From: Evan S. <ev...@dr...> - 2004-09-02 20:13:16
|
On Sep 2, 2004, at 8:58 AM, Richard Bullington-McGuire wrote: > 1. GAIM learns that a Jabber buddy has gone away > 2. GAIM queries Jabber for the buddy's idle time > 3. GAIM sets an internal alarm that updates the idle time each minute > without going across the network to ask again > 3. When GAIM learns that the buddy is Available again, it cancels the > timer. > Better than that, why not: 1. Gaim learns a Jabber buddy has gone away 2. Gaim queries Jabber for the buddy's idle time, and stores the time the buddy went idle 3. If the user requests an idle time display in whatever form, determine the time from the stored time until -now-. This has the advantage of not needing a periodic poll for every away buddy. Of course, it's still not an accurate measure - one can be away without being idle, or idle without being away, or go away and become idle and then become unidle without coming back from away, or.... so I think it'd be a bit misleading. -Evan |