eugene@pioneer.arpa (Eugene N. Miya) (06/24/88)
Sorry, I can't take it many more. I am getting tried of all the
marketing hype of "scientific visualization." Especially by marketing
people who have no idea what they are talking about. People just don't
wantonly display scientific data. It's not Enterainment Tonight.
Scientific data display isn't just adding pretty colors to data, nor is
it image processing, nor simulation, these may have parts of each of
these. But it has a different character from pure synthetic image
generation.
To reparaphase Hamming:
The purpose of computing is insight, not alternative realities.
I see some negative aspects of scientific fantasization bordering on
masturbation. Some good work is being extended beyond the useful.
What we have here, in some cases, are solutions looking for problems.
The problem is there are too few good people to guide the directions
data display should go. You don't just "look" at the data and get
insight. Thank goodness Cooley and Tukey had more sense than this.
Keep your marketing people at bay. It would be like having them say,
"XXX is a well anti-aliased image" when in reality XXX isn't. It takes
a graphicist to understand the issues of good image generation
(highlights, shading, refraction, etc.) so it also takes good people to
understand how scientists view and analyze data. I'm not trying to dull
people's enthusiasm for using computer graphics in science. It's just
that it's occasionally misdirected.
My comments are made independent of my employer, my ACM/SIGGRAPH
affiliation, and any other perceived association. They are mine.
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:
"Mailers?! HA!", "If my mail does not reach you, please accept my apology."
{uunet,hplabs,ncar,decwrl,allegra,tektronix}!ames!aurora!eugene
"Send mail, avoid follow-ups. If enough, I'll summarize."
More detail (specifics) if wanted.hacker@nucleus.UUCP (Thomas Hacker) (06/27/88)
In article <10763@ames.arc.nasa.gov> eugene@pioneer.UUCP (Eugene N. Miya) writes: >Sorry, I can't take it many more. I am getting tried of all the >marketing hype of "scientific visualization." Especially by marketing >people who have no idea what they are talking about. People just don't >wantonly display scientific data. It's not Enterainment Tonight. > I disagree with you. Scientific Visualization is a potentially important tool to be used in order to give form to literally mountains of numbers from simulation and experement. While it should not be used as the "end result" of the analysis of data, it can and should be used for giving a gross picture of what's going on in the system the scientist is looking at. These gross pictures are what give scientists insight into what's really going on and allows them to concentrate and direct their energies into looking at the right things. A good thing to read about what's going on in scientific visualization is in the May/June 1988 issue of "Computers in Physics" published by the American Institute of Physics. It contains several articles pertaining to the use of visualization using computers in Physics. -- Thomas J. Hacker ...!uunet!umix!nucleus!hacker (hacker@nucleus.UUCP) Physics/CS Undergrad Oakland University "Physics is the poetry of nature." Rochester, MI 48063
flip@pixar.UUCP (Flip Phillips) (06/29/88)
In article <10763@ames.arc.nasa.gov> eugene@pioneer.UUCP (Eugene N. Miya) writes: >Sorry, I can't take it many more. I am getting tried of all the >marketing hype of "scientific visualization." Especially by marketing >people who have no idea what they are talking about. People just don't >wantonly display scientific data. It's not Enterainment Tonight. [...] Nor can I. It seems that the marketing community has taken of on a buzzword and it will be a tough thing to stop. It seems now that any information represented graphically is 'Scientific Visualization'. In short, this is not true. [...] > You don't just "look" at the data and get insight. The point is presenting scads of data in a way that you CAN get insight. I.E. Reconstruction of a set of CT slices, mapping electron clouds, etc. Not necessarily performing false coloring of the CT slice, or presenting a plot describing characteristics of the electron cloud. The plot & false coloring IS useful... Its just that it doesnt seem to buy you as much as the other stuff. [...] > It takes >a graphicist to understand the issues of good image generation >(highlights, shading, refraction, etc.) so it also takes good people to >understand how scientists view and analyze data. [...] Well, actually it takes a scientist to understand how scientists view and analyze data usually. A mathmatician normally can't look at a rendered teapot and describe the associated equations. A graphics person can. At a bare minimum it requires scientists working with graphics people. Ohio State is heavily involved in this as are several other institutions (Ind. SuperComp Proj, etc) A supercomputer, mr/ct scanner, can spew out tons of data. Organizing this data in a way which is representative of the datas structure, content, or physical representation is a visualization. A plot of the numbers usually isnt directly, but like I said before, you use all of this information together. When a surgeon has a 3-d reconstruction of a broken hip he should and will refer back to the original CT slices. He also uses his eyes and hands when probing around in the actual patient. To do the operation blidfolded and only after looking at the 3-d reconstruction would be silly. (But fun to watch) Like all of the other tools, visualization tools are only part of the scientists tool kit, >--eugene miya, NASA Ames Research Center, eugene@aurora.arc.nasa.gov [...] -- Flip Phillips {sun | ucbvax}!pixar!flip Pixar - Marin County, California
trainor@lanai.cs.ucla.edu (Vulture of Light) (06/29/88)
eugene@pioneer.UUCP (Eugene N. Miya) writes: >Sorry, I can't take it many more. I am getting tried of all the >marketing hype of "scientific visualization." Especially by marketing >people who have no idea what they are talking about. People just don't >wantonly display scientific data. It's not Enterainment Tonight. hacker@nucleus.UUCP (Thomas Hacker, ACM) writes: > I disagree with you. Scientific Visualization > is a potentially important tool to be used in order to give > form to literally mountains of numbers from simulation and > experement. While it should not be used as the "end result" > of the analysis of data, it can and should be used for giving > a gross picture of what's going on in the system the scientist > is looking at. These gross pictures are what give scientists > insight into what's really going on and allows them to concentrate > and direct their energies into looking at the right things. I agree 100% with Eugene Miya. The hype will eventually hurt computer graphics. You only have to look at AI see what hype does to a field. I hope I don't puke in Atlanta this year... douglas [][] trainor@cs.ucla.edu [][] ...!{ihnp4,randvax,rutgers,sch-loki,ucbvax}!ucla-cs!trainor
lamaster@ames.arc.nasa.gov (Hugh LaMaster) (06/29/88)
In article <10763@ames.arc.nasa.gov> eugene@pioneer.UUCP (Eugene N. Miya) writes: >Sorry, I can't take it many more. I am getting tried of all the >marketing hype of "scientific visualization." Especially by marketing Well, in 1984 "Imagine" was a big marketing word, In 1986 "RISC" was in, In 1988 we have "Scientific Visualization", So, any bets on what the marketeers will have in store for us in 1990? (Salesmen have to eat too, you know.) -- Hugh LaMaster, m/s 233-9, UUCP ames!lamaster NASA Ames Research Center ARPA lamaster@ames.arc.nasa.gov Moffett Field, CA 94035 Phone: (415)694-6117
raveling@venera.isi.edu.UUCP (06/30/88)
In article <13973@shemp.CS.UCLA.EDU> trainor@lanai.UUCP (Vulture of Light) writes: >eugene@pioneer.UUCP (Eugene N. Miya) writes: > >Sorry, I can't take it many more. I am getting tried of all the > >marketing hype of "scientific visualization." ... > >hacker@nucleus.UUCP (Thomas Hacker, ACM) writes: > > I disagree with you. Scientific Visualization > > is a potentially important tool ... > >I agree 100% with Eugene Miya. The hype will eventually hurt ... I agree 100% with Thomas Hacker. I've just spent a LOT of time mucking with image fidelity measurements, wishing for a good 4-D graph, trying various alternatives, looking for tools and techniques that aren't obvious at all. And this is simple in comparison with many other problems begging for scientific visualization. There's a genuine need for visualization tools and techniques. --------------------- Paul Raveling Raveling@vaxa.isi.edu
jep@oink.UUCP (James E. Prior) (06/30/88)
In article <11026@ames.arc.nasa.gov> lamaster@ames.arc.nasa.gov.UUCP (Hugh LaMaster) writes: >In article <10763@ames.arc.nasa.gov> eugene@pioneer.UUCP (Eugene N. Miya) writes: >>Sorry, I can't take it many more. I am getting tried of all the >>marketing hype of "scientific visualization." Especially by marketing > >Well, in 1984 "Imagine" was a big marketing word, >In 1986 "RISC" was in, >In 1988 we have "Scientific Visualization", >So, any bets on what the marketeers will have in store for us in 1990? > >(Salesmen have to eat too, you know.) Why? -- Jim Prior {ihnp4|osu-cis}!n8emr!oink!jep jep@oink.UUCP Pointers are my friend.
eugene@pioneer.arpa (Eugene N. Miya) (07/02/88)
In article <5800@venera.isi.edu> raveling@vaxa.isi.edu (Paul Raveling) writes: > There's a genuine need for visualization tools and techniques. I don't disagree with the need. >>I said: >> >too much marketing hype on "scientific visualization." ... >> >>hacker@nucleus.UUCP (Thomas Hacker, ACM) writes: >> > I disagree with you. Scientific Visualization >> > is a potentially important tool ... >> >>I agree 100% with Eugene Miya. The hype will eventually hurt ... > > I agree 100% with Thomas Hacker. I've just spent a LOT > of time mucking with image fidelity measurements, wishing > for a good 4-D graph, trying various alternatives, looking > for tools and techniques that aren't obvious at all. > And this is simple in comparison with many other problems > begging for scientific visualization. No one disagrees with the use of the tool, the problem is the hype. People are creating tools in search of problems. I finished a 3rd re-reading of the CG Sci-Vi report (for use at work as well as the net). You said a key word "measurements." That's fine, but practically all tools to date are simply viewing tools, that's stupid (bit harsh, okay dumb). Anyway, give me a couple of more hours and I will polish up what's I'm composing on aurora. The problem is people who program or design systems who don't understand the process of scientific observation. I only learned some of this stuff in grad school years ago. Computer people make poor empirical scientists. BTW I know half of that committee, one's 2 doors away, Bob's up in the city, Blinn doesn't read net news, etc. If you don't see something in 24-hours, then next Wednesday after I get back from climbing in Oregon because I threw my hands up in disgust at work..... 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: "Mailers?! HA!", "If my mail does not reach you, please accept my apology." {uunet,hplabs,ncar,decwrl,allegra,tektronix}!ames!aurora!eugene "Send mail, avoid follow-ups. If enough, I'll summarize."
sewilco@datapg.MN.ORG (Scot E Wilcoxon) (07/02/88)
In article <271@oink.UUCP> jep@oink.UUCP (James E. Prior) writes: <In article <11026@ames.arc.nasa.gov> lamaster@ames.arc.nasa.gov.UUCP (Hugh LaMaster) writes: <>In article <10763@ames.arc.nasa.gov> eugene@pioneer.UUCP (Eugene N. Miya) writes: <>Well, in 1984 "Imagine" was a big marketing word, <>In 1986 "RISC" was in, <>In 1988 we have "Scientific Visualization", <>So, any bets on what the marketeers will have in store for us in 1990? <> <>(Salesmen have to eat too, you know.) < <Why? Because they are not AI (1987). -- Scot E. Wilcoxon sewilco@DataPg.MN.ORG {amdahl|hpda}!bungia!datapg!sewilco Data Progress UNIX masts & rigging +1 612-825-2607 uunet!datapg!sewilco
cjl@ecsvax.uncecs.edu (Charles Lord) (07/05/88)
In article <11026@ames.arc.nasa.gov>, lamaster@ames.arc.nasa.gov (Hugh LaMaster) writes: > In 1988 we have "Scientific Visualization", > So, any bets on what the marketeers will have in store for us in 1990? I'll bet Douglas Adams hit the nail on the head when he predicted musical data representation. It sounds like the IDEAL hype form of relation that means nothing to us now but could be learned if we listened to some huckster that was attempting to pan it off as the ultimate way to comprehend large data bases. With the growing trend for better sound generation a la Apple, it would be a simple programmimg/salesmanship progression. And you heard it here second, folks... -- Charles Lord Cary, NC cjl@ecsvax.UUCP Usenet cjl@ecsvax.BITNET Bitnet
spf@whuts.UUCP (Steve Frysinger of Blue Feather Farm) (07/05/88)
> In article <11026@ames.arc.nasa.gov>, lamaster@ames.arc.nasa.gov (Hugh LaMaster) writes: >> In 1988 we have "Scientific Visualization", >> So, any bets on what the marketeers will have in store for us in 1990? > > I'll bet Douglas Adams hit the nail on the head when he predicted musical > data representation. It sounds like the IDEAL hype form of relation that > means nothing to us now but could be learned if we listened to some > huckster that was attempting to pan it off as the ultimate way to > comprehend large data bases. Sorry, but you hit a sore spot here. You see, I've been doing research in Auditory Data Representation for about 8 years, as have a half a dozen or so friends/colleagues of mine. While I agree that hype is, in general, a Bad Thing, please don't vent your splein on our work - we don't hype it, and it happens to work VERY well for some particular types of exploratory data analysis problems (for both normally-sighted and visually-impaired analysts). It has been used quite effectively in seismic and economic analysis, and its superiority to conventional displays, for SOME types of data analysis, has been measured by carefully designed experiments. It just happens that people seem to do global pattern recognition better with their ears than with their eyes, while their eyes are better for scrutiny. Since I didn't notice you bashing the invention commonly known as the X-Y plot as 18th century "hype", please don't fire your flame gun when you don't know where it's pointing. Sorry, but as I said, you hit a sore spot, and I suspect you just fired a salvo in my direction inadvertantly. If you're interested, I can provide a bibliography of Auditory Data Representation, since I'm presently working on a review paper. Steve Frysinger
cjl@ecsvax.uncecs.edu (Charles Lord) (07/07/88)
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.
(If I am out of date on this, I apologize)
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. My apologies to those hard-working
individuals that are doing serious work in the field. FLAME OFF, OK?
--
Charles Lord
Cary, NC cjl@ecsvax.UUCP Usenet
cjl@ecsvax.BITNET Bitnetelf@dgp.toronto.edu (Eugene Fiume) (07/07/88)
In article <5352@ecsvax.uncecs.edu> cjl@ecsvax.uncecs.edu (Charles Lord) writes: >My point was that the data/music representation as portrayed >in "Dirk Benedict's Holistic Detective Agency" by Douglas Adams Sorry for the pedantry (or if there's some joke I didn't get) but that's "Dirk Gently's Holistic Detective Agency" by Douglas (DON'T PANIC) Adams. -- Eugene Fiume Dynamic Graphics Project University of Toronto elf@dgp.utoronto (BITNET); elf@dgp.toronto.edu (CSNET/UUCP)
spf@whuts.UUCP (Steve Frysinger of Blue Feather Farm) (07/07/88)
> 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.
rick@hanauma (Richard Ottolini) (07/31/88)
An important contention made at a recent Bay Area SIGGRAPH on `Ten Unsolved Problems in Computer Graphics' is the graphics goals of various disciplines are different. The entertainment industry desires realism; Art desires aesthetics; CAD desires accuracy. I contend that scientific visualization desires PRECISION. First, graphics makes a large set of numbers generated from data measurements or simulation more graspable than in raw form. Second, in my field-- earth observation and simulation-- the most interesting parts of a dataset are the anomalous regions. Extrapolating missing regions (e.g. splines) or smoothing glitches (e.g. median filters) can often distort or destroy essential features of an image if applied improperly. I warn against wholesale borrowing of Hollywood graphics techniques for scientific visualization as I see happening at many supercomputer visualization centers. Beauty is nice and often a result of scientific visualization, but should not be a major goal.
rl@cgl.ucsf.edu (Robert Langridge%CGL) (08/03/88)
In article <1201@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes: [...comments on scientific visualization and criticism of overemphasis on "realism"...] >Bob (Langridge) I know you read this, comments? OK Gene: This is the closing paragraph of an abstract I wrote for the upcoming Molecular Graphics Society 7th Annual Meeting (and the first to be held in the US - at the Cathedral Hill Hotel, San Francisco, 10-13 August 1988). I can post the rest of my abstract and more information on the meeting if anyone is interested. "...computer graphics is now an essential tool, but the rapid decrease in cost and increase in power of commercial workstations suggests emphasis on the *integration* of graphics with numeric and symbolic computing and with large databases. Molecular modelers should treat com- puter graphics as a means of interacting with their computations to generate insight and to aid in reasoning, and not be seduced by the heavy emphasis on "realism" in the computer graphics field. If real-time images can be generated which correspond to the investigators model of "reality" they should be used, but "reality" is not a necessary property of images for successful scientific visualization. Their most important property is the information they convey." I have spent a lot of time and effort over the last 25 years trying to persuade my colleagues in molecular biology that graphics is a useful tool. Now I risk being labeled a reactionary! Bob Langridge rl@cgl.ucsf.edu Computer Graphics Laboratory University of California San Francisco CA 94143-0446 +1 415 476-2630, -1540, -5128