Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!snorkelwacker.mit.edu!stanford.edu!eos!data.nas.nasa.gov!wilbur.nas.nasa.gov!eugene From: eugene@nas.nasa.gov (Eugene N. Miya) Newsgroups: comp.graphics.visualization Subject: Re: Out of the lab, into the classroom 2nd try Message-ID: <1991Apr3.215031.7763@nas.nasa.gov> Date: 3 Apr 91 21:50:31 GMT References: <1991Apr2.202227.13796@agate.berkeley.edu> <1991Apr3.172156.3960@nas.nasa.gov> <1991Apr3.193914.4569@agate.berkeley.edu> Sender: news@nas.nasa.gov Reply-To: eugene@wilbur.nas.nasa.gov (Eugene N. Miya) Organization: NAS Program, NASA Ames Research Center, Moffett Field, CA Lines: 122 In article <1991Apr3.193914.4569@agate.berkeley.edu> andyr@sag4.ssl.berkeley.edu (Andy Rose) writes: >I suppose a 'visualizer' needs a solid background in the sciences because >he/she needs to speak the language of science, for the 'visualizer' >in one role 'serves' the scientist and must understand what the guy >is getting at. Hopefully this >training is not exclusive of social science, because perhaps the data >sets from sociology (ie. census) is far more interesting and useful >than CFD. Actually, I don't personally think most scientists are going to afford "visualizers." The role is too specialized and will be restricted to companies and labs having $$ like Ames, like LBL, LLNL, IBM, etc. I think the average grunt grad student or prof is frequently lucky to have a personal computer (and I don't mean those doing simulation using computers in their day to day work). I mean the scientist who uses computers for word processing (daily) and maybe a weekly analysis run (if they have a shared computer, they might have to look their password up in a notebook). If I want to look at data, why wait? Why should a system be so complex I require an operator? I'm not certain what you mean by social science or by interest. You need the science not just for the language, but for the thinking. Modern science has many mathematical foundations independent of coordinate transform matrices. We held a meeting at Xerox PARC about what scientists do. Unfortunately a few of us got sick. Originally I invited A cosmologist A nuclear weapons designer A botanist A stable isotope bio-geo-chemist A acoustics physicist A computer performance measurer The common thread: with the exception of the designer we all skied together (the designer was once my Division chief here at Ames before returning to that type of work). All very interesting scientists. Right ring to left-wing. Men and women. Fun people to their colleagues. The cosmologist missed this meeting but we gave him a meeting at SGI (Feb 14, sorry, we took flak on that one) a newer version will be given on 25 April at the Dept. of Applied Sciences at LLNL. We (SIGGRAPH) are video taping this meeting {"Exploratory Computer Graphics"}. The stable isotope bio-geo-chemist has to travel aboard a ship in the Antarctica and take some data there. He can't afford another mouth to feed on his pitching ship (hazardous duty to some). The physicist doesn't use a computer everyday, she has a wind tunnel with a 120 ft by 80 ft cross section to study. She requires checking her notebook for a password; she has more important things to remember. The botanist is studying growth hormones and plants. He mixes chemicals and experiments take weeks to find out results. He was offered the lest help BTW (no slouch as a scientist). An awareness of these environments are essential. Some time ago Apple had a series of ads noting the teaching of science solely by computer. It had to do with a girl who refused to dissect a frog. A student isn't going to learn anatomy solely using a computer. "Being aware of language" is like thinking chemistry is just a bunch of symbols like H2S04. But the symbols don't tell you: are you supposed to mix acid into water or water into acid? Computers must fit into these research environments. Hazards and being squeemish are peripheral issues. >I wonder if the role of 'visualizer' will go the way of 'graphic artist' >who has in large part been supplanted by technology. With AVS and >apE and such scientists can do graphics on there own,thank you. I think so. Good analogy. However, I know scientists who don't like AVS. Asked what they want, they don't know. >!Neither. Yet. It is not a stand alone discipline by itself. >Why not? What does it take for a discipline to stand alone? Perhaps >if I call it 'visualization science' (remember physical science for >physics?, computer 'science') it can 'stand alone'. "Any field which has to call itself a science, isn't one." --Not my quote. "He who pays the piper calls the tune." If you want to make pretty pictures, then you are welcome to do computer graphics, entertainment, and maybe some mass education. $$ make a standalone field of research. If you want to do science, then graphics is a tool like a magnifying glass. But I believe we need to sharpen and refine our tools, especially if they are software tools. It's a means, not an end, and that point is made in the careers booklet which I cited. I am not trying to discourage you, I am just trying to find out where you are trying to emphasize. >!Perceptual psychology: has some use (human factors/engineering/cognition). >One basic problem which could be addressed by perception stuff is >how to map real values to colors. All kinds of questions arise when >you use color to represent values. Should I use intensity or hue, etc... >How about transparency, reflectivity, etc... As I just noted in a book review, color can be a distraction. The issue is more one of texture. Consider an Voyager image of a planet like Saturn or Uranus: very brown or very blue, no apparent turbulence from afar: change a color map for more dynamic range (simple remapping), structure stands out, but you must remember that Saturn is still brownish, not this new rainbow of colors. There are other real problems, more significant: information hiding, hidden or obscured objects and structures, optical illusions, etc. >Tactile feedback and stereoglasses can provide access to processing power >possessed by the brain which is untapped by conventional UIs. To a point. Physical models also have a place. >It may be that the only way to find structure and draw conclusions from >N-dimensional data is to build a machine to do it for us. A neural >network could process info with many more 'eyes' than we simple homonoids. >Indeed, this is far out. NN's are kind a fad going back to the 50s. Maybe see something in the future. >!Statistics: Yes, lots... >Here, here. A 'visualizer' should know the current thinking on reducing >large data sets. Every scientist is a little conservative when it comes to reducing data. He/she might never throw data away (intensionally). Also remember Huff. --eugene miya, NASA Ames Research Center, eugene@orville.nas.nasa.gov Resident Cynic, Rock of Ages Home for Retired Hackers {uunet,mailrus,other gateways}!ames!eugene