3 ms·

Trying to give a short reply, so think about two use cases: a) laptop as a back-end for complex measurement system. Connected to a high-voltage signal, sitting

Trying to give a short reply, so think about two use cases: a) laptop as a back-end for complex measurement system. Connected to a high-voltage signal, sitting in a crowded labs full of noise and cables. Would you like to reach out to type in a command every time, if you can speak or gesture instead? b) laptop put next to a rendering device (video, audio, whatever), close enough so a wide-band wireless becomes efficient. Would you like to have all rendering devices perform the same computation, closed-sourced, paid-licensing, full of non-addressed CVEs? Or you'd rather concentrate the same processing in one place -- your laptop -- source code open to public inspection etc?

coldtea 10y ago
>a) laptop as a back-end for complex measurement system. Connected to a high-voltage signal, sitting in a crowded labs full of noise and cables. Would you like to reach out to type in a command every time, if you can speak or gesture instead? First, wouldn't that be a totally niche use? I'm pretty sure people who can't type (e.g. no hand control due to some accident/condition) are far more than people in "crowded labs full of noise and cables" using their laptop as a "back-end for complex measurement system". So, I didn't say there are no uses for speech/gesture control (I even pointed to incapacitated users and such). I said it's gimmicky for most people, and this is not really a counter-example to that. But those things aside, if the lab is "crowded" and "full of noise" wouldn't that make speech commands either perform badly (due to the noise in the room), or further annoy the others in the lab (due to them adding to the noise)? That leaves gestures. But those wont work for complex commands, but just for stuff like "pick among a few options". Else you'd have to learn a whole gesture based language (like sign language). I'd take typing then -- or some special controller that encodes the options (e.g. a foot switch). >b) laptop put next to a rendering device (video, audio, whatever), close enough so a wide-band wireless becomes efficient. Would you like to have all rendering devices perform the same computation, closed-sourced, paid-licensing, full of non-addressed CVEs? Or you'd rather concentrate the same processing in one place -- your laptop -- source code open to public inspection etc? I'd call this out as another totally niche and contrived example. And my laptop is not "source code open to public inspection" anyway, much less the hardware in it. So why would I trust the processors in my laptop more than the processors in my rendering device? I bought both anyway, and have not inspected any of them.

coldtea 10y ago

>a) laptop as a back-end for complex measurement system. Connected to a high-voltage signal, sitting in a crowded labs full of noise and cables. Would you like to reach out to type in a command every time, if you can speak or gesture instead? First, wouldn't that be a totally niche use? I'm pretty sure people who can't type (e.g. no hand control due to some accident/condition) are far more than people in "crowded labs full of noise and cables" using their laptop as a "back-end for complex measurement system". So, I didn't say there are no uses for speech/gesture control (I even pointed to incapacitated users and such). I said it's gimmicky for most people, and this is not really a counter-example to that. But those things aside, if the lab is "crowded" and "full of noise" wouldn't that make speech commands either perform badly (due to the noise in the room), or further annoy the others in the lab (due to them adding to the noise)? That leaves gestures. But those wont work for complex commands, but just for stuff like "pick among a few options". Else you'd have to learn a whole gesture based language (like sign language). I'd take typing then -- or some special controller that encodes the options (e.g. a foot switch). >b) laptop put next to a rendering device (video, audio, whatever), close enough so a wide-band wireless becomes efficient. Would you like to have all rendering devices perform the same computation, closed-sourced, paid-licensing, full of non-addressed CVEs? Or you'd rather concentrate the same processing in one place -- your laptop -- source code open to public inspection etc? I'd call this out as another totally niche and contrived example. And my laptop is not "source code open to public inspection" anyway, much less the hardware in it. So why would I trust the processors in my laptop more than the processors in my rendering device? I bought both anyway, and have not inspected any of them.