Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!ames!sun-barr!apple!agate!eos!eugene From: eugene@eos.UUCP (Eugene Miya) Newsgroups: comp.cog-eng Subject: Re: Universal icons (Chernoff faces) Message-ID: <4991@eos.UUCP> Date: 4 Sep 89 23:48:32 GMT References: <343@aratar.UUCP> <4878@eos.UUCP> <5861@tekgvs.LABS.TEK.COM> <20045@datapg.MN.ORG> Reply-To: eugene@eos.UUCP (Eugene Miya) Organization: NASA Ames Research Center, Calif. Lines: 29 >Using faces as computer monitoring icons. Examples of such have appeared in previous SIGGRAPH slide sets. LANL made examples using faces and other icons (such as oil tank trucks for oil companies. I think it's the 1978 set. I thought about using a C-Cray computer shape (most other computers are just rectangular boxes, so you can't get much geometry from the multivariate data. I've not digitized a Cray-icon (it would have to be a 1, X or Y, 2s lack seats)). I thought about using them for performance measurement. There was an article in American Statistician I think (you can do a literature search, my copy is buried). Chernoff is or was at Stanford. The S stat package has a faces(1) function, see the S book. Chernoff tested faces up to 18 variables, S faces can go to 15 (so says the documentation). The gist of Chernoff's paper and my conversation with Rick Becker at ATT BTL about faces is that faces are able to show a greater dimensionality of multivariate data, but the interpretation is difficult because you have to associate the dimensions correctly to get the proper unambigous effect. They will likely remain a novelty without much practical use. Cute, but useless. Another gross generalization from --eugene miya, NASA Ames Research Center, eugene@aurora.arc.nasa.gov resident cynic at the Rock of Ages Home for Retired Hackers: "You trust the `reply' command with all those different mailers out there?" "If my mail does not reach you, please accept my apology." {ncar,decwrl,hplabs,uunet}!ames!eugene Live free or die.