On Thu, 2009-01-15 at 12:09 -0500, Benjamin M. Schwartz wrote:
> Simon McVittie wrote:
> > On Thu, 15 Jan 2009 at 09:35:54 -0700, Peter Saint-Andre wrote:
> >> Why do we need separate error/accuracy measures for horizontal and 
> >> vertical?
> > 
> > I'm not completely sure why Geoclue has them. Pierre-Luc, could you take
> > over here as geolocation liason? :-)
Geoclue has probably them because one of the provider provites such
information.  Vertical and horizontal error are only known when the
accuracy is set to DETAILED.

> I'm not Pierre-Luc, but GPS frequently has an order of magnitude more
> vertical error than horizontal error.  Trying to squeeze this situation
> into a single error number requires you to overstate one or understate the
> other.
This makes sens.

> Of course, there's no limit to perfectionism, and I'm not aware of any
> application that uses the error field at all.  The perils of design...
It all depends on the purpose of the application.  For IM clients, I am 
pretty sure that this field can be ignored as one could not really his
position to be too precise (for privacy reasons).  Although, this XEP could
technically be used to track vehicles or else... 

I think we should have both errors.  Then again, we might need to make explicit 
what
vertical-error: is it altitude error or latitude?

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
Telepathy mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/telepathy

Reply via email to