> I didn't have a way of running that code and only giving it access to a single USB device and nothing else.
To be honest I think that's the most compelling case for webUSB today. If desktop OSes had sandboxing tools (or more granular permissions in general) that are easily usable be everyday users, there would be no need to put webUSB in a browser sandbox. The cross-pltaform nature of it is nice, but that alone is not enough IMO. I think it would be interesting to see a Linux distribution where software that is not explicitly trusted (i.e. not installed by the system package manager) has no permissions by default. Interpreters make this more complicated but for binaries it could work.
To be honest I think that's the most compelling case for webUSB today. If desktop OSes had sandboxing tools (or more granular permissions in general) that are easily usable be everyday users, there would be no need to put webUSB in a browser sandbox. The cross-pltaform nature of it is nice, but that alone is not enough IMO. I think it would be interesting to see a Linux distribution where software that is not explicitly trusted (i.e. not installed by the system package manager) has no permissions by default. Interpreters make this more complicated but for binaries it could work.