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

> 1. GeomagneticFieldSensor can be constructed from Magnetometer, not the other way around - so low vs higher level

The distinction between low and high is that low-level sensors are characterized by their implementation and high-level sensors after their readings.

> 2. Options do mostly go ignored... don't show up in auto completion etc

Well, again if these options are critical for key use cases, they probably shouldn't be an option.

> 3. GeomagneticFieldSensor is a specialized (but probably common) use case - ie finding the direction of magnetic north - hence is why hard iron compensation is needed

It seems that this is the common use case for that sensor, and this kind of compensation is what people expect when they talk about a Magnetometer. But I might be wrong here.

Similarly, when people talk about a barometer, they're thinking about air pressure. The fact that such a sensor actually includes a pressure sensor and a thermometer (necessary to calculate the pressure), is sort of a side issue. That sensor still _feels_ low-level despite it being the product of sensor fusion.

> 4. Sensors postfixed by *Sensor are what is commonly used for specific use-cases, so fits with the above. When using a [raw] Accelerometer, Gyroscope, you need to know what you are doing - same with Magnetometer. All of these are calibrated a bit, even Magnetometer (soft iron calibration) which is fine - you wouldn't want otherwise (I dont know the use case).

Yes. I agree that there's a sweet spot between some level of calibration (and even fusion). There's no simple recipe for finding that sweet-spot. It is essentially a combination of existing practice, coherence across the platform, common use cases, etc.

-- 
GitHub Notification of comment by tobie
Please view or discuss this issue at https://github.com/w3c/magnetometer/pull/21#issuecomment-300761787 using your GitHub account

Received on Thursday, 11 May 2017 11:29:01 UTC