On Jun 5, 2009, at 03:05, Jonas Sicking wrote:
But I will note that I also pointed out the need to gather data. It's
easy to have an opinion, but we won't know whose opinion is right
until we get some data.
@summary has been specified for over 10 years, so there should be
plenty of data out there to show if it has been a good idea or not.
Wouldn't you agree?
I think the best data collection suggestion so far was made by Philip
Taylor on IRC yesterday:
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-1058
Of course, it wouldn't help unless people agree on a cut-off point.
E.g. whether a feature is a failure if the revealed preference of 80%
of the sample of the constituency is to route around the feature? 50%,
90%, 99%?
It's also possible that the answer isn't a boolean listens to
summary / doesn't listen to summary but depends on the page in which
case it would be most interesting to find out how a user decides
before listening to the summary whether to listen for it.
In general, yesterday's IRC log from http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-701
onwards is relevant on this topic.
Especially:
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-757 through
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-764
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-807
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-1028
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-1046
http://krijnhoetmer.nl/irc-logs/whatwg/20090604#l-1062
--
Henri Sivonen
hsivo...@iki.fi
http://hsivonen.iki.fi/
