Re: Location Determination vs. Location Conveyance

Marc,

The Firefox extensions that do HELD are another example of how the UA 
can do location determination (network-based, in this case) for itself. 
  Porting over Karl Heinz's mods to Zap! [1], it could probably do DHCP 
location as well.

--Richard

[1] <http://ecrit.labs.nic.at/>



Marc Linsner wrote:
> Andrei,
> 
> Thanks for the reply!
> 
> So, in devices where the OS doesn't manage/support location information, are
> you suggesting this API will access (for example) the GPS directly via a
> comm port and NMEA protocol?  Will this API have access to SUPL data in a
> mobile device?
> 
> Thanks,
> 
> -Marc-
> 
> 
> On 1/28/09 8:23 AM, "Andrei Popescu" <andreip@google.com> wrote:
> 
>> Hi Marc,
>>
>> On Tue, Jan 27, 2009 at 9:22 PM, Marc Linsner <mlinsner@cisco.com> wrote:
>>> Iıve been somewhat confused by some threads of late that have talked about
>>> the decision around how to make use of multiple location providers.
>>>  Admittedly, Iım not a coder, which Iım assuming leads to the majority of my
>>> confusion.  But Iım really wondering how this API is going to deal with the
>>> multitude of hosts, OSs, location determination mechanisms, etc.  Iıve long
>>> thought the host OS would provide location to applications via an API and
>>> would be up to the host OS to interface with the various location providers.
>>>  Granted, a host OS may indeed present multiple locations to an application,
>>> but in turn the host OS may make the decision as to which *one* location to
>>> present when multiples are available.
>>>
>>> In my mind, location determination will be 1 of 3 methods (or multiples):
>>> self-measured (GPS/Galileo), network provided, or user input.  It would be
>>> up to the host OS/user to decide which of them (or multiples) to present to
>>> applications.  I canıt imagine the API defined by this group talking
>>> directly to the location determination mechanism(s).
>>>
>>> What am I missing?
>>>
>> I don't think you're missing anything. If I understand you right,
>> you're saying that the API implementation may not even 'see' the
>> individual location providers, which are all hidden by the underlying
>> OS. And it's up to the OS to arbitrate between the various location
>> providers and return one location to the application (e.g. a UA that
>> implements Geolocation API in this case).
>>
>> While this is certainly true on some platforms, it isn't the case on
>> others so it is an implementation detail that a UA would have to deal
>> with on each OS it wants to support.
>>
>> Thanks,
>> Andrei
> 
> 
> 
>

Received on Wednesday, 28 January 2009 15:51:18 UTC