Hi Mike, Unfortunately, I think that the specification needs considerable work before it would be considered to be acceptable. I've started filing issues, but I am finding it rough going. It is very hard to follow, it lacks basic explanations of key features and their operation (this information is encoded in a labyrinth of algorithms), and I don't think that the design is appropriate.
I still don't understand the proposed consistency mechanism completely (see above), but it looks like the current design is one where issuer commitments are delivered by a user agent vendor to each user agent, using an update or remote configuration system. This is not a bad way to achieve consistency (and maybe correctness), but it sacrifices other properties that are important, such as avoiding having browser vendors as gatekeepers in the system. Also, this expectation is not documented. Perhaps greater familiarity with this will come with time, but as of this moment, I would have to say that this is not ready to be interoperable. I've opened a few issues with larger concerns, though there are lots of minor things I've found that follow on from each of these. On Fri, Mar 31, 2023 at 11:04 PM Mike Taylor <[email protected]> wrote: > Hey Martin, > > On 3/31/23 3:38 AM, Martin Thomson wrote: > > I will note that the current state of the specification does not seem to > match IETF Privacy Pass documents. I think that shipping is premature on > that basis. > > Mozilla deferred our position on this because the specifications were not > in a particularly healthy state at the time. That situation doesn't seem > to have changed much. > > I think the spec has improved significantly from where it was just a few > months ago, that said... > > More concerning is the lack of a widely acceptable key consistency and > correctness mechanism. A more rigorous analysis of the information > transfer properties of the proposed system will be needed before we can be > confident that this is OK to ship. > > It would be great if you could file issues with these concerns: > https://github.com/WICG/trust-token-api/issues/new > > (I'm sorry Steven that I didn't notice this before I had a chance to > discuss this in person this week, but I've been overwhelmed and blink-dev > isn't something I watch closely.) > > On Sat, Mar 18, 2023 at 4:29 AM 'Steven Valdez' via blink-dev < > [email protected]> wrote: > >> Folks from Mozilla have done some recent analysis on the privacypass >> protocol and some supportive of the general protocol, however we haven't >> gotten any newer signals on whether the PST system where some sites are >> issuers and other sites redeem tokens is of interest to them. Apple has >> been pursuing Private Access Tokens, which is a version of privacypass with >> the device vendor acting as the issuing party. >> >> On Fri, Mar 17, 2023 at 12:46 PM 'Mike West' via blink-dev < >> [email protected]> wrote: >> >>> I'm quite excited to see this ready to ship, thanks for the work you've >>> put into it over the years. >>> >>> Both Mozilla and Apple's positions seem dependent upon analysis of the >>> underlying Privacy Pass protocol. Have you had additional communication >>> with them about how things are going, since it's been a while since the >>> initial request in both cases? >>> >>> -mike >>> >>> On Friday, March 17, 2023 at 5:35:06 PM UTC+1 Steven Valdez wrote: >>> >>>> Contact emails >>>> >>>> [email protected], [email protected], [email protected] >>>> >>>> Explainer >>>> >>>> https://github.com/WICG/trust-token-api >>>> >>>> NB: We'll rename the repository to private-state-token-api when it's >>>> adopted by the antifraud CG. >>>> >>>> Specification >>>> >>>> https://wicg.github.io/trust-token-api >>>> >>>> Design docs >>>> >>>> >>>> https://docs.google.com/document/d/1TNnya6B8pyomDK2F1R9CL3dY10OAmqWlnCxsWyOBDVQ/edit >>>> >>>> Summary >>>> >>>> The Private State Token API is a new API for propagating user signals >>>> across sites, without using cross-site persistent identifiers like third >>>> party cookies for anti-fraud purposes. Anti-fraud methods that rely on >>>> third party cookies will not work once third party cookies are deprecated. >>>> The motivation of this API is to provide a means to fight fraud in a world >>>> with no third party cookies. The API prevents cross-site identification by >>>> limiting the amount of information stored in a token. Blind signatures >>>> prevent the issuer from linking a token redemption to the identity of the >>>> user in the issuance context. >>>> >>>> Private State Token API does not generate or define anti-fraud signals. >>>> This is up to the corresponding first party and the token issuers. The API >>>> enforces limits on the information transferred in these signals for privacy >>>> concerns. Private State Token API is based on the Privacy Pass >>>> protocol from the IETF working group >>>> <https://datatracker.ietf.org/wg/privacypass/about/>. It can be >>>> considered as a web-exposed form of the Privacy Pass protocols. >>>> >>>> The Private State Token API was formerly known as the Trust Token API. >>>> It is renamed to more accurately reflect its functionality. >>>> >>>> Blink component >>>> >>>> Internals>Network>TrustTokens >>>> >>>> NB: As a part of the process of renaming the Trust Token API to the >>>> Private State Token API, the blink component will also be renamed. >>>> >>>> >>>> TAG review >>>> >>>> https://github.com/w3ctag/design-reviews/issues/414 >>>> https://github.com/w3ctag/design-reviews/issues/780 >>>> >>>> TAG review status >>>> >>>> No concerns, aside from lack of clear interest from other browsers >>>> >>>> Risks >>>> >>>> Interoperability and Compatibility >>>> >>>> We intend to update the underlying cryptographic and token issuance >>>> protocols to align with the eventual Privacy Pass standard. This will >>>> affect compatibility with the small number of token issuers. Private State >>>> Token API fetch requests include a token type and version field that >>>> enables backward compatibility while allowing possible extensions for >>>> future token types and versions. While we will have a standard >>>> deprecation path of supporting multiple versions, we expect this to be >>>> easier with this API as each issuer using this API will need to register to >>>> become an issuer and will provide contact information as part of that >>>> process. >>>> >>>> >>>> Gecko: Defer >>>> <https://mozilla.github.io/standards-positions/#trust-token> >>>> >>>> WebKit: Pending ( >>>> https://github.com/WebKit/standards-positions/issues/72), already >>>> shipping similar technology >>>> https://developer.apple.com/news/?id=huqjyh7k (see PST vs. PAT >>>> <https://github.com/WICG/trust-token-api/blob/main/PST_VS_PAT.md> for >>>> more information about the differences in the technologies). >>>> >>>> Web developers: Positive >>>> >>>> A limited set of developers provided feedback on Private State Tokens, >>>> indicating that the tool was valuable for anti-fraud capabilities while >>>> also acknowledging some utility challenges (1). Other developers also found >>>> that Private State Tokens provided ability for authentication purposes (as >>>> illustrated by its use in the Privacy Sandbox k-Anonymity Server) (2). >>>> >>>> 1: >>>> https://github.com/antifraudcg/meetings/blob/main/2022/yahoo-trust-token.pdf >>>> >>>> 2: >>>> https://github.com/WICG/turtledove/blob/main/FLEDGE_k_anonymity_server.md#abuse-and-invalid-traffic >>>> >>>> >>>> Other signals: >>>> >>>> Ergonomics >>>> >>>> N/A >>>> >>>> >>>> Activation >>>> >>>> Using this feature requires spinning up a (or partner with an existing) >>>> Private State Token issuer that can issue and verify trust tokens, which is >>>> non-trivial. Verifying properties of the Signed Redemption Record or the >>>> client signature requires additional cryptographic operations. It would be >>>> beneficial to have server-side libraries that developers can use to help >>>> make using this API easier. Sample code can be found at >>>> https://github.com/google/libtrusttoken. >>>> >>>> >>>> >>>> Security >>>> >>>> N/A >>>> >>>> >>>> WebView application risks >>>> >>>> Does this intent deprecate or change behavior of existing APIs, such >>>> that it has potentially high risk for Android WebView-based applications? >>>> >>>> As this feature does not deprecate or change behavior of existing APIs, >>>> we don't anticipate any risk to WebView-based applications. >>>> >>>> >>>> Debuggability >>>> >>>> This API is debuggable via the DevTools Application Data panel and the >>>> operations are exposed in the Network panel. >>>> >>>> Will this feature be supported on all six Blink platforms (Windows, >>>> Mac, Linux, Chrome OS, Android, and Android WebView)? >>>> >>>> Yes >>>> >>>> Is this feature fully tested by web-platform-tests >>>> <https://chromium.googlesource.com/chromium/src/+/main/docs/testing/web_platform_tests.md> >>>> ? >>>> >>>> Yes >>>> <https://wpt.fyi/results/trust-tokens?label=experimental&label=master&aligned>*, >>>> some of the tests are currently failing as renaming/API changes in >>>> preparation for shipping these feature haven't propagated to those tests >>>> yet. Additionally, due to the requirements of having a server-side issuer >>>> (with bespoke crypto) to fully test the API, a majority of the testing is >>>> done in wpt_internal with a bespoke python implementation of a PST issuer. >>>> >>>> Flag name >>>> >>>> TrustTokens (in the process of being renamed to PrivateStateTokens) >>>> Requires code in //chrome? >>>> >>>> False >>>> >>>> Non-OSS dependencies >>>> >>>> Does the feature depend on any code or APIs outside the Chromium open >>>> source repository and its open-source dependencies to function? >>>> >>>> Yes. Token operations are dependent on having the key commitment >>>> information configured. Chrome (and Chromium implementations that consume >>>> components from component updater) supports this via a component, other >>>> clients will need to consume the component or come up with their own method >>>> of shipping the key commitment information to the client. >>>> >>>> Estimated milestones >>>> >>>> Chrome for desktop: 113 >>>> >>>> Chrome for Android: 113 >>>> >>>> Android Webview: 113 >>>> >>>> Anticipated spec changes >>>> >>>> Open questions about a feature may be a source of future web compat or >>>> interop issues. Please list open issues (e.g. links to known github issues >>>> in the project for the feature specification) whose resolution may >>>> introduce web compat/interop risk (e.g., changing to naming or structure of >>>> the API in a non-backward-compatible way). >>>> >>>> The major feature changes we expect are likely to be around the >>>> versions of tokens we support, as other use cases may need differing >>>> properties from those provided with the initial API and other format/API >>>> changes to align better with standardization and interop (see the >>>> Interoperability >>>> and Combatibility section up above). Most potentially web-observable >>>> changes in our open issues ( >>>> https://github.com/WICG/trust-token-api/issues) are around ergonomics >>>> of using the APIs and ways to use the API in more locations/manners which >>>> should pose minimal compatibility risk to existing users of the API. >>>> >>>> Link to entry on the Chrome Platform Status >>>> >>>> https://chromestatus.com/feature/5078049450098688 >>>> >>>> Links to previous Intent discussions >>>> >>>> Intent to prototype: >>>> https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/X9sF2uLe9rA >>>> >>>> Intent to experiment: >>>> >>>> >>>> https://groups.google.com/a/chromium.org/g/blink-dev/c/UIvia1WwIhk/m/stu7iXTWBwAJ >>>> >>>> Intent to extend origin trial: >>>> >>>> >>>> https://groups.google.com/a/chromium.org/g/blink-dev/c/fpfbKgJF8Vc/m/aC8HJfGdDwAJ >>>> >>>> >>>> This intent message was generated by Chrome Platform Status >>>> <https://chromestatus.com/>. >>>> >>>> >>>> >>>> >>>> -- >>> You received this message because you are subscribed to the Google >>> Groups "blink-dev" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To view this discussion on the web visit >>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/633d35ef-2bd9-43c3-9799-8243ff24d764n%40chromium.org >>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/633d35ef-2bd9-43c3-9799-8243ff24d764n%40chromium.org?utm_medium=email&utm_source=footer> >>> . >>> >> >> >> -- >> >> Steven Valdez | Chrome Privacy Sandbox | [email protected] | Cambridge, >> MA >> -- >> You received this message because you are subscribed to the Google Groups >> "blink-dev" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CANduzxB9WoK8nJ4J_R8L7cyb6JtidzJopoN5xxSdS4--4kbEuQ%40mail.gmail.com >> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CANduzxB9WoK8nJ4J_R8L7cyb6JtidzJopoN5xxSdS4--4kbEuQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- > You received this message because you are subscribed to the Google Groups > "blink-dev" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPLxc%3DXb_rRjf0PcjbvQH7i2ctwx1P-etgGsP4NfxJPbd42QtQ%40mail.gmail.com > <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPLxc%3DXb_rRjf0PcjbvQH7i2ctwx1P-etgGsP4NfxJPbd42QtQ%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > > -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPLxc%3DX0AWrt9pcS7XG5EVHkmFuGX%2BGWVtCX_w0VYCDxJfn3hQ%40mail.gmail.com.
