Explicitly spec out how non-k-anon auction works. #1080

morlovich · 2024-03-08T18:15:42Z

... This is a prerequisite to specing how it works with targetNumAdComponents, which has rather subtler
effects on what bids are made in each auction.

Preview | Diff

... This is a prereq to specing how it works with targetNumAdComponents.

morlovich · 2024-03-08T19:37:47Z

spec.bs

-          1. [=Score and rank a bid=] with |auctionConfig|, |generatedBid|, |leadingBidInfo|,
-            |decisionLogicScript|, |directFromSellerSignalsForSeller|, |dataVersion|, |auctionLevel|,
-            |componentAuctionExpectedCurrency|, and |topLevelOrigin|.
+            1. If |generatedBid| is not a failure:


Hmm, might need to register the original generatedBid for debugging if this one is a failure to get the reject reason = below threshold stuff out?

Though this appears that it may only work for PA, not for forDebuggingOnly?

https://source.chromium.org/chromium/chromium/src/+/main:content/browser/interest_group/interest_group_auction.cc;drc=3afa53375890e7144673d8e9e6356868295c2873;l=490

... which may make my pull request to the explainer a bit off :(

hmm, seems like a bug, that we may want to report this k-anon reject reason to forDebuggingOnly as well?

Yeah, probably. And I think my next refactors of this make us better suited for speccing it, so maybe don't need to worry about this for a moment.

qingxinwu

sorry for the delay. Some quick comments, and still need to take another look.

qingxinwu · 2024-03-13T01:43:30Z

spec.bs

@@ -5088,6 +5128,9 @@ result of [=evaluating a bidding script=], or an [=additional bid=] provided by
  : <dfn>no bid</dfn>
  :: A [=boolean=], initially false. True if there was a failure generating a bid, or the bidder
    chose not to bid. When true, [=generated bid/bidder debug loss report url=] is still considered.
+  : <dfn>for k-anon auction</dfn>
+  :: A [=boolean=], initially true. If this is false, the bid is only used to determine the hypothetical
+    winner with no k-anonymity constraints, which would be used to update k-anonymity counts only.


Can we link "update k-anonymity counts" to an algorithm, maybe [=Increment a winning bid's k-anonymity count=]?

What's the syntax for linking to an algorithm with text that's not just its name?

https://speced.github.io/bikeshed/#changing-lt

Thanks Dominic; done. (Though maybe I will want to tweak the name further)

spec.bs

qingxinwu · 2024-03-13T02:20:59Z

spec.bs

+            1. [=iteration/continue=].
+          1. Let |bidsToScore| be a new [=list=] of [=generated bids=]
+          1. If [=query generated bid k-anonymity count=] given |generatedBid| returns true:
+            1. Let |bidCopy| be a copy of |generatedBid|.


maybe "copy" -> "clone".
@domfarolino Is there a [=clone=] to link to for struct? I saw lists/sets/maps have one, but didn't find one for struct .

Yeah infra doesn't seem to have a definition for this and it probably should. Could you use the word "clone" manually, and then file a bug against the Infra Standard?

filed: whatwg/infra#634

spec.bs

qingxinwu · 2024-03-18T18:15:16Z

spec.bs

-          1. [=Score and rank a bid=] with |auctionConfig|, |generatedBid|, |leadingBidInfo|,
-            |decisionLogicScript|, |directFromSellerSignalsForSeller|, |dataVersion|, |auctionLevel|,
-            |componentAuctionExpectedCurrency|, and |topLevelOrigin|.
+            1. If |generatedBid| is not a failure:


hmm, seems like a bug, that we may want to report this k-anon reject reason to forDebuggingOnly as well?

spec.bs

dmdabbs · 2024-03-18T19:13:02Z

correctly I can order word.

😂

domfarolino

This mostly looks fine to me, I think I will defer to @qingxinwu on this PR.

qingxinwu · 2024-03-18T19:28:01Z

spec.bs

@@ -1700,7 +1700,7 @@ To <dfn>generate and score bids</dfn> given an [=auction config=] |auctionConfig
                 |generatedBidsForDebuggingReports|.
          1. [=list/For each=] |bidToScore| of |bidsToScore|:
            1. If |bidToScore|'s [=generated bid/for k-anon auction=] is true,
-                [=list/insert=] |bidToScore|'s [=generated bid/interest group=] in |bidIgs|.
+                [=list/append=] |bidToScore|'s [=generated bid/interest group=] in |bidIgs|.


nit: append ... to ...

Done, thanks

qingxinwu · 2024-03-18T19:41:50Z

spec.bs

+            1. [=iteration/continue=].
+          1. Let |bidsToScore| be a new [=list=] of [=generated bids=]
+          1. If [=query generated bid k-anonymity count=] given |generatedBid| returns true:
+            1. Let |bidCopy| be a copy of |generatedBid|.


filed: whatwg/infra#634

qingxinwu · 2024-03-18T19:43:02Z

This mostly looks fine to me, I think I will defer to @qingxinwu on this PR.

Just wanted to add you to review this, but you beat me. It LGTM % a nit.

SHA: 28abf64 Reason: push, by JensenPaul Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

SHA: 28abf64 Reason: push, by morlovich Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Explicitly spec out how non-k-anon auction works.

7148d90

... This is a prereq to specing how it works with targetNumAdComponents.

morlovich added the spec Relates to the spec label Mar 8, 2024

Fix various formatting mishaps

740a93d

morlovich requested a review from qingxinwu March 8, 2024 19:00

morlovich commented Mar 8, 2024

View reviewed changes

qingxinwu reviewed Mar 13, 2024

View reviewed changes

spec.bs Outdated Show resolved Hide resolved

spec.bs Outdated Show resolved Hide resolved

spec.bs Outdated Show resolved Hide resolved

Maks Orlovich added 3 commits March 13, 2024 09:38

Some easy parts of feedback

5b6b0c1

non-k-anon => non-k-anon-enforced

c5404fc

Link update algo

e7263f9

morlovich mentioned this pull request Mar 14, 2024

Spec multibid #1055

Merged

Maks Orlovich added 2 commits March 18, 2024 11:07

Merge branch 'main' into spec-k-anon

1030d77

Don't double-count duplicates.

9340c7d

qingxinwu reviewed Mar 18, 2024

View reviewed changes

Maks Orlovich added 2 commits March 18, 2024 15:03

Fix k-anon auction tie break; list-append

da739c3

correctly I can order word.

5a5f96d

domfarolino reviewed Mar 18, 2024

View reviewed changes

qingxinwu mentioned this pull request Mar 18, 2024

Add a clone method for struct whatwg/infra#634

Open

qingxinwu approved these changes Mar 18, 2024

View reviewed changes

to -> in

c021fb7

JensenPaul merged commit 28abf64 into WICG:main Mar 18, 2024
2 checks passed

github-actions bot added a commit that referenced this pull request Mar 18, 2024

Explicitly spec out how non-k-anon auction works. (#1080)

d4329c1

SHA: 28abf64 Reason: push, by JensenPaul Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Explicitly spec out how non-k-anon auction works. #1080

Explicitly spec out how non-k-anon auction works. #1080

morlovich commented Mar 8, 2024 •

edited by pr-preview bot

Loading

morlovich Mar 8, 2024

morlovich Mar 8, 2024

qingxinwu Mar 18, 2024

morlovich Mar 18, 2024

qingxinwu left a comment

qingxinwu Mar 13, 2024

morlovich Mar 13, 2024

domfarolino Mar 13, 2024

morlovich Mar 13, 2024

qingxinwu Mar 13, 2024

domfarolino Mar 13, 2024

qingxinwu Mar 18, 2024

qingxinwu Mar 18, 2024

dmdabbs commented Mar 18, 2024

domfarolino left a comment

qingxinwu Mar 18, 2024

morlovich Mar 18, 2024

qingxinwu Mar 18, 2024

qingxinwu commented Mar 18, 2024

Explicitly spec out how non-k-anon auction works. #1080

Explicitly spec out how non-k-anon auction works. #1080

Conversation

morlovich commented Mar 8, 2024 • edited by pr-preview bot Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

qingxinwu left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dmdabbs commented Mar 18, 2024

domfarolino left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

qingxinwu commented Mar 18, 2024

morlovich commented Mar 8, 2024 •

edited by pr-preview bot

Loading