[slavoingilizov]

I've seen this done multiple times too. It works until the human forgets about it, or knowledge is lost through collective organisational forgetfulness or churn. And the outages still happen because the lookup fails at some point.

[zie]

What we do to counteract that problem, is we generate organization wide calendars from this list. We provide .ics feeds for our organization, etc. ALL of it comes from this calendar, so if you want to know when a payday happens, you are just looking at a view of our calendar. If you want to know what the official org holidays are, you are looking at a view of our calendar. I mean it might display in Google calendar or Apple Calendar or whatever, but we are the data source.

So when a mistake happens, every employee that cares about that date will find someone to complain about it to, and ultimately someone somewhere goes, oh bother, we forgot to update the Calendar, and they promptly fix it.

Plus we avoid all blame, as it's very clear that we don't invent the calendar data. For us the HR dept is in charge of maintaining the calendar.

[neocritter]

Does it end up better or worse in practice than the pure computational approach? A system with the same faults as the system it replaces could still be an improvement if they occur less often.