Hi Aza, Thanks for the comments! On Fri, Jul 4, 2008 at 1:28 AM, Aza <aza@mozilla.com> wrote: > Hi Andrei, > > Excellent work putting together this proposal. > > My first concern is the synchronous nature of lastPosition. This is one of > those cases where API design directly influences the interface. E.g., what > happens when a website requests geolocation via lastPosition data but does > not yet have permission? Because this is synchronous, it must block as it > waits for the permision UI to return. From the Firefox UI standpoint, a > blocking/modal UI isn't acceptable -- it disrupts the users train of > thought, forces them to make a decision they may not care about, and doesn't > work as a security feature since users will habituate to slimply hitting > "OK"[1]. > Sure, that's a very good point. > The solution, I think, is using an asynchronous method for getting the > lastPosition. That way, waiting for the secuirty UI is exactly like waiting > for a slow location service. I propose something like: getLastPosition(), > which takes exactly the same arguments as getCurrentPosition (so either a > PositionCallback or a PositionOptions). The other method is to add a > "lastPosition:true" to the PositionOptions and pass that into > getCurrentPosition. > I think there is a third option here: navigator.geolocation.lastPosition could simply be null until the user has granted the appropriate permissions. What do you think? Many thanks, AndreiReceived on Monday, 7 July 2008 20:03:34 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 11 August 2009 20:54:07 GMT