Re: Forward/backward compatibility

On Thu, 13 Nov 2008, Erik Wilde wrote:
> instead of accepting valid scenarios where people say that they want to 
> use non-lat/long positions and this is how things are working now, this 
> argument asks them to use lat/long because that's how the API works. but 
> aren't we in a discussion about what people have as use cases and 
> scenarios for that API and what it should support?

No, we're months past that discussion.

As I've said before, there are certainly use cases for getting a postal 
address for the user. I'm not personally convinced that they need an API 
(we seem to do fine today with form fields); and I'm definitely not 
convinced that this API, which is optimised for a frequently changing 
location, should have anything to do with it. But irrespective of that, it 
belongs in a separate spec, if it belongs anywhere at all -- either a v2 
of this spec, or another one altogether.

Instead of arguing over and over about whether Geolocation v1 should have 
this, the people who want addresses in an API should write a separate 
specification optimised for those use cases.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Thursday, 13 November 2008 23:48:20 UTC