W3C home > Mailing lists > Public > public-geolocation@w3.org > October 2008

Re: PositionError Requests

From: Olli Pettay <Olli.Pettay@helsinki.fi>
Date: Thu, 30 Oct 2008 23:12:33 +0200
Message-ID: <490A2341.6040804@helsinki.fi>
To: Andrei Popescu <andreip@google.com>
CC: Doug Turner <doug.turner@gmail.com>, Chris Butler <cbutler@dash.net>, Greg Bolsinga <bolsinga@apple.com>, olli@pettay.fi, public-geolocation@w3.org

On 10/30/2008 10:58 PM, Andrei Popescu wrote:
> On Thu, Oct 30, 2008 at 6:16 AM, Doug Turner<doug.turner@gmail.com>  wrote:
>> On Oct 29, 2008, at 8:21 PM, Andrei Popescu wrote:
>>
>>> On Wed, Oct 29, 2008 at 4:14 PM, Doug Turner<doug.turner@gmail.com>
>>> wrote:
>>>> Looking at the defs., i do not see much of a difference between
>>>> PROVIDER_ERROR and NOT_FOUND (NOT_FOUND is just a special case of
>>>> PROVIDER_ERROR).
>>> Hmm, I think the difference between the two is more significant in the
>>> case of of a network-based provider: PROVIDER_ERROR would mean that
>>> the provider is not functioning correctly (the server is down, the
>>> client can't reach the network, etc), while NOT_FOUND would mean that
>>> everything is working fine, it's just that the server doesn't have
>>> enough data to tell where the client is. But I agree that maybe this
>>> is still not enough to warrant two separate error codes.
>>>
>>> Andrei
>>
>> That is what the message attribute is for, right?  From a web developer's
>> point of view PROVIDER_ERROR (something bad happened in the "thing" that
>> provides the geolocation position in the browser) and NOT_FOUND (the
>> position couldn't be determined) are the same.
>>
>
> Ok. So we have:
>
> UNKNOWN_ERROR = 1 (could be 0, but other error interfaces start at 1,
> e.g.  http://www.w3.org/html/wg/html5/#mediaerror)
> PERMISSION_DENIED  = 2
> POSITION_UNAVAILABLE = 3
> TIMEOUT = 4
Well, media error doesn't have "UNKNOWN_ERROR", nor could I find such
error elsewhere (DOM Core, DOM Range, DOM L&S ...).
Actually, it would be great if we could standardize 0 to mean always
unknown error, in every spec. (In a way it is already that, because it isn't
specified anywhere)

But if others think 1 is fine, that is ok to me too.


>
> Is that ok with everyone?
>
> Andrei
>
>
Received on Thursday, 30 October 2008 21:13:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 22 March 2012 18:13:40 GMT