4 ms·

>The whole point of Android certification is that you are following Google's rules. Whole point of WEI is that you are following Google's rules. You essential

>The whole point of Android certification is that you are following Google's rules. Whole point of WEI is that you are following Google's rules. You essentially acknowledged that YOU CANNOT become manufacturer - you become subcontractor for corporation that dictates what you can and cannot do. Which from point of trust is precisely that - anticompetitive behaviour.

charcircuit 3y ago
>Whole point of WEI is that you are following Google's rules. That would be the point of the Google Play attestor. With WEI sites can use anyone as an attestor. It is up for attestors to compete in providing a valuable signal to sites. It's not anticompetitive because you can come up with your own standard of a secure device and get sites to trust your attestor.
- zb3 3y ago
  Even if that'd be the case I'd still consider this API nothing more than a framework for enabling anticompetitive behaviour. But let's go back to the "anyone as attestor" argument. I don't see it in the API at all - I see only content binding..
  - charcircuit 3y ago
    When navigator.getEnvironmentIntegrity is called the browser can get an attestation from any single atestor willing to attest to it.
    - zb3 3y ago
      I'm unable to understand this. Since the API doesn't provide a way to force a specific atestor to be invoked, doesn't it mean it will be the browser that decides which atestor can run on a given platform? Of course then substitute browser for Chrome and we'll get the obvious outcome - Google will decide.
      - daurentius523 3y ago
        There are seriously so many conflicting informations that this should be hanged from implementation before it its fully formulated (aside fact that it is bad idea and should be scrapped). Firstly on Android Chrome implementation (testing) only Google Play is now implemented as attester and it does not seem that probable for it to change. (so presumably at least Chrome+Root wont work). Second and Most Important - the mechanism does not allow for "willing to attest it" - it actually contains this phrase "The web server then checks that the token came from an attester it trusts" so it is disingenuous to say that "any single atestor" is OK - it must be attestor that server decided beforehand (so if server decide we only trust Google and Microsoft - and you have Mac, then tough luck. [more realistically Linux will be at issue]) - so we are getting system that can exclude certain OS and Browsers permanently without recourse - by server side. @zb3 >doesn't it mean it will be the browser that decides which atestor can run on a given platform? That seems to be case in Chrome implementation* - and there is nothing about this in standard so unless other browsers would create another standard - this way seems likely. This would be disastrous for any decentralized OS like Linux or BSD - it would be completely impossible for it to realistically work as it requires single "platform" by design - and I should remark "there is no single uniform thing in Linux". * - note that server decides which attestors it trusts. Which means in essentia that sever decides which platforms it trusts. Again this must be scrapped.
- daurentius523 3y ago
  >It is up for attestors to compete in providing a valuable signal to sites. right because users will chose attestor - oh wait they will not. This system is designed (at least in Chrome) with PLATFORMS in mind not independent attestors - so you clearly didn't get the memo. >valuable signal rather highly intrusive signal. >It's not anticompetitive because you can come up with your own standard of a secure device and get sites to trust your attestor. Again it is anticompetitive when you produce a standard that benefits you (or biggest players) on market and makes nearly impossible for another competitor to emerge - it is even more anticompetitive when you create standard that "is impossible to implement" for architectural and 'branding' (ideological) reason by another competitor. Such situation can be understood as public attempt: to form cartel at best, monopoly at worst. Both are anticompetitive by nature - and would result in dissolution of company if it is registered in any country with working and not totally corrupt government. >secure device about "security": https://news.ycombinator.com/item?id=36985317 https://news.ycombinator.com/item?id=36985317

charcircuit 3y ago

>Whole point of WEI is that you are following Google's rules. That would be the point of the Google Play attestor. With WEI sites can use anyone as an attestor. It is up for attestors to compete in providing a valuable signal to sites. It's not anticompetitive because you can come up with your own standard of a secure device and get sites to trust your attestor.

zb3 3y ago
Even if that'd be the case I'd still consider this API nothing more than a framework for enabling anticompetitive behaviour. But let's go back to the "anyone as attestor" argument. I don't see it in the API at all - I see only content binding..
- charcircuit 3y ago
  When navigator.getEnvironmentIntegrity is called the browser can get an attestation from any single atestor willing to attest to it.
  - zb3 3y ago
    I'm unable to understand this. Since the API doesn't provide a way to force a specific atestor to be invoked, doesn't it mean it will be the browser that decides which atestor can run on a given platform? Of course then substitute browser for Chrome and we'll get the obvious outcome - Google will decide.
    - daurentius523 3y ago
      There are seriously so many conflicting informations that this should be hanged from implementation before it its fully formulated (aside fact that it is bad idea and should be scrapped). Firstly on Android Chrome implementation (testing) only Google Play is now implemented as attester and it does not seem that probable for it to change. (so presumably at least Chrome+Root wont work). Second and Most Important - the mechanism does not allow for "willing to attest it" - it actually contains this phrase "The web server then checks that the token came from an attester it trusts" so it is disingenuous to say that "any single atestor" is OK - it must be attestor that server decided beforehand (so if server decide we only trust Google and Microsoft - and you have Mac, then tough luck. [more realistically Linux will be at issue]) - so we are getting system that can exclude certain OS and Browsers permanently without recourse - by server side. @zb3 >doesn't it mean it will be the browser that decides which atestor can run on a given platform? That seems to be case in Chrome implementation* - and there is nothing about this in standard so unless other browsers would create another standard - this way seems likely. This would be disastrous for any decentralized OS like Linux or BSD - it would be completely impossible for it to realistically work as it requires single "platform" by design - and I should remark "there is no single uniform thing in Linux". * - note that server decides which attestors it trusts. Which means in essentia that sever decides which platforms it trusts. Again this must be scrapped.
daurentius523 3y ago
>It is up for attestors to compete in providing a valuable signal to sites. right because users will chose attestor - oh wait they will not. This system is designed (at least in Chrome) with PLATFORMS in mind not independent attestors - so you clearly didn't get the memo. >valuable signal rather highly intrusive signal. >It's not anticompetitive because you can come up with your own standard of a secure device and get sites to trust your attestor. Again it is anticompetitive when you produce a standard that benefits you (or biggest players) on market and makes nearly impossible for another competitor to emerge - it is even more anticompetitive when you create standard that "is impossible to implement" for architectural and 'branding' (ideological) reason by another competitor. Such situation can be understood as public attempt: to form cartel at best, monopoly at worst. Both are anticompetitive by nature - and would result in dissolution of company if it is registered in any country with working and not totally corrupt government. >secure device about "security": https://news.ycombinator.com/item?id=36985317 https://news.ycombinator.com/item?id=36985317