This is awesome -- it's looking like somthing really impressive may come out of this.
I've been following the discussion, and I'm a great believer in emergant properties. I think that using physical proximity to decide what to cashe would be a mistake. I think that cashing based on regularity of use/use of similar tiles/time since last use etc would work better. For example, the 2013 games in canada mentioned, when that goes up, many people from all over the world are going to be viewing the data, irrispective of thier physical localities. You would not want those tiles to be dropped from a server in south africa simply because a tile of the middle of the sahara was closer. I'd say connect every user to a server that's close to them (in terms of network latancy) and let the statistics determin what gets cashed where. Though, if you need a tie breaker, by all means fall back location. And when you build the sytem to connect users to servers, be sure to take into account server load. JR 2009/11/2 Paul Ramsey <[email protected]>: <snip> > I'm quite enthusiastic about this, though trying to remain cognizant > of the fact that people in Argentina do also view Australia sometimes, > so that use case also has to have acceptable performance > characteristics. > > P. > > _______________________________________________ > talk mailing list > [email protected] > http://openaerialmap.org/mailman/listinfo/talk_openaerialmap.org > _______________________________________________ talk mailing list [email protected] http://openaerialmap.org/mailman/listinfo/talk_openaerialmap.org
