it. Martin, all, please let me know how you want to proceed, for
> ma:hasLanguage, ma:hasPolicy, ma:date, ma:creationDate and ma:title.
>
> Regards,
>
> Felix
>
> 2011/9/12 Höffernig, Martin <Martin.Hoeffernig@joanneum.at<mailto:Martin.Hoeffernig@joanneum.at>>
>
>> Dear Felix, Thierry, all
>>
>> I have finished the recheck of the latest XMP example.
>> This example has passed the consistency check using Protege and Pellet.
>>
>> However, I have some remarks which should be discussed.
>>
>> The value of ma:hasLanguage is a rescource containing three different
>> labels:
>> <ma:hasLanguage rdf:parseType="Resource">
>>         <rdfs:label>en</rdfs:label>
>>         <rdfs:label>de-DE</rdfs:label>
>>         <rdfs:label>de-DE-bayrisch</rdfs:label>
>> </ma:hasLanguage>
>>
>> Since no cardinality restrictions has been defined in the MA ontology,
>> using multiple labels for the same resoure would be possible.
>> However, the question is if the above usage of mulitpe labels is correct
>> in
>> the MA ontology context?
>>
>> I would suggest to avoid the mulitple labels for the same resource and
>> describe multiple labels in the following way:
>> <ma:hasLanguage rdf:parseType="Resource">
>>         <rdfs:label>en</rdfs:label>
>> </ma:hasLanguage>
>> <ma:hasLanguage rdf:parseType="Resource">
>>         <rdfs:label>de-DE</rdfs:label>
>> </ma:hasLanguage>
>> <ma:hasLanguage rdf:parseType="Resource">
>>         <rdfs:label>de-DE-bayrisch</rdfs:label>
>> </ma:hasLanguage>
>>
>> Another resource (value of ma:hasPolicy) contains multiple labels
>> modelled
>> in the same as above.
>>
>> Furthermore, the media resource contains two different creation dates.
>> From my side of view there can be only one creation date.
>>
>> Best,
>> Martin
>>
>>
>> > -----Ursprüngliche Nachricht-----
>> > Von: Thierry MICHEL [mailto:tmichel@w3.org<mailto:tmichel@w3.org>]
>> > Gesendet: Donnerstag, 08. September 2011 20<tel:2011%2020>:03
>> > An: Felix Sasaki; Höffernig, Martin
>> > Cc: public-media-annotation@w3.org<mailto:public-media-annotation@w3.org>; Frank Biederich
>> > Betreff: Re: XSLT-Stylesheet for conversion XMP > MAWG
>> >
>> >
>> > Felix, Martin,
>> >
>> >
>> > I have added your 2 files (XML and RDF) into the testsuite.
>> > (replace the former one.
>> > Note that in the RDF I have added a comment as mentuionned in the
>> > guidelines.
>> >
>> > I have also added your Stylesheet for conversion of XMP (canonical
>> > form)
>> > to instance of "Ontology for Media Resources 1.0" in the list of
>> > Implementations/demonstrators
>> >
>> > see
>> >
>> > http://www.w3.org/2008/WebVideo/Annotations/drafts/ontology10/testsuite
>> > .html
>> >
>> >
>> > The last bit is that I have removed the former TTL file for XMP. Maybe
>> > you would like to provide a new one?
>> >
>> > Martin could you check the RDF file for XMP ?
>> >
>> > Thanks a lot for your input.
>> >
>> > Bets,
>> >
>> > thierry
>> >
>> > Le 08/09/2011 12:25, Felix Sasaki a écrit :
>> > > Thierry, all,
>> > >
>> > > I have made a slight update of the XMP examples. Attached ZIP file
>> > contains
>> > > a new input file MAWG_test_canonical.xml, and an output
>> > > file MAWG_test_canonical_mapped-to-rdf-by-xslt.xmp
>> > > can you replace input and output in the CR table for the ontology
>> > document
>> > > accordingly?
>> > >
>> > > The ZIP file also contains an XSLT stylesheet that I used for the
>> > generation
>> > > of the output file. Input needs to be XMP *in canonical form* (a
>> > certain
>> > > kind of XMP serialization). From my understanding an XMP processor
>> > needs to
>> > > be capable to produce that form. See the stylesheet also at
>> > > http://fabday.fh-potsdam.de/~sasaki/mawg/xmp/xslt/xml-canonical-to-
>> > mawg.xsl
>> > > Can you add the stylesheet (or maybe rather the link to it) to the
>> > mapping
>> > > tool list of implementations / demonstrators for the ontology? Many
>> > thanks
>> > > in advance.
>> > >
>> > >   I have put Frank Biederich from Adobe into the loop so that he is
>> > aware of
>> > > this.
>> > >
>> > > Regards,
>> > >
>> > > Felix
>> > >
>>
>
>
>
> --
> Prof. Dr. Felix Sasaki
> Senior Researcher, Language Technology Lab
> DFKI GmbH, Alt-Moabit 91c, 10559 Berlin, Germany http://www.dfki.de
> phone: +49-30-23895-1807<tel:%2B49-30-23895-1807> (fax: -1810)
> ------------------------------------------------
> Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
> Firmensitz: Trippstadter Strasse 122, D-67663 Kaiserslautern
> Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
> (Vorsitzender), Dr. Walter Olthoff
> Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
> Amtsgericht Kaiserslautern, HRB 2313
>
> Register for the W3C MultilingualWeb Workshop!
> Limerick, 21-22 September 2011
> http://multilingualweb.eu/**register <http://multilingualweb.eu/register>
>


--
Thierry Michel
W3C



--
Prof. Dr. Felix Sasaki
Senior Researcher, Language Technology Lab
DFKI GmbH, Alt-Moabit 91c, 10559 Berlin, Germany http://www.dfki.de<http://www.dfki.de/>
phone: +49-30-23895-1807 (fax: -1810)
------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Strasse 122, D-67663 Kaiserslautern
Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
(Vorsitzender), Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313

Register for the W3C MultilingualWeb Workshop!
Limerick, 21-22 September 2011
http://multilingualweb.eu/register




--
Prof. Dr. Felix Sasaki
Senior Researcher, Language Technology Lab
DFKI GmbH, Alt-Moabit 91c, 10559 Berlin, Germany http://www.dfki.de<http://www.dfki.de/>
phone: +49-30-23895-1807 (fax: -1810)
------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Strasse 122, D-67663 Kaiserslautern
Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
(Vorsitzender), Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313

Register for the W3C MultilingualWeb Workshop!
Limerick, 21-22 September 2011
http://multilingualweb.eu/register




--
Prof. Dr. Felix Sasaki
Senior Researcher, Language Technology Lab
DFKI GmbH, Alt-Moabit 91c, 10559 Berlin, Germany http://www.dfki.de<http://www.dfki.de/>
phone: +49-30-23895-1807 (fax: -1810)
------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Strasse 122, D-67663 Kaiserslautern
Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
(Vorsitzender), Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313

Register for the W3C MultilingualWeb Workshop!
Limerick, 21-22 September 2011
http://multilingualweb.eu/register
Received on Monday, 19 September 2011 12:00:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 19 September 2011 12:00:50 GMT