Tracking Protection Working Group teleconference -- 14 Nov 2012

<aleecia> please mute

<tl> I'll be able to spectate and tune in a little, but not materially participate.

<ifette> Out of curiosity, is anyone going to the workshop in two weeks?

<Walter> ifette: haven't been invited

<tlr> ifette, yes

<ifette> I don't think I can make it. It's literally right after thanksgiving

<dwainberg> Ian, I'll be there

<dwainberg> yeah, not great timing wrt thanksgiving

<npdoty> ifette, Walter, some of us have been behind on sending out email and invitations while traveling, but ping me again if I haven't been responsive on email

<Walter> npdoty: I was referring to the workshop in BXL, I don't even exactly know when it is

<Walter> npdoty: but I plan to be in BXL on 27th and 28th anyway

<ifette> not sure if I'm going to be with family in New Mexico or snapping photos of Angkor Wat in Cambodia, but I will not be in SF...

<ifette> When was the last time we approved previously scribed minutes?

<aleecia> scribenick: dsinger

<ifette> I see all minutes since July marked as draft

<ifette> no minutes from October

<dtauerbach> 301 is me, but i'm unable to scribe

<ifette> or November

aleecia: we'll start with overdue action items

…good progress, which is great

<ifette> aleecia...

<ifette> agenda q

<ifette> did we stop reviewing past minutes?

…for the first 4 (Justin), he reports he cannot join, these are editing, and he'll be doing these in this next week

<tl> Not with voice!

<tl> I'm "multi-tasking".

<npdoty> action-323?

<trackbot> ACTION-323 -- Thomas Lowenthal to share results of what-the-response-is-for discussion -- due 2012-10-22 -- OPEN

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/323

…next Tom Lowenthal

<npdoty> tl, do you have an update on 323 to share via irc?

…we need Tom's actions done

<npdoty> I think it's just one action open on tl

<tlr> action-268?

<trackbot> ACTION-268 -- David Singer to edit the TPE document to make sure that the final definition of parties is in sync across the two specifications -- due 2012-11-12 -- OPEN

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/268

<rigo> I think that was done

<tl> npdoty: Can't talk much now, sorry.

<npdoty> marked 268 pending review

…dave singer to coordinate with Justin on the joint one

<rigo> action-330?

<trackbot> ACTION-330 -- Rigo Wenning to send a reminder about http://lists.w3.org/Archives/Public/public-tracking-international/ -- due 2012-11-07 -- OPEN

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/330

…then Rigo's on international action 330

Rigo: thinks it is done, sent a reminder to the public list, with lots of responses. We have 36 registered for the task force. French are invited.

action 320 can close

<trackbot> Sorry, couldn't find 320. You can review and register nicknames at <http://www.w3.org/2011/tracking-protection/track/users>.

<ifette> we can hear

Nick: the minutes are still being automatically drafted, and cleaning up is taking a lower priority.
... not prioritizing manual cleanup

Ifette: would like the minutes cleaned and available, approved for accuracy and made available. worried that our ability to correct will get worse as they get older

<npdoty> while it happened in theory, I think we had problems with confirming accuracy, even with short term approvals

…has not been looking at the raw minutes for accuracy

<susanisrael> 215286aann is susan israel

aleecia: one vote for more rapid cleanup; perhaps in reverse chron order, so we can check the recent ones

<laurengelman> i just joined too

ifette: is fine asking people to review the draft/raw minutes, but don't want to put it all on Nick

<susanisrael> i came in late to the discussion of minutes but i know there was some errors in the minutes one day when i scribed. i would have been willing to go back and clean them up.

tlr: maybe the easy way out is dirtier minutes more rapidly available

<npdoty> hearing: make minutes only relatively clean and more relatively quick

<susanisrael> i think dirty minutes would be ok

<Walter> +1

dsinger: would like them indexed on the web site to track down what action items and issues really mean

<Chapell> +1 to susanisrael and nick

aleecia: let's go for sooner+ dirtier if needed and see how people like it

<eberkower> aamm = eberkower

<Chapell> zakim aamm is Chapell

aleecia: checking callers...

<Simon> that's me

<npdoty> at least one IP caller

<laurengelman> I am the IP

<laurengelman> I was one of the VOIP

<laurengelman> :)

Texts nearing a call for objections poll

aleecia: now getting close to calls for objections and decision process; let's get sync'd up

<aleecia> http://www.w3.org/2011/tracking-protection/drafts/tracking-compliance.html#def-unlinkable

…(1) unlinkable data: 3 possibilities; 2 are in the compliance document.

….3.6.1 and 3.6.2; and the 3rd is action 286

action-286?

<trackbot> ACTION-286 -- Brendan Riordan-Butterworth to propose DAA text regarding de-identification (for unlinkability discussion) -- due 2012-11-22 -- OPEN

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/286

…action 286 is the DAA text

….we're now five weeks in; the owner changed the due date :-(. We need it completed immediately or it'll be closed for lack of action

Lmastria: we feel it's a blank check, we don't understand how the defn will be used

<dtauerbach> +q

…a concern is that folks are trying to take bits and pieces of a whole cloth and pick the bits that they like; the DAA program is not suited to that

<npdoty> after we decide on a definition for a term, it's possible that that would cause people to have concerns about related compliance requirements, right?

…until we have clearer defns around what tracking means, how this will be used, it's hard to find any language

<ifette> At some point, it seems like we're moving towards a vote on different text for unlinkability

<ifette> issues aside, i'd rather have a DAA option to vote on than no option

aleecia: we'll take the Q

<WileyS> Inappropriate!

jmayer: some comments on process

<WileyS> The concern was raised on the public mail list

<WileyS> 2+ weeks ago

…is concerned about a problem being surfaced 5 weeks after the action was raised

<Lmastria-DAA> The issue of lack of fundamental definitions has been raised before. it is being raised again. Nothing has changed

…on the substance, the point is to find constraints where consumer privacy concerns are really reduced. Am concerned about the idea that that idea should be flexible

…the point of unlinkability is to find stuff that, well, cannot be linked

<aleecia> this *is* a definition.

<aleecia> we're trying to nail down the definition of unlinkability

<fielding> while I think the concern was obviously raised many times during, before, and after the action was assigned, let's just drop the action until someone does it.

<tlr> lmastria, "unlikability" seems like a fundamental definition to me. And I'm confused about about the current due date on the action item if you have a fundamental concern.

walter: I share some of these concerns. (garbled)

<ifette> cant really understand walter

<aleecia> very hard to understand you, walter

<ifette> very bad audio quality

<jchester2> can someone tell Walter there is a sound problem, please

…(())) monolithic solutions…linkability…is willing to take on the action

<eberkower> Having a hard time understanding what he's saying

<Chris_IAB> unlinkability, like other things, need to be considered in the context of the definition of "tracking" -- which has been raised (but not opened by the co-chairs) for FAR longer than 5-weeks

<Walter> Ok

<rigo> Lou, the lack of fundamental definitions can be seen as a bug or as a feature. For the definition of "tracking" I consider it a feature. At best, we will define "tracking" as a definition of DNT:0

<rigo> I consider it a feature because it will create more problems than it solves

aleecia: suggests you take the action, see 3.6.1 and 2, and then see if a new text is warranted

<jchester2> Question for the DAA and DMA. Does the DAA still support its proposal from Amsterdam that its language related to advertising and marketing be a permitted use? Thanks

<Walter> Basically, what I said was that it is not the nature of this process to approach the text as a monolithic thing. This is a conversation.

dtauerbach: it does seem kind of weird that we can't settle this without the rest of the document. volunteers non-normative text, with Shane

<WileyS> Dan - most of the non-normative text already exists - see the full draft

<Walter> Unlinkable data has been defined

action dtauerbach to write non-normative text on unlinkability, with Shane

<trackbot> Sorry, couldn't find dtauerbach. You can review and register nicknames at <http://www.w3.org/2011/tracking-protection/track/users>.

<Walter> What lacks a definition is a description on how to get there

<npdoty> Section 3.6 doesn't have non-normative text

<ifette> use auerbach

<ifette> instead of dtauerbach

<WileyS> Okay - its on the mailing list so I'll dig it up and send it to Dan. I believe the non-normative text works with either definition :-)

<vincent> Walter, not sure we agreed on a definition

<ifette> dsinger, use "auerbach"

<jmayer> My point: I don't see why context should matter. Unlinkability is about the set of constraints under which consumer privacy risks are reduced because data cannot be linked. I'm greatly disappointed if participants want to distort the plain meaning of that term and the accompanying computer science out of crass economic interest.

action auerbach to write non-normative text on unlinkability, with Shane

<trackbot> Created ACTION-335 - Write non-normative text on unlinkability, with Shane [on Dan Auerbach - due 2012-11-21].

<npdoty> ACTION: auerbach to write non-normative text on unlinkability (with Shane) [recorded in http://www.w3.org/2012/11/14-dnt-minutes.html#action01]

<trackbot> Created ACTION-336 - Write non-normative text on unlinkability (with Shane) [on Dan Auerbach - due 2012-11-21].

<npdoty> close action-336

<trackbot> ACTION-336 Write non-normative text on unlinkability (with Shane) closed

<Walter> vincent: the definition has been shared with DAA over a week ago with zero response so far

<Walter> vincent: feedback would be very welcome

<npdoty> action-336: duplicate of 335, sorry for the confusion

<trackbot> ACTION-336 Write non-normative text on unlinkability (with Shane) notes added

<fielding> rigo, not defining tracking is not an option because it is being used to instruct users to set DNT. If we don't define tracking, this WG is over.

<jmayer> The correct response, if you don't like how the spec uses unlinkability, is to not use unlinkability. Don't fight over the definition.

<susanisrael> +1ianfette

ifette, gets that it is hard to come up with this in isolation. there is a rat-hole here. had we defined tracking, that would have an effect on the rest of the document. that said, I do see aleecia trying to move to some conclusion.

<BrendanIAB> For the record, I updated due date when the issue was assigned to me with a due date that was before the current time (assigned to me on Oct 31 with due date of Oct 10)

<rigo> fielding, we define "Not-tracking"

…we need some option from the DAA, it would be good to have that perspective. I realize it's not as simple as cut-n-paste, but there is a plea here to find an alternative option we can look at

<rigo> and in DNT:0 we define "tracking"

<fielding> regardless, it is not necessary to define tracking in order to define unlinkable -- they are not dependent concepts

<vincent> +1 fielding

rachel: I hear the frustrations, but the concern about the definition of tracking is also long-standing. We understand wanting to continue the conversation, and we can give the DAA defn, but to decide on this without a definition of tracking maybe premature. We need tracking defined as a basis first.

<npdoty> the DAA document uses the term "de-identified" rather than "unlinkable", yeah?

<jmayer> The procedural problem is that there's a claimed dependency five weeks after the action was assigned. That != the group not taking up an issue on its agenda.

<jchester2> Rachel: Are you repesenting the DAA or the DMA?

aleecia: a conversation on 'tracking' would be very long and would log-jam too many conversations, and it's not my choice as chair to do things in that order

<Lmastria-DAA> +q

aleecia: please provide text or the action will be closed

rachel: will offer some text before friday.

aleecia: many thanks

rigo: Two things. The unlinkability discussion is 'how to get off the hook', so there is some sensitivity around this (partic. around re-identification). We need a decent std of when people are off the hook

…and it shouldn't be dependent on the definition of tracking, because it's about how you behave afterwards

<susanisrael> is it possible that we don't need to define unlinakable? Maybe this is a bigger project for elsewhere in and of itself?

lMastria: with all respect, I know it's frustrating, and challenging, but the reality of the conversation about '

<npdoty> susanisrael, I think the "unlinkable" definition is used importantly in the document, for declaring data that is out of scope for compliance requirements

lMastria: about what tracking is, may be long and cause log-jams, but I represent a whole eco-system, and I can't say I am signing up for a way to look at things, that may go counter to some business models, and also counter perhaps to some things we offer as choice

<BerinSzoka> This conversation about whether or not to grapple with definitions of tracking and unlinkability in their full context reminds me of the wisdom of that great sage Homer Simpson: "If something's hard to do, then it's not worth doing."

…I can't say I'll waiy for later

<fielding> rigo, defining tracking (or not tracking) is necessary to give the header semantics because that is how the user expectation is set

…I do need to know what I am negotiating with. I appreciate that increases frustration,

<Walter> BerinSzoka: could you be a bit more constructive, please

<npdoty> lmastria, I think we hear your concerns about compliance, but is it necessary for us to avoid definitions at this time?

aleecia; can you clarify how defining tracking might go against the choices for transparency? online or offline

<susanisrael> npdoty, I understand, but it's also a big question that could consume a fair amount of time and energy in and of itself. Does need to be defined, but maybe defer that?

lmastria: none of us seem to know what tracking is, at the moment; none of us know the impact without the defn

aleecia: would like to understand more, perhaps an offline conversation?

<npdoty> susanisrael, I think a high-level definition of "unlinkable" is achievable, and would be useful for the document

<laurengelman> my understanding is that compliance requires a company to do X in response to getting a Y. None of that requires understanding the word tracking.

<ifette> It may not show up in the normative language but it's in the name of the header

tlr: 2 points. We have a few other action items where the mailing list feels that the defn of tracking is a key element. It might be good to talk about the tracking defn esp. since the word doesn't show up in normative text!

<BrendanIAB> +q

<fielding> tracking is normative is TPE and a charter requirement for Compliance

<susanisrael> npdoty, ok i accept that. My technical team owes me some answers on this anyway and I will use that to evaluate or try to help.

<BrendanIAB> The assertion that “tracking” does not show up in the current normative language is incorrect.

<ifette> Saying "tracking" doesn't show up anywhere may be true in the language of the spec, but the otuside world thinks we're defining "do not track" and so tracking is definitely a relevant term

<laurengelman> so it is a messaging issue, not a compliance issue

<Brooks> The name of the spec is therefore non-normative language?

<BrendanIAB> In the TPE document, the term “tracking” is used 296 times, including in the title. It is used extensively in the simple form, and in two primary compound forms, “tracking preference” (49 times) and “tracking status” (119 times).

<aleecia> The name of the spec is not normative

<rigo> fielding, I don't think we are far away from each other

<ifette> aleecia, that may be true, but so far all the implementations present the name of the spec (and nothing else) to users

<ifette> so it seems germane

<Walter> I'd be happy to provide a definition for tracking

…second, it is always possible that text contributed (early) ends up inconsistent with what gets put in later on, and then it's totally fine for the chairs to re-open issues. We shouldn't take that as a deterrent, and there is risk of circular dependency. If there are concerns about context, let's be open about that, and get it on the table

<laurengelman> unlinkability is a word with meaning. It is used when the writer INTENDS to comit MORE than the current use at anonymous

<WileyS> +q

<aleecia> note that unlinkability has a different and specific meaning in europe. we'll need another name no matter what.

jmayer: propose a possible way ahead: I hear people saying "if unlinkability is a tough defn, they wouldn't be comfortable using unlinkable data for some things". Those a fine conversations to have. Let's not front-load those, and not load them into unlinkable; let's define it as it's understood generally, and then if there are concerns that more is needed in some contexts, let's have that conversation.

<npdoty> currently we use "unlinkable" in the sections on Minimization, Persistent Identifiers and in proposals around Short-Term Collection

aleecia: when we have a defn, we can have conversations, so I am taking the DAA's point to heart, and take a defn, and see how we use unlinkable data

…(please don't repeat points)

brendan: notes that the term 'tracking' is used in normative TPE text. it should be included

<WileyS> It should be removed from both document titles then as well

<jmayer> The term "tracking" isn't substantively used in the compliance document. If it remains in the TPE document, it's on the way out.

<jmayer> New points please, Rachel.

rachel: is concerned about a strategy of removing the word from the text, rather than defining it. We are overall defining what it means to 'not track'.

<laurengelman> what it means to track is what the spec says you cannot do when a user sends DNT:1

<BrendanIAB> jmayer - it's used 296 times in the TPE

<aleecia> thank you, Rachel

<npdoty> I think BrendanIAB's point is that "tracking" isn't used in the compliance document, but is used in the TPE document which expects a normative definition from Compliance

<Chris_IAB> The word "track" is in the Charter; "Do-not-track" is also in the charter.

<efelten_> The TPE uses terms like "tracking status resource" which is a precisely defined term. That contains the word "tracking" but not in an undefined sense.

<jmayer> Brendan, as you pointed out, the majority of those instances are in a defined term that happens to include the word "tracking."

<BerinSzoka> +1 to Rachel

<WileyS> 'Tracking" Preference Expression & 'Tracking' Compliance and Scope

<jchester2> For me, many of the industry have tried over next few weeks to undermine this process, by adding proposals designed to weaken the final outcome.

…absolutely hears that we can change things if the context changes, but notes the context that we're on a short time-frame, and we're trying to close down, not open up, so that's the context, and that the chairs might decide, and that opportunities to revisit may be limited

<laurengelman> we cannot rewrite history

<Brooks> Can someone tell me if the title is normative text?

<Lmastria-DAA> +1 Brooks

<WileyS> Appears "Normative" in context to me...

aleecia: we knew the schedule was aggressive when we started; we hoped DNT could relevant for the EU context, and it's also seeing implementation in browsers, and users are seeing something that does comparatively little right now, and the chairs do get yelled at for being behind

<npdoty> I think this is just a curiosity, but I don't think the titles of documents add normative requirements on implementers

<BrendanIAB> jmayer - compound form still requires that both terms be defined. Else, why not refer to it as the "Ablative Preference Expression"?

<Chris_IAB> efelten, the group is called the TRACKING Protection Working Group (top of the Charter)

rachel: the EU context is not being answered directly in the WG, but in the global considerations

<ifette> +1

<efelten_> That's not a normative part of the spec.

<Walter> rigo: I would agree

<WileyS> Nick, seriously? You feel its appropriate to have two documents that start with the term "Tracking" and never define it? I'm not seeing how that is rational to expect.

<jmayer> If you'd like to call it something without the word "tracking," like "DNT Protocol Preference Expression," totally fine by me.

aleecia: some things are happening in the EU because we're late, and there is talk of a workshop and a non-normative document right now

(queue jumping to Rigo)

<WileyS> jmayer, DNT carries the term Tracking - remove it completely or define it.

<Walter> npdoty: I think we may indeed get to the point where we can not postpone defining tracking any further

<ifette> Perhaps this is not the right venue to solve global policy issues

<npdoty> WileyS, I wasn't taking any position, I was responding to the question of whether document titles were normative text, which I pointed out was a curiosity

<Lmastria-DAA> +1 ifette

<Walter> ifette: the web is worldwide...

<Chris_IAB> efelten, I could find over 30 instances of the word "track" or "tracking" on this group's Charter page... Not all are part of terms that have been defined by this working group. Taking it out of the spec, would to me, seem like we need to re-charter.

<jmayer> We've been over this many, many times: the documents do not need to define the magic word tracking. If you need a definition, tracking is whatever the documents say you can't do.

<WileyS> Nick, I don't believe they are directly but in context they are (as you're basically saying "this document titled XYZ requires these rules be followed")

rigo: there is some pressure from the EU from high up, and they were talking in transatlantic talks. Don't forget the next step is candidate, which is implement and find out what goes wrong etc. Then we have the Global Consids document, and if we can't make it, because of mis-match, we'll come back

<jmayer> That doesn't in the slightest turn on whether the magic word "tracking" is in the document title or defined phrases.

<WileyS> jmayer, feel free to put forward that definition then :-)

<rigo> Global considerations will be a Task Force

<jmayer> Anyways, off to Corporations again.

<rigo> of this Working Group

<jmayer> I'm putting forward that we actually get work done.

aleecia: no, the chairs don't 'decide'; we'll have objections come from the group, and the chairs try to find the least objectionable (i.e. the closest to consensus)

…finally: when we get to last call, how fixed is the text?

<ifette> Aleecia, you said one of two things. Either we do another LC or we make minor tweaks and move forward. There is a third option, which is that it may simply be unworkable

…we need another draft: the LC draft goes out for wider review than the 90 in the WG, and we'll get comments from other WGs, from the world, and from implementers. That's all useful feedback. We might do another LC document, if we weren't even close; or if we feel we can fix the issues, we go through the process to final rec.

….we might have 2nd or even multiple LC drafts

<npdoty> Process info on Last Call is available here: http://www.w3.org/2005/10/Process-20051014/tr.html#last-call

…there is lots of opportunity for feedback from WGs, external bodies, implementers, and so on: we test and make sure it's do-able

<tlr> nope

<npdoty> I think that was a fine description of Last Call

…comments from the staff? No

<Chris_IAB> jmayer, most if not all of the ad industry is not in agreement with your position that we should not define track and tracking here in this group. In so much as we are part of this working group, there is disagreement. Just wanted to note that for the official record.

dwainberg: on the process, (q was closed), that was helpful. 2 questions: what is the standard for the least strong objections?that they cannot live, or the substantive merits of the objections.

aleecia: the chairs do try to judge the strength of argument, not the volume

<Chris_IAB> we don't see this as a "magic" term. It's just a term, like other terms. Just needs to be defined precisely to give this group's work context and meaning.

<berinszoka_> Aleecia, could you elaborate briefly on the criteria by which you assess objections?

aleecia: stating 'we can't live with this' is not strong, arguments need to be grounded in technical merits; the document tries to make this clear
... the chairs are seeking to find the consensus of the group.

tlr: notes we are off a discussion on unlinkable

dwainberg: on implementations: I get a sense of judging success of a technical spec.,m but what criteria apply to the compliance document?

<npdoty> I believe we have an issue/action postponed about exit criteria for CR

<berinszoka_> great question! the more clear we are now about criteria, the fewer problems we'll have in the future

<ifette> it's ISSUE-131

<rigo> david: Karl said we have to have it testable. I think that is a very important question!

<ifette> ISSUE-131?

<trackbot> ISSUE-131 -- What should implementations look like to satisfy our exit criteria for CR? -- raised

<trackbot> http://www.w3.org/2011/tracking-protection/track/issues/131

<tlr> correct

aleecia: ifette has an open question 'what are our exit criteria for the compliance document' which is excellent to consider

(back to Q) unlinkable data

<Walter> +1

fielding: First, there is no direct dependency of unlinkability on tracking; there is one for the completion of the spec.

…second, on what thomas was saying, there is not a defn right now, but that's not in accordance with the charter, and that would be a problem

<rigo> roy, would you work with me on such a definition. We need it for DNT:0

…browser UIs use that one word right now, and I can't continue with this tour which feels like the compliance document doesn't define anything useful

aleecia: to summarize, sees defining tracking as extremely important, because T stands for Track in DNT

<tlr> He means that, since the charter is phrased in terms of "tracking", we need to explain how the document responds to that.

<tlr> I suggest we move on in the agenda.

…does not see the charter question. But we had an open question on whether we should use 'do not track' at all; that's postponed

shane: quickly. to refute jmayer's claims that there is a standard, accepted, defn of unlinkability. keep that in mind

<berinszoka_> wait, Aleecia, did you just suggest that we might reconsider the "Do Not Track" brand? I think there's a lot to be said for that but, as you point out, it's already being implemented in the field...

aleecia: maybe we need a term other than 'linkable' as that's a term in use in europe

…by Friday, we'll have DAA text. We'll have 3 options on the table.

<WileyS> Aleecia - that is not a EU legal term, correct? I'm familiar with our use of the term in our work with the A29WP with respect anonymizing search log data - is that what you're referring to? If yes, the A29WP put out an opinion paper on this topic specifically.

…3.6.1, 3.6.2, and action-286 output.

<npdoty> BerinSzoka, that was suggested early on (a different brand) with the acknowledged limitation that "Do Not Track" is already a popular term

…monday, we'll put out a request to review, and see if there is anything they don't expect. after that, we''ll go through the call for objections process

<fielding> TPWG charter is at "http://www.w3.org/2011/tracking-protection/charter.html"

…if we can reduce before the call for objections, great.

…but at this point, we have 3.6.1 vs. 3.6.2

…oops, two more pieces of this agenda item.

<aleecia> http://lists.w3.org/Archives/Public/public-tracking/2012Oct/0652.html

<berinszoka_> again, I think there's a lot to be said for rebranding, but... if that's still a live item, I hope we'll give it the time it deserves

…one is regulatory hooks, action ???

…one is the option of silence; however, power is back on the east coast

<rigo> roy, the trouble of the definition of tracking is that tracking only covers personal data. And to cover, we would have to define personal data. And Wham you're back to an impossible definition and we would say cookies, IP address etc are not/are personal data. => terrible

<fielding> action-286?

<trackbot> ACTION-286 -- Brendan Riordan-Butterworth to propose DAA text regarding de-identification (for unlinkability discussion) -- due 2012-11-22 -- OPEN

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/286

…four options are on the table...

<npdoty> action-324?

<trackbot> ACTION-324 -- Aleecia McDonald to compile public compliance commitment options for potential Call -- due 2012-10-24 -- PENDINGREVIEW

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/324

<npdoty> s/???/324/

…looks like people are thinking or ready for moving on

<Walter> rigo: not necessarily so, I disagree. When you look at the latest definitions floating around in BXL, it is pretty clear that for the most part cookiets etc. are personal data

<WileyS> Aleecia - can you enumerate the 4 options now, please?

<robsherman> I think the end of the thread said that Jonathan and Tom were going to try to combine their proposals. Did that happen?

<Walter> rigo: also, look at the proposed definition of unlinkable data

aleecia: will be sending the options and question around as final summary, and head towards call for objections here

<WileyS> Aleecia - please see my question above...

<Walter> rigo: which means that anything that does not meet that definition is linkable

<robsherman> Thanks

<tlr> issue-45?

<trackbot> ISSUE-45 -- Companies making public commitments with a "regulatory hook" for US legal purposes -- pending review

<trackbot> http://www.w3.org/2011/tracking-protection/track/issues/45

aleecia: jmayer and Tom were going to combine, but…neither can comment now and we have not seen text. doesn't seem to have happened

…will ping them both in email and see if we can reduce

…it's great to reduce options

<rigo> Walter, but we will bicker around what is personal data and what isn't. And that is endless, because you'd have to solve the transatlantic divide. If you reassure me, the better. We will have to define tracking in DNT:0 to make it possible in Europe

<aleecia> +1 david

<susanisrael> wileys i think aleecia posted the link to the 4 options- http://www.w3.org/2011/tracking-protection/track/actions/324

<npdoty> "likes it when we reduce to one [option]" :)

<Walter> rigo: I think we cannot avoid the inevitable, otherwise we'll erode the credibility of the process

<Walter> rigo: but we may have to take this conversation offline

fielding: the 2nd 1/2 of the options seems to assume the response header field, so there is a dependency there

<susanisrael> wileys, i think aleecia posted link to the options though it couldn't hurt to talk through them. http://www.w3.org/2011/tracking-protection/track/actions/324

<rigo> but we can try to avoid it

<npdoty> fielding, you're pointing out that one of the options would allow indicating that they don't comply with the Compliance requirements, yeah?

aleecia: if we have a response saying " I don't comply", then saying "any response indicates compliance" is a contradiction.

<susanisrael> sorry i repeated myself--i thought the first comment had not appeared.

aleecia: two things: Roy will fix the text, and Tom/Jon will be pinged to combine. one week on that

<npdoty> ACTION: fielding to clarify text around compliance hook options and indications of non-compliance [recorded in http://www.w3.org/2012/11/14-dnt-minutes.html#action02]

<trackbot> Created ACTION-337 - Clarify text around compliance hook options and indications of non-compliance [on Roy Fielding - due 2012-11-21].

amyc: on Roy's question, we also discussed some sort of token 'some token to say implementation in process'

<npdoty> ACTION: aleecia to work with tl/jmayer on combining options on compliance hook [recorded in http://www.w3.org/2012/11/14-dnt-minutes.html#action03]

<trackbot> Created ACTION-338 - Work with tl/jmayer on combining options on compliance hook [on Aleecia McDonald - due 2012-11-21].

<fielding> actually, my point only applies to (2) in Aleecia's message

aleecia: group has shot that down?

<rigo> +1 to dsinger

<ifette> i still support a "testing" state

<npdoty> aleecia, I have heard that on the calls, but occasionally hear differently from people offline

<rigo> even P3P had a testing state

dsinger: thinks that it was just a misunderstanding

aleecia: but we heard that industry didn't feel it needed it on a call

<npdoty> aleecia definitely asked on more than one call, and we didn't hear any interest

aleecia: maybe open as a TPE issue

tlr: among the 3 or 4 I have heard express an interest (amy, dave, ian, ??) maybe take on an issue

<npdoty> tlr: hearing interest in a beta flag from amyc, ifette, dsinger -- can one of you define the issue?

<fielding> we have a issue already

<rigo> which one?

<amyc> thanks dsinger

need to link the issue to singer's email on responses

<tlr> if we have one, even better

<tlr> I though we didn't, and wanted to be sure we don't lose this.

<WileyS> Permitted Uses

<aleecia> 6.1.2.6

<aleecia> action-260

aleecia: back on 6.1.2.6 (!), within the compliance, which has an alternative

<fielding> issue-161?

<trackbot> ISSUE-161 -- Do we need a tracking status value for partial compliance or rejecting DNT? -- raised

<trackbot> http://www.w3.org/2011/tracking-protection/track/issues/161

….from action 260 therte is another text (from Nick)

<aleecia> http://www.w3.org/2011/tracking-protection/drafts/tracking-compliance.html#security

<aleecia> http://lists.w3.org/Archives/Public/public-tracking/2012Oct/0466.html

<aleecia> http://www.w3.org/2011/tracking-protection/track/actions/293

…thinks that these should be put all in one place under the action

<aleecia> http://lists.w3.org/Archives/Public/public-tracking/2012Oct/0506.html

<aleecia> http://www.w3.org/2011/tracking-protection/track/actions/293

<fielding> under an issue, please

…two possibilities for graduated responses, action 279 and 293; gave examples

<npdoty> ah, "partial" compliance, I hadn't thought about that as the beta/testing flag, but maybe it is

<tlr> action-279, action-293

<tlr> (for tacker's benefit, the dash is required)

…thinks we need to talk through 279 and 293, which are pending review. does anyone think we're in a different state?

action-279?

<trackbot> ACTION-279 -- Ian Fette to write an explanation of graduated response and a list of explanatory use cases -- due 2012-10-10 -- PENDINGREVIEW

<trackbot> http://www.w3.org/2011/tracking-protection/track/actions/279

walter: would like to propose some refinements, as the current text(s) have problem

<WileyS> Please do that on the mailing list - will create a 5th option

<Walter> will do so

<npdoty> Walter, can you clarify on IRC what the problems are?

<aleecia> hi, David!

<Walter> npdoty: there are no limits on retention and no remarks on proportionality

<npdoty> ACTION: walter to propose a refinement to debugging permitted use [recorded in http://www.w3.org/2012/11/14-dnt-minutes.html#action04]

<trackbot> Created ACTION-339 - Propose a refinement to debugging permitted use [on Walter van Holst - due 2012-11-21].

dwainberg: can we separate grad. response from security/fraud and debugging? since it is only 'preferred' it's not really normative.

<npdoty> walter, I think the minimization limitations are general requirements which apply to all permitted uses

<Walter> npdoty: even then a retention term would be in order

<npdoty> s/npdoty:/npdoty,/

aleecia: I think we have a dependency, we need to figure out what is grad. response esp. for security/fraud. it's not clear if grad. response is ssensible here

<Walter> npdoty: if only as non-normative text

s/ssensible/sensible/

<npdoty> Walter, worth looking at the existing options and minimization about retention before proposing a new option

dwainberg: but the text only says 'preferred' (which is not a must)

aleecia: if we believe that there is nothing sensible about grad. response, we should delete the text entirely.

<npdoty> dwainberg, aleecia, it's possible that we can define the debugging permitted use without a definition of "graduated response"

<Walter> npdoty: also boils down to whether we repeat stuff or not, I'm currently of the opinion that implementers will not familiarise themselves with the text as a whole so some redundancy is in order

…we had an understanding what that means for debugging, but we're less clear on sec/fraud

<ifette> ifette: How is it that here we have a dependency but that elsewhere (see our discussion at the beginning of the call on unlinkable data) we think we can define the term as a separate exercise from its usage in the document

…suggests a defn before we move through

<fielding> +1 ifette's text on graduated response

<npdoty> ... for example, it's only non-normative in my updated text on debugging: http://npdoty.name/w3c/middle-way#debugging

ifette: see thoughts in IRC

<fielding> http://lists.w3.org/Archives/Public/public-tracking/2012Oct/0506.html

<npdoty> Walter, agree that we need to settle on repetitions, though I personally think we can separate that question (and probably leave it to the editors)

<fielding> well, "graduated response" is not an independent term, since it depends on context to have any meaning at all

aleecia: people were having a hard time imagining how grad response could be even relevant to sec/fraud, whereas on unlinkable we have different issues (is it a get out of jail free pass?), so that makes a big difference to the discussion

tlr: an admin question, what issue?

<Lmastria-DAA> +1

<Lmastria-DAA> +q

<rigo> Graduated response (ACTION-279)

<Zakim> rigo, you wanted to say that graduate response is just an application of the already agreed principles

<Walter> npdoty: and also, in my experience a mere data minimisation requirement doesn't work in the real world, implementers will need additional guidance

<npdoty> I thought ifette's point in that email was that "graduated response" *could* be defined separately, and then explained in context for both debugging and security, perhaps with non-normative examples

<aleecia> Thomas, that's ISSUE-24 OPEN Possible exemption for fraud detection and defense

<tlr> thanks.

<fielding> closest would be issue-24?

<tlr> issue-24?

<trackbot> ISSUE-24 -- Possible exemption for fraud detection and defense -- open

<trackbot> http://www.w3.org/2011/tracking-protection/track/issues/24

<aleecia> It is, yes

<susanisrael> i understood ifette's point the same way npdoty did.

<ifette> npdoty, i agree it would be cleaner were it defined separately and explained in context. I was just saying that we seem to be taking a different approach here than w.r.t. unlinkable data

<ifette> rigo, sounds good

rigo: maybe no issue, but action 279. thinks we have violent agreement; the issue is that grad. response as a term is saying what youre kinda expected to do under the data minim and relevance requirements, it's a spelling out of the general principle. the text does the trick; whether we have the term, is less important.

<npdoty> ifette, hmm, I thought you were making an opposite point in email and over voice

<aleecia> yes, was short hand and a reference to the game "monopoly"

lmastria: worth pointing out: thinks none of us are here to get out of jail, or other similar purposes; can we stop using this as a short-hand? we're all doing this delb. and in good faith; the term doesn't help us move ahead

<Walter> Actually, I don't mind "get out of jail" because for truly unlinkable data I don't mind if you ignore the spec

aleecia: apologizes and references monopoly

<ifette> npdoty, in email I was saying that this particular term could be better defined _in the document_ as a separate term. That's a different matter to "when in the discussions of the WG, in relation to other discussions, do we talk about its definition"

aleecia: skipping to end of agenda and admin

…first, next week is t'giving in the US

<JC> +1

<ifette> I will not be able to join a call next Wed

admin

<ifette> I will be traveling for thanksgiving

<Chris_IAB> -1

<ifette> -1

<robsherman> -1

<ChrisPedigoOPA> -1

<Joanne> -1

<laurengelman> -1

<adrianba> -1

<WileyS> -1

<hwest> -1

<dwainberg> -1

<jeffwilson> -1

<Chapell> -1

…-1 - you cannot join

<Brooks> -1

<cOlsen> -1

-1

<marc> -1

<johnsimpson> +1

<Chris_IAB> -100

<Lmastria-DAA> -1

<rigo> -1

<jchester2> +1

<amyc> -1

<Walter> +1

<peter-4As> -1

<eberkower> +1

<vinay> -1

<npdoty> +1

<efelten_> +1

<Simon> +1

<fielding> -0

<moneill2> +1

<vincent> +1

<rvaneijk> +1

<Chris_IAB> -5 in this room at DAA

<susanisrael> -0

aleecia: call looks kinda thin

<suegl> -1

…cancel call? will talk to staff and conclude

<ifette> Also cannot attend hte workshop. It's literally the day after the biggest travel weekend in the US...

<fielding> fyi, -1/-0/+0/+1 are from Apache

<WileyS> Aleecia - X-mas and New years Holidays? Another face-to-face?

aleecia: the week after is the w3c workshop in Berkeley (beyond DNT) 26/27 and then we call on the 28th

how about the 28th?

<Walter> -1

<JC> +1

<ifette> -1

<Joanne> +1

<johnsimpson> =1

<efelten_> +1

<vinay> +1

<jchester2> +1

<vincent> +1

<WileyS> +1

<jeffwilson> +1

<ChrisPedigoOPA> -0

<Brooks> +1

<moneill2> +1

<adrianba> +1

<robsherman> +1

<npdoty> can you make a call on the 28th? -1 if you can't, +1 if you can

<dwainberg> -0

<fielding> +1

<npdoty> +1

<Lmastria-DAA> what is the agenda for berkeley?

<susanisrael> +1

<Chapell> -1

<tlr> -1

<peter-4As> +1

<cOlsen> +1

<eberkower> +1

<Simon> -1

<amyc> +1

<johnsimpson> +1

<rvaneijk> -1

<BrendanIAB> +1

<Chris_IAB> +/-1

<rigo> Lmastria-DAA: "out of jail" is not intended as an offense (my apologies) and a term we use regularly on this side of the Atlantic without further afterthought

<rigo> -1

<marc> -1

<johnsimpson> yep. can't type...

<Walter> Lmastria-DAA: indeed, it is merely a reference to the Monopoly board game

<npdoty> s/Lmastria-DAA:/lmastria-daa,/

<laurengelman> :)

<WileyS> Aleecia - thoughts on X-mas and New Years weeks? Another face-to-face?

<amyc> and new schedule?

aleecia: as we start to go through the calls for objections etc., we'll be creating a new 'announce' mailing list for announcements that are not for discussion, so you can see the next steps etc., what's officially happening

<ifette> Current charter expires 1/2013

aleecia: no f2f until the new year,

<ifette> Will we discuss charter issues?

<WileyS> Aleecia - understood - I meant do you expect another f2f at all - in 2013

<Chris_IAB> SO, another f2f next year?

<ifette> Putting it out there that I would object to another extension w/o group discussion on scope and progress

aleecia: thomas is cleanign up actions and issues, and so on.

<WileyS> Aleecia - would be helpful for me to know ASAP for budget planning purposes

<Chapell> +1 to Ian and charter renewal

<rigo> there will be a F2F on Global considerations in January. Kimon is planning for a nice southern european location

aleecia: we are adjourned. scribe faints

<npdoty> ifette, sure, definitely follow up with us

<Walter> rigo: my vote would go to someplace in France, the other countries will have too many riots

<npdoty> chair: aleecia

- DRAFT -

Tracking Protection Working Group teleconference

14 Nov 2012

Attendees

Contents

Texts nearing a call for objections poll

admin

Summary of Action Items

Scribe.perl diagnostic output