RE: Position Heading Questions?

This thread has probably worked its way to completion, but I thought I would weigh in nevertheless. I agree that it would be much better to think of 3-axis orientation as different from position, the focus of this particular group. A compass may give you one of the orientation axes, while accelerometers may give the other two. A great scenario to think about is to imagine standing at the stern of a ferry. You're heading west but facing east. A navigation application would be interested in the heading and speed in order to update trip statistics, while a tourist application would be interesting in the orientation (to help you identify what you're looking at).

Speed and heading are derivatives of position. But they are so useful for most developer scenarios that it makes sense to keep them in the geolocation spec. However, orientation should be clearly separated from heading because that is not a derivative of position and there are many scenarios where you need only device orientation and not position (e.g. orienting or controlling objects on screen based on how the device is held). Like the ferry scenario there are many where you need both orientation and position together, but from a developer point of view it is not unnatural to fetch both types of information from different navigator.* interfaces.

If folks are passionate about this we should consider starting a spatial orientation working group so that some time in the future we can script to navigator.orientation just like navigator.geolocation.


-----Original Message-----
From: [] On Behalf Of Andrei Popescu
Sent: Monday, November 17, 2008 11:50 AM
To: Thomson, Martin
Cc: Greg Bolsinga; public-geolocation
Subject: Re: Position Heading Questions?

On Mon, Nov 17, 2008 at 5:09 PM, Thomson, Martin
<> wrote:
> Fair point.
> I think that will continue to disagree on the "very different" point.  Conceptually, I see that orientation is another aspect of position.
> Let us instead look at the alternative: a separate "physical-orientation" working group with a new draft.

Or maybe have the new draft done by this WG? It would seem the nicest
option to me.

>  Our would you instead recommend that it be punted to "v2".  My view is that the cost/benefit of this feature is demonstrably favourable.

I tend to agree. If more people want it in this spec, v2 could work as well..


Received on Tuesday, 18 November 2008 08:57:09 UTC