See also: IRC log
<kerry> rragent, make logs public
<kerry> scribe: frans
<kerry> scribenick:frans
<kerry> scribenick: frans
<Linda> Thanks Kerry!
<joshlieberman> webex password?
<ChrisLittle> v4qtEh56
<scribe> chair:kerry
<scribe> chair: Kerry
<scribe> chair: kerry
<joshlieberman> Thanks ChrisLittle
<KJanowicz> Hi Kerry, can I very briefly talk 2 min about the workshop proposal?
PROPOSED: Accept last week's minutes: https://www.w3.org/2016/01/13-sdw-minutes
<jtandy> +1
<ClemensPortele> +1
+1
<kerry> +1
<KJanowicz> +1
<ahaller2> I was attending last week, probably forgot present
<ChrisLittle> +1 minutes
<Linda> +1
<MattPerry> +1
<ahaller2> no problem
<ahaller2> +1
RESOLUTION: accept last weekÅ› minutes
<kerry> patent call: https://www.w3.org/2015/spatial/wiki/Patent_Call
kerry: no comments on patent call
KJanowicz: SSN workshop will be held
KJanowicz organizes together with Josh
kerry: where and when?
<ahaller2> \ Can you post the link?
<KJanowicz> http://giscience.geog.mcgill.ca/
KJanowicz: Montreal 27 september
- 30 september
... please send in ideas for the workshop
kerry: I made a draft
agenda
... copying the agenda in the IRC, please have a look
<kerry> http://www.w3.org/2015/spatial/wiki/Agenda_F2F3
kerry: people who are not going:
please look at the times
... enough presence of SSN people is questioned
<SimonCox> I'll be a bit hit and miss on the wire
comments can be added to the wiki
<joshlieberman> Later in the day is certainly better for those in the US east.
ChrisLittle: need more time
<KJanowicz> can somebody post the wiki link to the list of participants for the F2F?
<AndreaPerego> Participants: https://www.w3.org/2015/spatial/wiki/Attending_F2F3
<ahaller2> sounds good
jtandy: BP editors have some
topics for the agenda
... they will ask for real world examples
kerry: chris, please make a
suggestion for a time slot for time
... frans, please place request for UCR time on the wiki
<kerry> ack
Linda: start and end times?
kerry: I think it depends on the venue
Linda: I have the key so anything is possible
kerry: nine to six?
... and a lunch break
<ChrisLittle> +1 to 9-6 and lunch and coffees
<AndreaPerego> +1 from me
<ChrisLittle> * any hagelslaght?
ClausStadler: not certain of attendence yet
kerry: please put attendence on the wiki
kerry: objection to SSN was
complexity
... it is common to just use a bit of the vocabulary
... do people have ideas on good modularization?
<KJanowicz> +1
kerry: modularisation may be a topic in coverage and time as well
ahaller: horizontal and vertical modularization
<SimonCox> Good modularization: limit external dependencies
KJanowicz: event model (which?) is a good example
kerry: please post a link of the example
<KJanowicz> http://semanticweb.cs.vu.nl/2009/11/sem/
<KJanowicz> see, for instance, http://semanticweb.cs.vu.nl/2009/11/img/Fig1.jpg
<KJanowicz> +1
ahaller2: generally agrees, but there should not be too many modules
<ahaller2> ... and put all upper level relations in one module
kerry: what is modularisation? Is it multiple files?
aharth: different namespaces are probably needed
<ahaller2> +1
<KJanowicz> This is how we addressed the problem in Geolink: http://schema.geolink.org/
joshlieberman: imports could be a problem
<KJanowicz> +1
kerry: the point is being able to use just a bit, without bother from other stuff
<KJanowicz> it is more than that
<SimonCox> Namespace issue has nasty interaction with linked-data assumptions
joshlieberman: perhaps it is just a question of documentation and visualization
<KJanowicz> +1
ahaller2: memory consumption on lightweight devices is a factor
<SimonCox> disagree with @ahaller regarding motivation for modularization.
<KJanowicz> also logical implications
(I am sorry, I did not get the message)
<SimonCox> agree with KJanowicz - need to be able to remove unintended/unfortunate entailments
<KJanowicz> modularization also important to avoid strong 'ontological commitments'
kerry: no clear solution yet, but worthwhile to hear different views
ClausStadler: what about the data cube vocab?
<SimonCox> Yes - overlap with several other ontologies. OBOE, BCO
kerry: SSN predates data cube vocabulary
<SimonCox> BCO is aligned with BFO, so takes fundamental continuant/occurrent split as a starting point
kerry: not sure whether data cube affects modularization
<KJanowicz> also: avoid problems like whether we allow for virtual sensors or now
ahaller2: GitHub can be used, but webprotege has limitations
<KJanowicz> Why not use DL functional syntax first?
ahaller2: suggestions for online tools?
<ClausStadler> just one note: with git it would be possible to add e.g. a pre-commit hook which prevents being able to commit a non-validating file, but not sure if existing validators would fit your needs
jtandy: phil has looked around
how others do it
... has one person been responsible. or were there more
editors?
<KJanowicz> Lets not be driven by tools too early in the process
KJanowicz: start simple, like pen and paper
<KJanowicz> https://www.yworks.com/products/yed
Linda: in a Dutch project no
online solution was found
... Topbraid and subversion was used in the end
<SimonCox> For collaborative editing/merging need effective diff mechanism!
<KJanowicz> Agreed but this is what you would have yed and other concept maps for
ahaller2: warns against shutting out people
<KJanowicz> the same will happen in protege, the axioms require an understanding of OWL
<SimonCox> Its not Turtle that is the issue, it is RDF!
ahaller: use webprotege only for classes and properties
<joshlieberman> It seems possible to work on both a graph level and an axiomatic level. The graph level is more accessible.
<KJanowicz> agree with Josh
joshlieberman: just the graph view could make the ontology more accesible
<SimonCox> How to ensure graph/axiom views are correctly synced?
<KJanowicz> We could use VOWL for that
kerry: axioms are probably the most of the work
<KJanowicz> http://vowl.visualdataweb.org/protegevowl.html
kerry: what is VOWL?
<joshlieberman> There is time-zone specific humor, apparently.
<ahaller2> the problem is with the Protege desktop version is that it will always scramble the source code
<SimonCox> +1 ahaller2
aharth: no strong opinion, no magic solution either
<SimonCox> Would be good if Protege joined the 21st century and used turtle instead of rdf/xml
kerry: afraid that this topic is not going to be resolved either
<Linda> Why not Topbraid then?
<KJanowicz> This is merely a serialization
<SimonCox> Canonical serialization ?
<ahaller2> does any23 produce a reliable serialisation?
<kerry> ACTION: ahaller2 to look at tools for ontology editing [recorded in http://www.w3.org/2016/01/20-sdw-minutes.html#action01]
<trackbot> Created ACTION-133 - to look at tools for ontology editing [on Armin Haller - due 2016-01-27].
<KJanowicz> IMHO our main challenge will be to develop a good formalization. Tools and serializations are secondary, lets make sure they do not hold us back.
Linda: topbraid has a free version
SimonCox: there is a limit on
visualisation (diagrams)
... Topbraid does better serialization
<KJanowicz> Do they have a linux version?
<ClausStadler> i don't know about any23, but for comparing changes in RDF files I was using raptor to convert arbitrary rdf syntaxes (http://librdf.org/raptor/ | sudo apt-get install raptor2-utils) to ntriples followed by a sort - not sure what happens with blank nodes though (i try to avoid them, but with owl axioms it would be a pain)
<joshlieberman> http://lodlaundromat.org ?
<ClausStadler> (what andreas just says ;) )
<joshlieberman> Also possible to use one of the "alignment" tools as the diff generator.
aharth: if we use canonical RDF we can use different tools
<SimonCox> N-Triples is always most reliable route to canonical form, but least inspectable visually!
kerry: that would be an ideal solution
<KJanowicz> Aidan's paper http://www.www2015.it/documents/proceedings/proceedings/p430.pdf
sorry josh, I did not get that
<aharth> a quick way would be just sort an N-Triples file, if there are blank nodes, use the algorithm from: Aidan Hogan, "Skolemising Blank Nodes while Preserving Isomorphism", WWW conference 2015
kerry: josh, please note the idea in the wiki
Linda: the LOD Laundromat could be used to clean up RDF
<joshlieberman> Alignment tool example: LogMap -- http://csu6325.cs.ox.ac.uk/
kerry: time is up, next topic
will be postponed
... reminder: please let us know about involvement in F2F
meeting
<ChrisLittle> bye and thanks Kerry
<KJanowicz> bye bye
<AndreaPerego> Thanks, and bye
<joshlieberman> bye and thanks
<ClemensPortele> bye!
<MattPerry> \quit
<kerry> bye!
bye
<kerry> trackbot, end meeting