The incompelete fuzzy logic 101
by Dark_Skye

Problem: you need to kick-start a meeting tomarrow for your team. You mail everyone asking if 08:00 Est is a good time. You get back 20 odd replies that say either "yes" or "no". Well now does that help you? All you can logically conclude is 08:00 can not be a meeting time. Now you email again asking for 08:30, and still get a "no". By this method you'll only be able to exclude one time per email, that's an efficency of N. (Meaning we'll have to guess the time, or we'll be in an infinite loop of email tag.)

Sometimes true and false aren't enough, and that's when you need "fuzzy" answers like "maybe later", "ok, but I'd rather it be later", or "ok". (Think of these in the terms >, =>, and == respectively.) These fuzzy answers paired with a reply that gives the excluded range will yeild a much better efficency (log2 N) like a binary search would. (In other words, if they all mail back "08:00 should be 'maybe later", you can not only exclude 08:00. You can exclude at times before 08:00.) This helps us save a lot of time and effort.

Saving time and decreasing CPU and RAM useage by our program is important for the end user. You shouldn't make programs that cause a need to upgrade your system with each release. This wouldnn't be a good idea, since you'd lose customers, and in time market share. MicroSoft will faulter eventually if they keep going on the same track, because who will buy a new $2000+ system for win95 with a web broswer interface (win98)? Well our government friends will stop this I hope, by pushing back on the monopoly they've already lost a good bit of worth. (Asking price per share of MS is down.)

Lets say you wanted to make a water pump safety guage. Well, you do... trust me. You want the pump to cut off if the pressure is Very low or if it's Very high, but if the the pressure is OK let it keep running. This triangle graph shows if the pump should cut off from the current pressure. (0.0 = False, don't cut off. 1.0 = True, cut it off. The pressure vaules are dummy values with no units.) The value of OK = 75% vs the value of Very low = 25%. The output is "It's ok, but pressure is "lower than optimal".

fig. 1-1

With the addtion of Fact Tables fuzzy values allow for a vast improvement for event handling. I'll try to write a piece on Fact Tables later. I bet you feel there's more to this, well you're right...