Re: [magnetometer] Add RawMagnetometer, for uncalibrated readings.

>> The term raw is sort of weird as the data is partially calibrated.
> Agree, I also feel that 'Raw' term is confusing, unfortunately don't have proposal for a better name.

That might be a sign it shouldn't be a name. ;)

>> We'd need a term for this dynamic calibration...
> I think we should 'pull-up' terms as soon as there is more that 1 place where it is defined (applicable). If this is a generic term, absolutely. If it is only for magnetometer, maybe we can keep it in magnetometer spec, wdyt?

I see your argument. But because that's clearly generic terminology, I think it belongs in generic sensor. else, it implies editors of subsequent specs will have to hunt for terminology in all concrete sensor specs all the time.

>> As an option, we need something like keep HardIronBias or avoidHardIronCorrection
> This makes magnetometer interface calibrated and (almost)uncalibrated at the same time, introduces side-effects. If developer constructs magnetometer with 'avoidHardIronCorrection = true' and platform does not support uncalibrated magnetometer what would happen at start()?

Good question.

Maybe we can introduce a third kind of error for such cases, e.g. `NotSupportedError`?


-- 
GitHub Notification of comment by tobie
Please view or discuss this issue at https://github.com/w3c/magnetometer/pull/21#issuecomment-292889658 using your GitHub account

Received on Monday, 10 April 2017 08:57:18 UTC