<wseltzer> https://lists.w3.org/Archives/Public/public-web-adv/2020Sep/0002.html

<wseltzer> https://lists.w3.org/Archives/Public/public-web-adv/2020Sep/0001.html

Agenda-curation, introductions

<scribe> Scribe: Karen Myers

Wendy: Let's start with some agenda curation and introductions, given a relatively light agenda
... We have messages from Arnaud and James with additional agenda suggestions
... Suggest that we talk about floc availablility; Turttledove and Gatekeeper simulation
... James wanted to bring our attention to decentralized web interest group
... Any other items for agenda suggestion; highlights people would like to raise?
... All right, I would like to start
... with a brief conversation...start with introductions
... Do we have anyone new to the group who would like to introduce yourself?
... Hearing no one

Agenda planning for TPAC vF2F and breakouts.

Wendy: Briefly talk about the virtual F2F
... I've been collecting ideas
... having lots of conversations 1:1
... and will be reaching out to more people
... and invite others to reach out to me
... We have proposed two, four-hour blocks
... from time zone poll
... I did not see any conflicts raised on 21-22 October, Wed/Thursday of TPAC meeting week
... I proposed that as our dates
... we will be concentrated in UTC +2 to -7
... for Europe/Americas segment
... and we have a few people from elsewhere

<wseltzer> https://github.com/w3c/web-advertising/blob/master/meetings/TPAC2020.md

Wendy: I think the block of time listed in the proposed agenda page is probably a reasonable match
... with apologies to those for whom it will be late at night or early morning
... hard to get a time that's good for everybody
... That leaves us with a block of time to fill
... I heard suggestions that we do some backgrounders, some roadmapping
... of technologies; think about where we are and where they might be going in the W3C standardization process
... Some technical discussions, issues that would benefit from sustained time
... that would benefit more time to talk through
... welcome others who want to lead discussions, or raise questions they would like to hear answered
... See if we can bring those up

<btsavage> Where should be propose ideas for topics to discuss?

Wendy: I've also heard interest in hearing from a broad swatch of browser implementers, including those who are not participating in this group
... See if we can invite them to the discussion
... listen to various browser perspectives

Wendell: you covered a lot of what I was going to tee up and propose
... At Verizon we found the session from major browsers, prefiguring directions
... they had slides and said they were going to do it
... we found it super useful; explains mindset and directionality
... what we face at our company
... and around the industry is
... seeing it in slides in the tradedress of companies, with or without dates, is good
... Good to get vendors to voice
... or for someone who reads trade press to review
... but there is a desperate need to figure out what is going on
... Like suggestion about browsers
... What is unclear is whether any of this stuff works
... not a peer reviewed paper; but if it works at scale for organizations
... goes under diversity heading for wider participation

<Chapell> +present

Wendell: Web doesn't have an identity layer today
... two proposals
... isLoggedIn and webid
... specific track on that would be interesting
... Advertising trade is moving to use email names as the identifier
... talked about if you don't value the content, then putting name next to what you are reading is the prosposal
... Would be good to review the discussion
... lots of discussion on WICG
... haven't heard too much
... will wrap up there

Wendy: Thank you, Wendell
... key is finding the intersection of what people can freely talk about
... without getting @ promises
... how can we build multiple sides of the API
... so builders on both sides can synchronize their rollouts and build components necessary for browsing machinery and adtech to work together
... and so on the other side of the API
... I've also wondered whether getting conversation on the use case document
... from the perspective of the users
... or where is there strong interest in seeing features to support use cases
... which use cases are critical

Kris: I was going to add another topic I would love the browser vendors to add on to their presentation
... Third party v first party context
... Adv is focused on third party
... also concerned about changes when things are in first party context
... Like to get more thoughts from browsers where they are going there
... that's it

Ben: I'd like to propose a session where we talk about the current state of the art with MPC
... secure multi
... FB has done a lot of open source reserach
... talk about state of the art
... what can be accomlished in MPC
... what browsers can / cannot do

<dkwestbr> +1 on Ben’s suggestion of state of the art MPC

Ben: and aggregator reporting API

Wendy: thanks, Ben
... to the extent that some of the
... people might be outside the regular participants of the group
... if we get agenda put together early enough, we might get others to respond and participate
... potentially a topic at a different technical level

Brian May: I wanted to add to what Wendell said about IDs

scribe: I know of two other IDs

<wbaker_> +1 MPC state-of-research

scribe: Trade Desk 2.0 and LiveRamp identity link
... would be useful to get some kind of survey of major addressability
... uptake, restrictions
... and accountability aspects
... which most addressability proposals I have seen includes
... Accountability being how to include users' consent and intent

Wendy: thanks
... if you have suggested participants to address that or to help frame the questions, that would be great
... Excellent, thank you for that input
... any further comments on this?

Kris: Along with webid
... and LoggedIn, would be great to hear where we are with trust tokens

Wendy: We have lots of proposals in our repository links
... where they are now and understanding landscape
... good to hear

Kris: Just add onto
... I definitely like the updates
... on what we have listed
... but it would be lovely to get from the browsers
... some sort of ranking
... on how far along these proposals are
... as in which ones they think will be coming out sooner rather than later
... a roadmap

<wbaker_> +1 orderly transition to the future

Kris: or if we could weight them for how far along things are

Wendy: some of that is in the W3C context, consideration of if this is incubation or req track
... are there multiple, interoperable implementations
... getting a sense of where we would likely find the ultimate implementations is helpful in looking at what to move into working groups for standardization
... as well as for what to expect for those who might be using them
... Thank you for all of those suggestions
... Suggest we move on
... Arnaud, you suggested several different topics
... I invite you

Arnaud: Hi everyone
... Let's start with the first one

*FLOC: Availability for experimentation: *

Arnaud: Let Lionel go through floc

Lionel: We saw floc
... want to know goal of these developments
... run live tests; what is goal, timeline around that

Josh: I'm Google developer working on Chrome and doing floc development
... give you status
... we are building out code
... the the clustering algorithm is simple; proof of concept
... for now intent is to write some simple code and then do the privacy analysis
... cannot do anything else if we don't meet the privacy goals
... figure out how to do the analysis
... once we are further along, we will publish an update to the explainer
... assuming it goes well, we will
... Where we are; still very early

Lionel: Thank you for the answer
... do you have any idea of the timeline

Josh: So, we expect in the next few weeks to be able to talk about the privacy experiment publicly
... beyond that, no

<lbasdevant> Thanks

Robert, Neustar: Good question on privacy analysis

scribe: two parts; how it stays private on Google server
... if differential on cluster outcomes can be maintained
... is privacy analysis about secure multiparty computation
... what is centralized is not revealed to Google
... or differential privacy determined level

Josh: My first meeting

Wendy: please respond

Josh: How are we going to do the privacy analysis
... short term and long term answers
... short term to figure out how many people are in the clusters and differential privacy
... I think that will be different from a long-term solution; keep in the wild
... and not know about he clusters
... figure out if it is plausible
... and make sure no one can see anything about the clusters

Robert: You are testing the outcomes?

Josh: We are trying to nail that down in the next couple of weeks

Josh Koran: You mentioned counting how many people; I'm assuming browser IDs rather than real people?

Josh: Good question
... probably this needs to be at the browser ID level; I think you are right; not actual ID, individual browsers

Robert: Must exist for cluster label to be returned to the browser

<kleber> The last comment in this FLoC issue talks about some of this: https://github.com/jkarlin/floc/issues/18

Robert: if it is purely anonlymous you cannot read back; must be a browser ID

Josh: yes, like a browser instance, I agree

Joshua: browser ID rather than a real person

Josh: Right; we have no interest in capturing information about a person

Wendy: Any further question on floc; thanks Josh for joining us

Kris: I understand, not identifying back to a person
... but are you trying to group across different devices, being associated with one another
... or browser instances

Josh: I cannot speak to all of it
... at the moment we are trying to understand if a device browser can come up with a cluster that has privacy properties
... none of this includes cross-device integration at the moment

Kris: ok, thanks

Aram: Can you hear me?

Wendy: yes

Aram: I was wondering do any of the Google experiments plan to use LoggedIn to browser experiments
... sounds like you don't want to use those, but if not the case, it's a no

Josh: That is the kind of detail we are hammering out now and plan to publish very soon

Aram: Thanks

Wendy: Thank you for that status update
... anything at end of the queue on that one?
... Arnaud, do you want to take the next item

*TURTLEDOVE simulation:*

Arnaud: Second one also related to testing
... RTB House to do implementation
... huge effort
... we think it is running in the right direction
... what we would really like to see
... is something to allow publishers and advertisers to do a simulation
... of how everything works
... how to fill the pipes
... and we are wondering if Chrome or Google team is working on a simulation, or game-like simulation for actual users to see how they would work with it

Michael: I don't think we have anybody on the call
... from the Google team that published the RTB-experimental proposal on Github
... let me see if I can find the link to that
... there we go

<kleber> https://github.com/google/rtb-experimental

Michael: There is a group inside of Google that has said they are working on discussions with other people in RTB
... real time binding with how to experiment with versions of these proposals
... not yet available in browser
... but where people in adtech want to experiment
... when APIs actually exist
... I don't think anyone is on the call today
... but that is the area for potential simulation for how idea could be taking place
... Not a Chrome effort but between Google and other adtech partners

Wendy: several people queuing up
... may need to wait until we have that Team here to join us

Arnaud: second thing is the initiative
... what they propose
... is not really on the reporting side of things; what we are missing is some visibility
... why we tried to do with the scripts that we did; not very user friendly
... having something that would really show
... how they would access business data would be useful to them
... and start to build strategies
... This Google RTB initiative does not use browser
... they act as gatekeeper
... Google RTB would be some kind of gatekeeper
... I see some problems with simulation
... does not reflect the fuller inbrowser experience
... Would be curious for your thoughts on that

Michael: for first point about reporting
... you must be thinking about Turtledove
... floc doesn't have huge impact on reporting
... I agree; what aggregate reporting to get out of TD
... issues on the Sparrow proposal
... seems up in the air to me
... answer is pretty ambiguous so far
... Don't now if RTB experimental folks have efforts on what reporting will look like, I don't know
... Second question you asked about Google doing simulation
... you understand it is difficult to provide...without experimental efforts in the browser
... RTB is excellent; some integration with RTB House for how things are implemented in browser and what server comms pipeline could look like
... could approach the more end to end simulation you are looking for

Arnaud: Were we to go on RTB simulation...it relies on classic cookie tracking
... give a false idea
... affects performance
... as such question
... not really sure how useful such an experiment would be
... second things is technicalities; conclusion for what it means for the ecosystem; not sure what insights we can draw

Michael: I understand
... until Chrome has something ready to do in the browser that would require the plans to be more subtle, not sure what we can offer

Angelina: I want to make a request
... if we could have the RTB team in the next call
... Love to understand their measurement work
... I lead IAB group
... hard to get documentation on how they measure
... some feel it's proprietary
... lots of steps, tedious to document
... but it's important to get feedback from publishers and ad community on how they will get data; and how they want to leverage data
... and how to target people. It's a big concern right now

Michael: I am happy to pass along to RTB experimental people that they are invited to chat with this group at a future meeting
... Caution
... that talking publicly about some sensitive issues is difficult
... and a roadblock to getting people to share information
... Outcomes aimed at influencing future roadmap inside of Chrome, those need to be made public
... Chrome has hard time with data that cannot be seen
... But if people want to do their own simulations without providing details of what they are doing
... it's hard for us to take a position, or to have much influence there

Angelina: I totally understand that
... but also some sensitivity to just get these folks to participate
... I deal with it every day
... are there surveys, focus groups we can do, not necessarily an open forum
... to get feedback from buy and sell side on these issues

Michael: I will pass along this request
... to get the RTB experimental folks to come to this call

Wendy: Tried to close cue, but there is more interest

Westbrook: Just I will be quick
... are you able to say about RTB group
... only asking here, scope out on our end who would attend
... can you clarify further
... are you referring to RTB holistically
... I think only proposal I have seen so far
... is Turtledove
... is that what you are scoping out

<kleber> look at https://github.com/google/rtb-experimental

Westbrook: if premature, I understand, but if you can clarify

Michael: just dropped link on what was already published on experimental github repository
... list of different proposals they are experimenting with
... floc, TD, trust budget privacy tokens

D: Thank you

Kris: I actually reached out to the Google folks
... and offered Salesforce help
... instead of asking Google to disclose, perhaps ask if there are companies in an experimental stage
... I spoke a month ago, and they were at early stages
... I would not expect RTB experimental work right now is that farther along than any of the rest of us

Wendell: I'll take a different tack
... I think what everyone is experiencing here, showing industry frustration on where we are now; those of use who are not browser builders
... directed to the browser builders who have chosen to show up; thank you
... we are looking for orderly way for transition
... some folks writing long essays and issues
... they don't have kind assessments; assumptions of a vast conspiracy
... quote Feburary 2022 when Google announced cookies would go away
... Apple said @
... All feeling frustrated
... for a lot of companies on this call, whole lines of business will go away
... if there is some way to channel that frustration and anger to give more information
... or give complete certitude that a certain practice will not be allowed
... this weird netherworld
... maybe it will stop, may be not
... wrap up with the frustration level is increasing for a lot of folks here

Wendy: Thank for keeping the tone constructive even as we express frustration

<alextcone> +1 to Wendell's comments

Brad: Respond to request on how to get propropietary information into sensitive forum
... TAG group, Internet architecture board from IETF
... if IAB wants to co-host something, we could likely set up the right structures
... and have a report to publish the salient bits

Wendy: Thank you
... maybe that is another brainstorming topic for the virtual F2F

<AramZS> Hi, we're running low on time, if there's a link to the distributed web group can we share that here before we run out of time?

Wendy: are there places where we need a Chatham House Rules type forum to discuss data sharing
... thank you, Aram for reminding us of the time
... recognizing that Arnaud, has one more issue
... move to James to suggest the decentralized IG

Decentralized Web Interest Group

James: A big thank you to everyone who joined the meeting
... and for the drafting; multiple angles and inputs
... we decided on Decentralized Web IG name
... it is now in charter draft
... where charters go to be discussed
... and goes to Advisory Committee for their review shortly

Wendy: Just more internal coordination to send out the "advance notice"
... discussion among AC, then it can go for formal AC review; the Advance Notice will happen shortly

James: a bit of gap with W3C team
... as soon as the Advance Notice comes out, we can let this group know

<wbaker_> Decentralized Web? https://www.w3.org/2019/09/did-wg-charter

James: Primary objective is to take subjects that impact centralization or decentralization of certain design choices
... provide capacity; input; guidance
... through the group, commenting as that goes to AC
... and finally we want participation to go beyond engineering; include policy
... short text that took time to write
... Want to thank Wendy and others who contributed

<wbaker_> Decentralized Web ? https://internetpolicy.mit.edu/research/decentralized-information-group/

James: interested to get your feedback through Github, email

Aram: can I suggest changing the name
... the decentralized web has other meanings, other groups exists; things around blockchain that is tagged the decentralized web
... I realize that is not your intent, but wanted to raise the issue
... Understand this conversation came out of the success criteria
... does it make sense to remove the success criteria doc from this group's repository?

Lionel: Thank you
... one question; how do you see the decentralized group interacting with what we do here

James: great questions and observations
... another IG within W3C, would work with any other group that wants to interact
... would produce guidance
... and skills available to inform work; this group or any other group
... it takes time for a group to be chartered
... preference would be to keep docs within this group
... continue to progress
... in parallel with this group being chartered

<AramZS> That sounds fine to me, leaving it in.

James: at least 50% of discussion was about the name
... we felt that summarized multiple issues
... won't go through that now; welcome advice on other names
... we may have already considered it, but thank you for raising it and we welcome other comments

Wendy: the next step, Advance Notice, opens it up to conversations that might include input from some in this group and others
... who may suggest others
... maybe someone will have a brilliant name that reduces confusions
... that dialogue is welcome in the charter drafts repository where this draft is sitting

<jrosewell> Comments can be raised here as issues with the prefix [decentralized]. https://github.com/w3c/charter-drafts/issues

<jrosewell> https://w3c.github.io/charter-drafts/decentralized-charter.html

<lbasdevant> https://github.com/WICG/sparrow/blob/master/gatekeeper_certification.md

<wseltzer> [adjourned]

<wseltzer> wseltzer: notes for future meetings: Arnaud re Gatekeeper certification process

<wseltzer> chair: wseltzer

This is scribe.perl Revision of Date 
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Succeeded: s/logged in/isLoggedIn/
Succeeded: s/Ryan:/Brian May:/
Succeeded: s/@/the clustering algorithm/
Succeeded: s/else/else if we don't meet the privacy goals/
Succeeded: s/to the/an update to the/
Succeeded: s/@/differential privacy/
Succeeded: s/@ at/differential privacy/
Succeeded: s/@/Robert/
Succeeded: s/individual users/individual browsers/
Succeeded: s/really/really on the/
Succeeded: s/ID/idea/
Succeeded: s/who they are talking to/their measurement work/
Succeeded: s/that/the Advance Notice/
Succeeded: s/that will/the Advance Notice will/
Present: blassey mlerra ionel arnaud_blanchard dialtone seanbedford kleber dinesh-pubmatic lbasdevant mjv bleparmentier pl_mrcy ajknox AramZS pbannist hober Karen cpn Chapell gjlondon dkwestbr jrosewell hcai
No ScribeNick specified.  Guessing ScribeNick: Karen
Found Scribe: Karen Myers
Agenda: https://lists.w3.org/Archives/Public/public-web-adv/2020Sep/0000.html

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)