Should superevent t_start/t_end be based on the trigger duration?
cc: @leo-singer, @erik-katsavounidis, @shaon.ghosh @peter-shawhan
From a discussion with @deep.chatterjee: !556 (comment 104776)
The current implementation of superevents' starts and ends come from https://git.ligo.org/emfollow/gwcelery/blob/master/gwcelery/tasks/superevents.py#L116. For all pipelines other than cWB and oLIB, these d_t
s are based on defaults in the config (https://git.ligo.org/emfollow/gwcelery/blob/master/gwcelery/conf/init.py#L55-79). This means that a high mass BBH trigger, lasting some 0.5 seconds, would have the same superevent duration as a BNS trigger which could last over 200 seconds.
Should we instead be using template durations for CBC events as reported in their extra_attributes
fields? Then, the superevent durations would accurately reflect the duration of the trigger.
Since this would represent a nontrivial change, it would be nice to have input from people.