<scribe> scribe: simonstey

<scribe> chair: benws

<Brian_Ulicny> Can someone post the Webex link

<scribe> agenda: https://www.w3.org/2016/poe/wiki/Meetings:Telecon20161121

<renato> https://mit.webex.com/mit/j.php?MTID=m26bd8696de908ff6bd04482e85fa34e7

<renato> OR ;-)

admin

benws1119: approval of last week's minutes

<renato> UCR Editors Draft: http://w3c.github.io/poe/ucr/

ucr

<renato> https://www.w3.org/2016/poe/wiki/Constraints

simonstey: missing links between ucs & reqs; played around with Ivan's script and got it to work
... will continue working on that

constraints

michaelS: Victor and I started to work on a wiki page
... victor is the "green guy"
... there is a link to a second page where i've rewritten some definitions

<smyles> https://docs.google.com/document/d/1qGlBEpRgi60WyNnHLCfMuXlnTV1zzpoRz8DakIiOcP0/edit#heading=h.ai7fbfrg2s0a

michaelS: leftoperand & rightoperand, where the leftoperand is the more "flexible" representation

second page: https://www.w3.org/2016/poe/wiki/Constraints_in_the_Information_Model

renato: what do you mean with "defining how a value must be retrieved"?

michaelS: what has to be measured in this constraint
... it's about having a very clear definition of what the constraint is about
... less room for interpretation

smyles: I have a similar point -> for each constraint operator, where do you get the values from?
... and we have to distinguish between duty constraints and "normal" constraints

victor: either you leave this undefined
... you also want do define some additional information for each constraint, right?
... e.g. domain/range of properties?

michaelS: maybe.. there are some pros & cons for that

<victor> (I am referring to my fourth comment at the bottom of https://www.w3.org/2016/poe/wiki/Constraints)

michaelS: I think this could be solved in two ways.. either as ben said
... defining it in the free text
... or using "anonymized" properties
... e.g. "value is found at property1"

renato: we should define domain/range, as long its valid for all uses of that constraint
... do you want to split "absolutposition" into multiple constraints michael?

michaelS: yes
... maybe even more!

renato: also regarding "language"
... what do you mean with before and after?

michaelS: when the constraint should be evaluated
... either before an action is performed or afterwards

<victor> (are the new ones truly supported by Use Cases?)

renato: so you are proposing to deprecate some of those constraints an only use the new ones?
... the other option would be to clean up the definition of some of those things

michaelS: well.. depends on the users of such constraints

smyles: could we try (right now) to define one of those constraints more precisely
... i.e. taking one of those we think isn't defined precise enough
... and following all way through

michaelS: lets try "spatial"
... "A code representing a geospatial area." -> what does that mean?
... it should refer to the area respective action is exercised in
... but how to measure that?
... kind of difficult, so lets break it down into more simpler terms
... e.g. name of country or alike
... so I came up with "namedGeoArea"

smyles: this is actually a very common type of restriction
... 1) how the AP currently deals with geographic restrictions
... some of the restrictions are country based, but most are city and/or region based
... "my competitor across town cannot use this"
... we deal with those restrictions, by assigning a geography to each recipient
... we make a decision upfront
... what I did was, to built a geography hierarchy
... so you would have to say "in paris -> france -> europe.."

renato: could you use a getty thes. for that hierarchy?

smyles: AP has both an editorial & business hierarchy

renato: my question is: how far do we want to go wrt. supporting implementations?
... using conditions as some form of machine readable representation
... or having all of them also automatically verifyable
... for some constraints we can't know how the community will interpret them

<victor> +1

benws1119: there are two streams: 1) we have to increase precision thus include more and more constraints
... 2) do the opposite and leave most of those constraints for profiles
... some of those are very domain specific

Brian_Ulicny: we use geonames a lot in our group

<victor> (we also use and love geonames

<victor> )

Brian_Ulicny: we also need to be able to say "this is constrained to western europe and by western europe we mean the union of "...""

smyles: one way forward is to say: here is the def. of what we mean with geospatial
... and those are the different choices of implementing this

<victor> useful discussion? Yes but... which are the conclusions?

upcoming f2f

renato: [explaining his email regarding constraints on constraints]
... will send an update of the example around

benws1119: we have a virtual f2f coming up soon

renato: depends on our deliverables
... the week after the VF2F we should publish the 3 documents we have for now

benws1119: what about progress on constraints?

<renato> https://github.com/w3c/poe/issues

renato: yes, we should focus on open issues

<victor> +1

benws1119: maybe we should agree on a set of decisions we want to make at the meeting

renato: probably constraints on constraints/assets/...

<smyles> also constraints on duties

renato: if we aim to get constraints fixed/updated

yes

<renato> https://www.w3.org/2002/09/wbs/87755/F2F-2-prefs/results

[trying to interpret the results]

renato: seems like there's a slight tendency towards europe

This is scribe.perl Revision: 1.148  of Date: 2016/10/11 12:55:14  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found Scribe: simonstey
Inferring ScribeNick: simonstey
Present: Brian_Ulicny benws renato scribe smyles victor
Regrets: Phil Sabrina Ivan Caroline
Agenda: https://www.w3.org/2016/poe/wiki/Meetings:Telecon20161121
Found Date: 21 Nov 2016
Guessing minutes URL: http://www.w3.org/2016/11/21-poe-minutes.html
People with action items: