Extremely high or low t_done values

[chmac]

We're seeing some outrageously high t_done values from BlackBerry devices. Here's a few examples:
t_done=1332824651142, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9900; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.585 Mobile Safari/534.11+"
t_done=1334395941323, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9860; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.585 Mobile Safari/534.11+"
t_done=1334395918226, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9860; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.585 Mobile Safari/534.11+"
t_done=1335025506071, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9900; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.585 Mobile Safari/534.11+"
t_done=1335036029722, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9810; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.585 Mobile Safari/534.11+"
t_done=1335036029867, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9810; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.585 Mobile Safari/534.11+"
t_done=1332729963010, user_agent="Mozilla/5.0 (BlackBerry; U; BlackBerry 9360; en-GB) AppleWebKit/534.11+ (KHTML, like Gecko) Version/7.0.0.530 Mobile Safari/534.11+"

Any idea what's going on with these values? Here's a few others from the other end of the spectrum, these all returned t_done=-536870912, some of the user agents are:
"Mozilla/5.0 (Linux; U; Android 2.1-update1; en-gb; SonyEricssonE15i Build/2.0.2.A.0.24) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Mobile Safari/530.17"
"Mozilla/5.0 (Linux; U; Android 2.1-update1; en-us; WX445 Build/VZW) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Mobile Safari/530.17"
"Mozilla/5.0 (Linux; U; Android 2.1-update1; en-gb; GT-I5700 Build/ECLAIR) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Mobile Safari/530.17"
"Mozilla/5.0 (Linux; U; Android 2.1; en-us; pandigitalopp12/sourceidDL00000013) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Mobile Safari/530.17"
"Mozilla/5.0 (Linux; U; Android 2.1-update1; en-us; T-Mobile_Espresso Build/ERE27) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Mobile Safari/530.17"

Those t_done times are over 40 years, so it seems unlikely our pages took that long to load!

Any advice or articles on how to filter and analyse boomerang data? I read the slides of a few presentations (maybe?) by boomerang authors. What does the "one number to rule them all" slide mean btw?

[bluesmoon]

:)

Chances are the start time reported by NavigationTiming was 0. We've moved development of boomerang to https://github.com/lognormal/boomerang/ and I know I've fixed something similar in that fork.