Consider collaborating with UACH on standardized "enum" values #174
Comments
domenic
pushed a commit
to whatwg/html
that referenced
this issue
Mar 30, 2022
navigator.platform was permitted to return the empty string due to a privacy-motivated spec change in ecefa6e. Since 2008, the understanding of changing things like this for privacy has evolved from returning special privacy values to returning a value that is already commonly returned for non-privacy reasons. Therefore, change the spec not to say "the empty string" but to permit returning a string commonly returned on another platform. Additionally, change the examples to be from the set of values given at https://www.chromium.org/updates/ua-reduction/ to avoid suggesting more-identifying and potentially-less-compatible values. But, for now, avoid actually constraining the permitted values to a reduced set. Further work on constraining these values is tracked mostly in https://github.com/whatwg/compat; see especially whatwg/compat#174.
|
Thanks for the suggestion, I think it's a good idea. I'll try to find the time to tackle this in Q2 (not putting a year, just in case™). |
mfreed7
pushed a commit
to mfreed7/html
that referenced
this issue
Jun 3, 2022
navigator.platform was permitted to return the empty string due to a privacy-motivated spec change in ecefa6e. Since 2008, the understanding of changing things like this for privacy has evolved from returning special privacy values to returning a value that is already commonly returned for non-privacy reasons. Therefore, change the spec not to say "the empty string" but to permit returning a string commonly returned on another platform. Additionally, change the examples to be from the set of values given at https://www.chromium.org/updates/ua-reduction/ to avoid suggesting more-identifying and potentially-less-compatible values. But, for now, avoid actually constraining the permitted values to a reduced set. Further work on constraining these values is tracked mostly in https://github.com/whatwg/compat; see especially whatwg/compat#174.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
https://wicg.github.io/ua-client-hints/ has gone further than https://compat.spec.whatwg.org/#ua-string-section by trying to lock down some "enum" values. Namely:
It might be worth having a common foundation, probably living in the Compat Standard, which these two share. I'm envisioning something like:
and then the UACH spec says to use the platform name, and the UA string part of the Compat spec says to pick a platform row and use the corresponding
<platform>and<oscpu>columns.The main thing here is not to waste work on stuff that we think will not exist in the UA string in the future. But I think platform is likely to stay, as is the mobileness bit.
The text was updated successfully, but these errors were encountered: