Recognition dates are misleading

[geofft]

Several groups have recognition dates of 17 November 2003, which is when some information was imported into the previous incarnation of the database and not their actual recognition date. Quentin suggests that we simply replace those with "unknown".

[geofft]

165 groups were allegedly recognized on 17 November 2003; another 9 were on 20 November 2003, and those are also groups that I'm pretty sure all were in existence before then (they include MITSFS, SIPB, and Tech Squares). No other groups were recognized that month, and the December ones look legitimate, so I'm inclined to just say that any November 2003 recognition date should become "Unknown".

All other dates where we recognized more than 5 groups on a single day are from 2005 or later.

[geofft]

I think the right way to handle this is the commit in the trac-64 branch of ​git://geofft.scripts.mit.edu/asa-db.git, including running groups/remove_bad_recognition_dates.py, but I'd like code review.

[adehnert]

NAK. Very much NAK.

My preference would be to do this at the display layer anyway, because I hate throwing away data. (Probably display "recognition date" as "entered into DB on" or something, I guess...) I can deal with this eventually.

[adehnert]

Hmmm... I feel like I'd be slightly happier with this as a migration.