Thursday, March 28, 2013

Thursday: So that was a serious problem

But I think it's fixed now, or at least in any case, I think there's a workable solution to make it work better.  I think my solution also explains why it preferentially impacted one filter over the others.  I think this also explains the non-commutativity I saw yesterday: instead of taking the average or median of the valid values, we just take the first from the list.  Therefore, if the first entry is bad for some reason, we introduce an order dependence.

Look! A puppy!

No comments:

Post a Comment