WebID Community Group

09 Nov 2012


See also: IRC log


Bob_Powers, bblfish, Sandro, deiu, Kingsley_Idehen, domel, scor, MacTed, Jonathan_Dray, Andrei, Sambra, kidehen, Sandro_Hawke, Kingsley, Idehen, bergi, bobP
Andrei Sambra


<bobP> scribenick deiu

thanks bobP

<bblfish> +present

<sandro> bblfish, are you calling back?

I will scribe

What's the syntax?

<sandro> yep. I was just suggesting full names for that.

<turnguard> hi everybody, i'll be following in passive mode... gotta take care of kid...

<scribe> scribe: Andrei Sambra

Overview of what happened at TPAC

bblfish: we had a good week, all around

<bblfish> http://www.w3.org/community/rww/wiki/TPAC-Lyon-2012

bblfish: Monday was for introduction / IRC / getting people to create WebIDs
... there is a group.n3 file with the participants from TPAC

<bblfish> http://www.w3.org/2005/Incubator/webid/tpac/group.n3

bblfish: we had 25 people, if you are missing, please let bblfish know
... TimBL was there for both days
... people should use the queue when asking questions
... typing q+ will add you / q- will remove you from queue
... we had redefined WebID

<bobP> http://lists.w3.org/Archives/Public/public-webid/2012Oct/0265.html

bblfish: there was a picture drawn by TimBL about WebID integration
... left side has authentication systems / right has RWW
... we separated Identity from Authentication and Authorization
... had some good meetings on Wednesday
... 50 people interested in talking about identity
... most people were in agreement
... Thu and Fri was LDP f2f
... good to see how professional chairs handle things

scor: was there anyone from Google / Mozilla - browser vendors?

bblfish: the identity meeting had a lot of people
... no one really remembers if any browser vendors people were there

LPD Access Control

bblfish: we can now look at LDP and access control
... LDP chair agrees that AC is essential at this point
... but he is worried about sidetracking work from main LDP direction

sandro: it's not really up to the chair

bblfish: LDP chair is a but jumpy when someone mentions authentication

<sandro> sandro: The chair is reasonably trying to stop people from trying to solve the problem -- that's out of scope for the group.

<bblfish> http://www.w3.org/wiki/WebAccessControl

bblfish: we can have a session every two weeks to talk about WebACL

<bblfish> Proposal: we have some regular meetings here on this http://www.w3.org/wiki/WebAccessControl

<bblfish> any ideas on this?

deiu: I think WebACL is more important for LDP than it is for WebID

<bblfish> +1

<bergi> +1

<sandro> +0 sounds reasonable to me; doubt I can help

<domel> +1

<jonathandray> +1

<kidehen> +1

<scor> -0

+0 I think it's out of scope for WebID

<kidehen> this is RWW-0, I hope

bblfish: this can be part of the RWW

I agree with that

<bblfish> so we want to share this space with RWW people

<bblfish> and topics

<kidehen> that's my assumption, at this point

<scor> +1

kidehen, I agree with RWW-0

Is this a good time for regular meetings?

<scor> +1

<bergi> +1

<bblfish> +1

<domel> +1

<bobP> +1

<turnguard> +0

Today's time: 16:00 CET / 10:00 EST


<jonathandray> +1

RESOLUTION: We have some regular meetings here on this http://www.w3.org/wiki/WebAccessControl / Friday at 16:00 CET / 10:00 EST

The WebID definition

<bblfish> http://lists.w3.org/Archives/Public/public-webid/2012Oct/0265.html

bblfish: we need to look at the picture again
... we don't have a good process yet
... we should try to push for changing from a CG to a WG
... there was some disagreement on the mailing list about the definition of WebID

<bblfish> http://lists.w3.org/Archives/Public/public-webid/2012Nov/0020.html

bblfish: we need to reach a consensus
... TimBL argued about using turtle and # URLs

<bblfish> Proposal: A WebID is "A URL that denotes an Agent - Person, Software or Organisation."

<kidehen> +1

<bblfish> +1

bblfish: A WebID is "A URL that denotes an Agent - Person, Software or Organisation."

<turnguard> +1

<bergi> +1


<sandro> -0 but I wasn't in this conversation, coming late

<domel> +1

<turnguard> @kingsley : maybe you could also explain to others what is meant by "denote" and why it is used

<kidehen> lets finish

sandro: I have two issues
... first is about branding WebID: similar to OpenID, etc.
... second is having URLs that denote webpages is a "can of worms" http-range-14 issues

<kidehen> sandro: doesn't URI work ?

<kidehen> henry: ideally, yes.

bblfish: is there a way to word it in the spec so that we differentiate URL from URI?

<MacTed> sandro - that's where "WebID" has been erroneously used as shorthand for the "WebID Protocol" ... which has been distinct from "WebID" (a URL/URI/IRI denoting...)

<scor> Antoine had suggested to using IRIs instead in the spec

<kidehen> henry and sandro: we can negate the can of worms. We just need to agree about the least controversial i.e., de-referencable URI or URL

RESOLUTION: A WebID is "A URL that denotes an Agent - Person, Software or Organisation."

scribe: we should add conditions

<bblfish> "The WebID when dereferenced MUST return a document/representation that describes the URL referent uniquely."

scribe: a WebID that doesn't provide info about the thing it describes is not very useful

<kidehen> henry: are these conditions or attributes?

scribe: we should be open when defining conditions

<kidehen> bblfish: are these conditions or attributes, thus far you've outlined two attributes of a WebID

scribe: there is a third condition: the user has to have control over the WebID
... otherwise it's just a URL on the Web

<sandro> +0 I'm getting more comfortable with this.

<kidehen> :1

<kidehen> +1

<bblfish> Proposal: second part of WebID definition: ""The WebID when dereferenced MUST return a document/representation that describes the URL referent uniquely."

<bblfish> +1

<bergi> +1

+0 it's too ambiguous: "describes the URL ref uniquely" can mean anything in terms of representation

<sandro> +1 (assuming it's okay to have redirects, of course)

<bblfish> so something related to inverse functiona properties, owl2. keys

<MacTed> I would suggest s/the URL referent uniquely/the URL referent/

<turnguard> +0 document/representation is something one can argue about

deiu: there is no mention of linked data (which is the basis for WebID)

<jonathandray> +0 what about somthing like "... MUST return Turtle; servers MAY deliver other formats using standard HTTP content negotiation; If the client doesn't indicate a preference, Turtle MUST be returned"

<jonathandray> ok

<jonathandray> ok then +1

<domel> +1

<kidehen> +1

<MacTed> ... MUST support Turtle; MAY deliver other formats; SHOULD handle content negotiation...

<scor> +1

<kidehen> MacTed: how did we get to Turtle?

<scor> deiu: Linked Data is not a spec, can't be used normatively

scor, but Turtle is

<MacTed> kidehen - from jonathandray's comment.

<melvster> jonathandray: careful of setting defaults because browser hrefs cant do this, and you may want to serve e.g. XHTML+RDFa

<bblfish> Proposal: second part of WebID definition: "The WebID when dereferenced MUST return a document/representation that describes the URL referent uniquely."

<bblfish> +1

<kidehen> +1

<sandro> +1

<jonathandray> +1

<scor> +1

<MacTed> I don't understand "describes the URL referent uniquely"

<bergi> +1

<domel> +1

<melvster> 0 -- not pro or against, happy to go with it ...

<MacTed> "describes the URL referent" seems sufficient


<turnguard> +0

<sandro> "primarily describes the URL referent" ?

<kidehen> MacTed: we are getting to it. The description is a graph, the minimal set is as outlined in bblfish mail

<sandro> "provides identifying characteristics of the referent" ?

bblfish: it should have a cert:key relation

MacTed: the description must include self-reference

bblfish: the WebID document should give an inverse function property that uniquely identifies the WebID

<sandro> bblfish: The webid document should speak about the webid, in some way which uses an owl:InverseFunctionalProperty to uniquely identify the referent

<kidehen> MacTed: this is about the description. This is what makes the description one that uniquely describes the subject

sandro: can I have a WebID that is for EITHER me or my son?

MacTed: what if I don't use the WebID for authentication, but only for identity

<kidehen> You need > = 1 IFP relationship in the description document

<MacTed> dereference WebID URI, get document which describes referent *including* statements pointing to that same URI as WebID...

<sandro> sandro: I'm unclear what kind of RDF triples have to be there for it to really qualify.

<bblfish> http://lists.w3.org/Archives/Public/public-webid/2012Nov/0020.html

<kidehen> bblfish: your email covers it, yes

<kidehen> bblfish: we can make an example. melvster is a nice test case :-)

bblfish: we're moving forward with this, but it's difficult
... might want to continue the discussion on the mailing list

<sandro> I'm inclined toward SHOULD on this. I don't see why this kind of fuzzy thing should be a MUST.

Web Identity Interoperability spec

<MacTed> dereference WebID IRI, get description document which includes statements saying that this WebID IRI identifies the entity described by the document(s) which you get by deferencing this WebID IRI...

<bobP> The Agent could be the union of you and your son

<scor> the ultimate test is whether you can use your WebID in the protocol...

<bobP> but the referent to the Agent should be unique

bblfish: reuse existing protocols (BrowserID / OpenID) for WebID authentication
... focus on interoperability, starting from ACL
... use ACL on the wiki

<kidehen> +1 for Wiki notes showcasing reasoning based tricks etc..

bblfish: we already have a group.n3 file

<melvster> just a thought ... "A WebID is a URI that denotes an Agent. The WebID *should* but an HTTP URI. that when derferenced, returns turtle describing that Agent."

<kidehen> melvster: we don't need Turtle in the definition. We are beyond that, I hope

<melvster> ok sure ...

bblfish: we need to defuse the tensions between different groups and reach consensus

<kidehen> bblfish: yes

<kidehen> bblfish: yes, we have run the gamut re. good, bad, and the ugly.

<kidehen> melvster: your route to signing emails goes through one of these IFP reasoning examples :-)

bblfish: we can get people from different communities to help write a Wiki page about ACL use cases

<MacTed> thinking aloud - "A WebID is a URI that denotes an Agent. When a WebID URI is dereferenced, it MUST return a description of that Agent. The description MUST include statements which state that the WebID identifies the same Agent."

<bblfish> hippies

<kidehen> MacTed: we should put meat of <http://lists.w3.org/Archives/Public/public-webid/2012Nov/0020.html> into said Wiki doc

<kidehen> bblfish: yes. logic is the unifier

bblfish: we can use semantic web reasoning to help us

<MacTed> kidehen - yes

<kidehen> k

<kidehen> bye all

<scor> montly or bi weekly meetings please

<scor> bye all

<melvster> +1 monthly

<melvster> bye all

<sandro> bye, thanks

I propose we have monthly formal WebID meetings, and informal ones whenever we want

<melvster> +2

<melvster> oops

<melvster> +1

<bblfish> ok.

<bblfish> I think we could have monthly RWW one too

<jonathandray> bblfish: +1 for both

<melvster> -1 from me, id prefer to attend just 1 monthly

<melvster> (speaking personally not as chair ... others are free arrange calls as wanted)

I think ACL is orthogonal to WebID

as long as we emphasize on the _ID part of WebID

<bblfish> http://my-profile.eu/

my-profile.eu still has an issue with RDFa

<jonathandray> melvster: then 1 monthly meeting for both webid and rww ?

there's an open issue ticket on the EasyRdf github page

<melvster> jonathandray: sounds good, if it gets mega popular, we can increase the frequency by demand

<melvster> and informal meeting when people want

<bblfish> melvster you don't need to come to the WebID monthly meetings.

<bblfish> you can meet deiu

we should taskify someone to write an app for voting on proposals

users can cast their votes and add a description for their reasoning

bobP, bblfish don't forget the guys in NL want to give a WebID to 15.000 students

was it 50k?

bblfish, biweekly for wording the proposals (informal)

monthly formal for voting on proposals

<turnguard> ok bye

<melvster> 15k

Meeting adjourned.

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.137 (CVS log)
$Date: 2012/11/09 17:06:17 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.137  of Date: 2012/09/20 20:19:01  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/scribe: Mr. Deiu//
Succeeded: s/scribe: deiu//
Succeeded: s/MacTed: can/sandro: can/
Succeeded: s/reference/referent/
No ScribeNick specified.  Guessing ScribeNick: deiu
Found Scribe: Andrei Sambra
Default Present: Bob_Powers, bblfish, Sandro, deiu, Kingsley_Idehen, domel, scor, MacTed
Present: Bob_Powers bblfish Sandro deiu Kingsley_Idehen domel scor MacTed Jonathan_Dray Andrei Sambra kidehen Sandro_Hawke Kingsley Idehen bergi bobP
Agenda: http://lists.w3.org/Archives/Public/public-webid/2012Nov/0023.html
Got date from IRC log name: 09 Nov 2012
Guessing minutes URL: http://www.w3.org/2012/11/09-webid-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]