We would like to do some testing with real users where we can find difficulties in using the discernatron software to grade search results. In addition we would like to get a nice bump in scored results that we can use when judging changes to search scoring.
To help this, setup a lunch at the WMF offices where we give people lunch and they pay us back by grading things in discernatron. Best day is probably a Thursday when Erik and Stas are in the office. We might want to look into if there is a good way to integrate remoties into this as well, although obviously we can't give them lunch. There might be some people outside WMF we could invite as well, this isn't an exclusive event. Justin has indicated interest, are there others?
Thinking we should start this with a 5-10 minute presentation demoing the software, then have people available to help guide people through the process. Maybe try and convince a few people to grade a result each day on their commute where possible.