Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ames!ncar!boulder!sunybcs!rutgers!att!whuts!spf From: spf@whuts.UUCP (Steve Frysinger of Blue Feather Farm) Newsgroups: comp.graphics Subject: Re: Re: Scientific visualization Message-ID: <4518@whuts.UUCP> Date: 7 Jul 88 14:02:50 GMT References: <5352@ecsvax.uncecs.edu> Organization: AT&T Bell Laboratories Lines: 54 > I did not intend to raise a sore point with Steve Frysinger et al > when I predicted "hype" regarding audo data representation. I > was stating that this mode of representation that has been used > in scanning massive databases for trends, etc, has demonstrated > limited usefulness in the interpretation of more 'routine' databases. Actually, most of the successes have been with 'routine' databases. One of our favorites was the 7-dimensional set of economic indicators (taken from 1948-1980). These are monthly data, so we're talking about a time-series of ~384 7-vectors. The advantage of novel representations (not just auditory) comes in the ability of the analyst to apprehend patterns involving all of the dimensions, without resorting to "composite" (weighted sum) indices which throw away information and rely on the validity of some economic model. By the way, my ex-research partner in this is now a VP with Soloman Brothers as a direct result of this, so I guess it works (but I'm not an economist). Our chemist friends have had great success with spectral data as well, though most of their interest is for the visually impaired. > My point was that the data/music representation as portrayed > in "Dirk Benedict's Holistic Detective Agency" by Douglas Adams > IS a "hype" adaptation of that technology, and would succeed as a > standard output device for non-visually impared people only through > a vigorous marketing system. I haven't read the book (but I think I will!). You might also want to read "Hunt For Red October", by Tom Clancy - his character "Jonesy" is doing sonar the old-fashioned way, still very relevent today. > My apologies to those hard-working > individuals that are doing serious work in the field. FLAME OFF, OK? > Charles Lord No offense was taken; I wasn't pissed off, just protecting some territory (I only recently evolved :-). Incidently, there are lots of examples of novel data representations that might sound ridiculus (usually because they're easy to do), but work quite well. For example, consider the astronomer's trick of flickering back and forth between two sky photographs to cause the relatively few bodies in motion to "jump out" from the ground of many fixed bodies. This may be simple, but it also happens to be quite effective (and if you read the vision literature, you'll recognize the relationship to "abrupt onsets", the detection of which has a strong ecological foundation). Steve "Why not use ALL of our senses?" Frysinger P.S.: We did get some unsolicited media coverage which might have been interpreted as "hype". The Toronto Globe and Mail did a science section of the work first. Science News then picked it up and did a cover interview. And finally, last year, WNPR's "Sounds of Science" program interviewed me. While these were not "scholarly" presentations, I think they served a purpose in making people aware of the possibilities.