- From: Tobie Langel via GitHub <sysbot+gh@w3.org>
- Date: Fri, 02 Sep 2016 11:46:44 +0000
- To: public-device-apis-log@w3.org
[Section 6 (Concepts)](https://w3c.github.io/sensors/#concepts) and
[Section 10
(Extensibility)](https://w3c.github.io/sensors/#extensibility) are the
key ones here.
The differentiation made between low and high-level sensors in GS
aren't to do with sensor fusion (though generally, most high-level
sensors are fused), but about whether the Sensor object describes a
kind of physical sensor (e.g. a pressure barometer) or the kind of
data you get out of it (e.g. altitude), _regardless of by which mean
you actually get that data (in the case of altitude, it could be using
a pressure barometer or GPS for example). _(As a sidenote, whether
fusion happens are the SF or HW level is irrelevant for this
classification.)_
With that distinction in mind, LinearAcceleration would fall in the
high-level sensor category (there's fusion going on between a
gyroscope and an accelerometer, you're describing the kind of data you
get and not the means by which you get it, that data could be
obtained through different means, albite with much less precision,
e.g. extrapolating from GPS readings).
So in the accelerometer case, the only thing the `Accelerometer`
sensor object should provide is raw data accelerometer output
including gravity. Note that some platforms (e.g. Android) do offer a
smoothed out output for certain sensors (e.g. gyroscope is calibrated
for drift), so it might make sense in that case to offer that as an
option on the `Gyroscope` object (e.g. `new Gyroscope({
correctForDrift: true });`), again defaulting to raw output.
Note that I'm more than happy to have this distinction re-discussed
and modified, but, again, this needs to be done at the GS level, so
all extension spec have the same understanding and apply this
similarly.
--
GitHub Notification of comment by tobie
Please view or discuss this issue at
https://github.com/w3c/accelerometer/issues/3#issuecomment-244353852
using your GitHub account
Received on Friday, 2 September 2016 11:46:52 UTC