See also: IRC log
<trackbot> Date: 12 May 2010
<ktk> hhalpin: no I'm new
Fabien - if you're online, it would also be great if you could give an update on the Open Graph Protocol to those who missed WWW2010...
<ktk> I'm adrian gschwend
<ktk> got introduced to it at last weeks LIFT conference in geneva
<ktk> so thought will have a look :)
<pgroth> good i'm in the right place
<scribe> scribe: hhalpin
<FabGandon> hhalpin: I'll try to join but can't promise, sorry.
<ktk> any sip available?
<ktk> hhalpin: k, will check irc then this time.
<DKA> Chair: DKA
<DKA> Scribe: Harry
<DKA> ScribeNick: hhalpin
<Yuk> zakim +??P20 is yuk
<cperey> forget it then
PROPOSED: Accept http://www.w3.org/2010/04/21-swxg-minutes.html as a true record of our last meeting
ACCEPTED: http://www.w3.org/2010/04/21-swxg-minutes.html are a true record of our last meeting
<danbri> i just joined, zakim seems sleep
<cperey> do we have invited speaker for meeting of May 19?
That is the most accurate
record of actions
<cperey> what about spending a meeting on Social Web Frameworks (second edition) paper and preparations for the final XG report?
it seems that christine's update was the most interesting thing that happened over the last bit of time.
so I'd be happy to discuss that next week.
DKA: We're doing an update on the final report
<cperey> how will we integrate all the inputs from different groups which have been contributed over the past year?
<danbri> (am i connected?)
<cperey> can we include this in the discussion about final XG report?
DKA: we have a paper on XMPP from
the OneSocialWeb work from vodafone
... so we have running code
... to let people set up their own node
cperey: we could discuss the final report next meeting
<cperey> the story...
DKA: the storyline of the final report must be discussed
<cperey> the narrative
I'm happy to review that, although it's been discussed extensively in the past.
<cperey> yes, and not include the XMPP detailed discussion in that
<cperey> is there a Table of Contents?
<mischat> hello all, sorry i am late
<cperey> I don't know the narrative so, I think we need another meeting on it
PROPOSED: next meeting to focus on final report?
<mischat> hi pgroth
<cperey> editorial meeting
ACCEPTED: Next meeting will focus on narrative and structure of final report
<ktk> hi henry
pgroth: the best way to proceed
is to take a look at the work we've done on the use-cases
... we gathered a use-case at the provenance XG
... does this reflect the work the issues from the Social Web XG?
... talk about how we've organized the requirements in our XG
... and if that captures what you guys think?
... if you look at that site
... we have a scenario about news aggregator
... imagine a news aggregator (BlogAd) that integrates information from Twitter and blogs
... to make a nice neat site, like an automated Huffington Post
... the key question here is how to determine where it's content is coming from
... the example we use is the trending topic on twitter
... a tweet about pandas being moved from chicago to florida
... so the aggregator wants to find information about it
... the particular story should be aggregated
... so we can track origin and give credit to the originator of the story
... and we'd also like for the originator to maintain anonymity
... not just republishing a tweet
... but adding metadata to the tweet
... follow the tinyURL on the tweet
... might find an image of the panda
... wants to display that image and figure out license associated with that image
... and whether or not we can license that image from the re-use.
... the originator of this panda tweet
... may want to retract that information
... and the blog aggregator should be able to determine
... if someone has retracted information
... or if someone dynamically wants to be dis-identified with a particular piece of information.
... this is the broad outline, trying to outline.
<bblfish> yes, makes sense
are there any technical solutions that could help solve this use-case?
<cperey> yes, the question we would like to have answered is "what are the types of technologies in Provenance used to solve issues such as set forth in this use case?"
I'm happy with your overview :)
<mischat> can RDF, as defined in the rec without named graphs, be used to track provenance information from multiple sources?
<cperey> Open Provenance Model
<cperey> Provenance vocabulary
pgroth: there are a number of open solutions
<cperey> Provenance Mark up ???
pgroth: in particular, the OPM, the Provenance Markup Language (PML)
<cperey> Provenance Markup Language (PML)
pgroth: and these solutions
... and there are then systems that *implement* the capture of provenance
pgroth: some involve
instrumenting the actual operating system (PASS?)
... other systems look at workflow engines, where workflow engines capture everything that happens
... and then there are programming languages that have provenance libraries that can then capture the flow of the program
... the hard part of all of these is that their isolated to one system
... so there is no clear open system
<paul> Somewhat related:http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=orms
pgroth: the common representation
... how do we instrument database system for capturing provenance?
<YolandaGil> Paul: you may want to describe the overall framework for provenance that we have developed in the group: http://www.w3.org/2005/Incubator/prov/wiki/User_Requirements#Requirements_for_Provenance_on_the_Web
pgroth: then the final work is
formally understaning provenance
... such as special operators for tracking provenance
... but overall, we are still in the space of understanding the scenarios
<YolandaGil> I think the framework helps think about provenance issues. There is no comprehensive solution space for all these issues, only pockets of work.
DKA: I'm struggling to get the
social aspect of it
... when your talking about assigining provenance which means that we have to identify the owner?
... what are your options for identifying the owner, particular people, facets of people's identity?
pgroth: alot of these solutions
ground out in an identity framework
... in the end, we have to know whose responsible for a particular item or entity
... we need to ground out in identity
... what that identity solution is should be more in social web space
mischat: I have a question
regarding the suitability of RDF as is
... whether or not it can track provenance at all
... are you contributing to the RDF next steps workshop?
... are standardizing named graphs a good idea?
pgroth: we submited a position
... we didn't endorse any solution
... but we came up with some existing solutions
... that we thought were kinda of good
<FabGandon> by the way, we also submitted a paper on that subject http://www-sop.inria.fr/members/Fabien.Gandon/docs/NameThatGraph/
<pgroth> yolanda are you on the phone?
<bblfish> you mean rdf/xml does not supprt named graphs. But it is in the semantics I believe (or else it would not be possible to do SPARQL)
yolanda: we have decided not to
endorse a particular solution
... we focussed on a few requirements
... in particular
... 3 major issues
... 1) content of what you are presenting
... 2) how do you manage provenance
<cperey> what about trust of the source of provenance?
yolanda: 3) how do you use provenance (merging it)
<mischat> bblfish: n3 is not a rec as far as i am aware
<FabGandon> +1 to look at the many possibilities existing in the literature
<mischat> rdf/xml is
<bblfish> Also I wonder if named graphs is needed on the publication side. If you want to say that something was written by someone you can say it quite easily <http://bblfish.net/> a foaf:HomePage; dc:created <http://bblfish.net/poeple/henry/card#me> .
yolanda: we would be able to
refer to groups and unions of triples
... and tell whether or not two entities are the same or not
... we also need to support evolution
... as any artifact will change over time
... this comes up in the retraction of tweets
<FabGandon> bblfish: RDF model doesn't have named graphs, RDF/XML does not have a syntax for that (as a matter of fact) only SPARQL has GRAPH and source primitives
... including any transforms of content in republishing
... the last requirement is the support for entailment
... some things you can't assert
<bblfish> the rdf semantics document speaks about graphs, merging them, etc… The rdf/xml syntax has a restriction, but that is something else
<mischat> thanks FabGandon, and yes am aware of your namedgraph paper
yolanda: so you could try to
associate if someone like hhalpin said something, he is a w3c
... these are the three requirements
... the two requirements that seemed relevant to baseline RDF
... one needs to make provenance assertions
... to publish a subset
... of provenance records
... I don't want to publish my name
... but I don't mind publishing some characteristics, a position, an affliation, an insitution I'm at.
... fifth requirement is querying
... we want to bring up these issues.
<bblfish> btw, for change over time, one could use the atom ontology http://bblfish.net/work/atom-owl/2006-06-06/AtomOwl.html
yolanda: so we did refer to particular solutions, such as named graphs
<FabGandon> mischat: one of the scenarios comes from Social Web and is called "Named graphs to embody the social act of tagging"
mischat: this is similar to what we've been talking about in the Social Web XG
<mischat> he is
<paul> I'm on the line now
so now I'll summarize that
oh, perhaps you should speak up now paul!
Three main identity technologies we've been looking at
1) OpenID - see notes from meeting with Dick Hardt, still seems like things may be revised in vNext.
2) WebID - using SSL plus FOAF, see notes from meeting with Henry Story
3) Infocards - see work on Higgins project, no meeting with them yet, but hoping to get one soon!
<mischat> Thomas and our chair Dan are hosting a privacy related workshop in london http://www.w3.org/2010/api-privacy-ws/
cperey: how do you manage
... verify it, source it, etc.
yolanda: there is one underlying
model of provenance
... possibly a very long chain, following all these tweets and blogs
... around all these tweets and blogs
<DKA> Also note that the OneSocialWeb/XMPP work represents another model for identity...
<mischat> higgin's project ^^ ?
scribe: from our work, we seem to think trust is very subjective
<cperey> trust is subjective, a judgement, about if an entity is reliable, etc
scribe: if a particular entity or assertion is reliable/believable
yolanda: so we think of trust as
one of the uses of provenance
... so we have separated the scope of our group
... and so we hope that many different approaches (like using social nets, algorithms) to derive and propagate trust over adequate provenance information.
... whether or not we trust particular provenance assertions
... is where signatures and authentication come in.
(notes that otherwise things get circular, i.e. we have to trust provenance in order to build trust!)
(so we need some grounding out...)
<cperey> I took my turn :-)
fabien: happy to hear the
provenance xg that there are many options
... so we are happy to see group wants to keep option open
... such as pat hayes' surface proposal, an alternative to named graphs
<mischat> http://eclipse.org/higgins/ Higgins projects : Personal Data Store (exciting), Identity Services, and Active Client
fabien: as far as I'm concerned,
we used the use-case
... of the social web xg in our paper
... so we want to embody the act of social tagging
... and we wanted to use named graphs in social tagging
(note that we also need to have a session discussing tagging, i.e. CommonTag and NiceTag work!)
paulgroth: we're a bit newer than
the Social Web XG
... so we're just starting to look at technical space of work
paulgroth: identity is a really important part of grounding out
<mischat> paul: are you from the higgins project?
<mischat> pgroth: = provenance xg
<bblfish> can't hear harry anymore
<paul> I talk, but nobody can hear me
<danbri> what happened?
yes we can hear you!
<bblfish> it was very interesting
sorry, there was a mute button on, due to keyboard clacking issues towards beginning of telecon.
paul: the space is still
... we are starting on OpenID vNext
<danbri> http://www.sociallipstick.com/?p=239 is an interesting read, w.r.t. oauth2 as a foundation for openidng
paul: but of course WebID and FOAF stuff is less mainstream
<bblfish> there's a train
<bblfish> that's the sound of foaf+ssl leaving the station :-)
DKA: Any perspective on XMPP-based identity out there?
Paul: it will be interesting to
see for us
... what the Social Web XG figures out
... so we're trying to figure out
... from the space of current provenance work connects the existing social web technologies
... alot of our technology comes from different spaces
paulgroth: a lot of work from databases and workflows
paul: I wish I had better
... the identity community is struggling with its own internal fragmentation
... the general belief is that provenance and reputation have a pre-condition of reliable identity
<YolandaGil> I agree!
paul: which I think is consistent with what the provenance xg is saying
paulgroth: the ability to embed
... so we could embed it in header of image
... and the whole notion of signatures
... that is part of this story
... but we haven't actually finished this off
yolanda: the only thing I would
want to say
... is that we've been trying to understand what different groups want
... so when we met with harry at WWW2010 conference
... there was some requirements that were new from the Social Web XG use-cases
... in particular, the deletion of records in a social networking site
<mischat> hehe, and for reasoning :)
yolanda: and what that has to do with provenance
<cperey> is there a problem with the bridge? or ...
+1 good point mischat
<cperey> I"m back :-)
yolanda: there are aspects that
we haven't considered.
... so if we hope to address all your requirements in our document
... all of us are very aware that we need to present solution-space, not just requirement-space
<cperey> agree also that this applies to SWXG work!
mischat: the whole concept of
... there is this notion of inferred triples
... we don't do reasoning in our triple store as we update and delete data constantly
... does the XG deal with entailment/reasoning in dynamic knowledge-bases?
... should I be putting all my inferred triples in the ontology
yolanda: entailment is one of the
big issues in the RDF next step papers
... absolutely definitely the case
... no record will be exhaustive or complete
<Zakim> FabGandon, you wanted to ask about multiple provenances in reasonning
<mischat> pgroth: ?
FabienGandon: Can I combine provenance from multiple sources?
<mischat> multiple personas online :)
<bblfish> "Beatnik: Change your mind" a simple blog post I wrote about trust http://blogs.sun.com/bblfish/entry/beatnik_change_your_mind
yolanda: so different provenances from different repositories
<mischat> media standards trusts are looking at as far as i am aware
<mischat> provenance in newss is a dream
<cperey> thank you!
<pgroth> yeah provenance in news is good
<pgroth> thanks everyone
Next call - final report narrative!!
trackbot, end meeting
<mischat> bye bye pgroth
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/2/ng/ Succeeded: s/is new/in news/ Found Scribe: hhalpin Inferring ScribeNick: hhalpin Found Scribe: Harry Found ScribeNick: hhalpin Scribes: hhalpin, Harry Default Present: +1.781.416.aaaa, hhalpin, +0774811aabb, DKA, MacTed, yoshiaki, Yolanda, melvster, cperey, +0773779aadd, +0752818aaee, mischat Present: +1.781.416.aaaa hhalpin +0774811aabb DKA MacTed yoshiaki Yolanda melvster cperey +0773779aadd +0752818aaee mischat Found Date: 12 May 2010 Guessing minutes URL: http://www.w3.org/2010/05/12-swxg-minutes.html People with action items:[End of scribe.perl diagnostic output]