Path: utzoo!attcan!uunet!cs.utexas.edu!usc!apple!bionet!ames!amelia!wilbur.nas.nasa.gov!eugene From: eugene@wilbur.nas.nasa.gov (Eugene N. Miya) Newsgroups: comp.graphics Subject: Re: SV phil. question (some collected responses) Message-ID: <6743@amelia.nas.nasa.gov> Date: 15 Jun 90 16:30:50 GMT References: <6719@amelia.nas.nasa.gov> Sender: news@amelia.nas.nasa.gov Reply-To: eugene@wilbur.nas.nasa.gov (Eugene N. Miya) Organization: NASA Ames Research Center, Moffett Field, CA Lines: 198 PLEASE NOTE in posting: DO NOT EXPOUND ON SV, the question had to do with scientific discovery. A few are doing such, PLEASE re-read the question. Skip, I liked your posting. Here are others. From: kassover@ra.crd.ge.com (David Kassover) In article <6719@amelia.nas.nasa.gov> you ask, essentially: Would Watson and Crick have made any significantly different discoveries (or more, or faster) if they had had at their disposal a graphics workstation to aid in visualizing molecular structure? My opinion: Yes, if and only if the advantages of that tool were not used up in the maintenance of that tool. Let me clarify. I am a consultant, and I and my company do much work at the above named Research Center. This entire letter, by the way, is *my* opinion, and has nothing whatever to do with the way my client runs his shop, as it were. I don't think it contains any proprietary information, nor "dirty laundry", or I would not have said anything. This Center was established a long time ago with a particular organization in mind: That of Ph.D. holders doing research, with lab technicians to handle the mechanical aspects (building experimental rigs, etc.) Things have changed in the last many years, to the extent that, since the doctors are no longer competent to maintain their tools, (we are a *long* way from being able to do research with a sliderule and a suitably sharpened pencil) a third class of employee must be included. The alternative is to "waste" a Ph.D. on computer system maintenance, or network maintenance, or document preparation. I have observed both solutions, and it is my opinion, therefore, that if Watson or Crick were to decide that playing with the computer were more fun, or more necessary, than observing models, much *less* research results in their specialties would be forthcoming. On the other hand, we have C and Unix because Bell Labs didn't mind their personnel doing large amounts of undirected, unchartered work. Which is more valuable? Which would be more to have first? Hmm. Would Watson and Crick have done more work with a workstation available? W&C accomplished quite a lot, in part, *because* they were well suited to their tools (clay, sticks, etc) Perhaps, with a different tool set, a substantially similar result would have been had from a different set of people. Is a Scott Joplin rag inferior to a Bach concerto? What would Bach have produced if his instrument were the square piano instead of the custom built church organ? > Should I offer prize money for the best answer? Um, the consensus at my entrepreneurship group is that "if someone offers you money, TAKE IT". So therefore, I will not discourage you from offering prizes. I hope this diatribe is somewhat what you wanted. If not, I'm sorry for wasting your time. [No, fine. --enm] -- David Kassover "Proper technique helps protect you against kassover@ra.crd.ge.com sharp weapons and dull judges." kassover@crd.ge.com F. Collins ----------- From: Ritchey Ruff Would Watson and Crick have learned more, faster using a workstation than their simple ball and stick models? Why? How? good question!!! [Well, not that good, I'm re-forming ala Skip.] I've been discussing exactly this type of thing with a friend of mine who is doing computational modeling of supercoiled DNA binding up at Univ of Washington. we came to the conclusion that 2-D screens don't cut it for the initial discovery phase ("ok, so *thats* basically what it should look like"), but does help lots to do selective verification and comparison of models (using data slices, binding strength plots, etc) IF (and a BIG IF) you have LOTS of cpu cycles AND the proper software AND a good ploting package (PLOT-10 just dont cut it ;-). What I think is really needed is a 3-D headphone setup w/data glove and moderate resolution (better than scott's setup). then you could get "inside" the model, directly manipulate it, have the computer check/verify against x-ray crystallography results, etc. we are talking MUCH cpu cycles here, of course... sum up: currently the hardware/software we use for "scientific visualation" is like the original internal compustion engines. big, cluncky, hard to use; BUT there were people back then who could see the possibilities and worked towards them (and the acuality actually went far beyond what was hoped/expected). [Good point] enough rambling, --ritchey ruffwork@cs.orst.edu ps-houdy! so, your .sig shrunk, eh? I'm still waiting for my prof to get around to proofreading my dissertation (he's had it since March 30th and still hasn't gotten done proofreading). I'm planning on getting involved with Tom Furness up at the Human Interface Technology Lab (I've been up talking to him and to William Bricken...there's an interesting guy). ------------ From: hmueller@wfsc4.tamu.edu (Hal Mueller) My hunch is that they would not have learned more, but might have done it faster given a suitable environment--one which enabled them to do 3-d manipulations intuitively. I'm excited by the possibilities of building connection rules into each component, specifying what other components it is allowed to attach to, and in what geometric configurations, and in the presence of what third-party components. Your comment about irreplaceable solid models is well taken. As an undergrad I spent a fair amount (too much!) time doing technical theater work, mostly sets and lighting. I had a chance to do a few set designs. The process was very collaborative--designer is concerned with visuals, lighting designer wants a result that won't make his job difficult, director wants physical space that will fit his plans for movement of the actors. Usual process was many iterations of rough sketches, drawn by one party with the others watching and kibbitzing. This worked well until I did a design for a blind director. Instead of pencil sketches we worked with foam rubber models, carved, glued, and moved around as we talked. After doing 1 design that way I never went back to sketches again, even when working with a sighted director. So perhaps this means that visualization tools aren't there yet. We need good depth perception, fast change of point of view, fast and easy assembly/dissassembly. I don't think visualization is a new approach to the problems its used on--simply a new tool. -- Hal Mueller hmueller@cssun.tamu.edu n270ca@tamunix (Bitnet) Graduate Student, Department of Computer Science Research Assistant, Department of Wildlife and Fisheries Sciences Texas A&M University, College Station, TX 77843 ---------- From: Jay Hersh I have seen other ball and stick models used for visualization for lack of access to graphics since the advent of CG. The real point is speed, data can be acquired and graphical models of 3D structures extracted from that data much, much faster via computers than via hand. If you really feel the need for touchy feely stuff you can always go to stereo lithography techniques to convert the 3D computer models to a ball and stick. As long as you don't care what the material is since stereo lith. uses funky plastics. This will dissapoint those who like to eat their models when done (like eating their young). You might try contacting Muriel Ross of NASA AMES (is that where you are) to ask here how 3D reconstruction from serial slices has helped her discover new physical relationships im cellular structures in rat macula (what helps you acheive balance) in order to develop a neural net model of how they actually work. I have done work (see my survey paper in Proceedings of 1st conf on Visualization in biomedical computing IEEE Comp. Soc. Press or other papers in same) for ways in which 3D graphics has really helped to solve structure extraction and functional inter-relation problems that no ball and stick could. Molecular topology is relatively simple so ball & stick seems easy. Try building models to visualize structure and inter-relation of things like cellular structures, you'll see computer graphics really, really does have a place despite what I agree is load of hype. My real affiliation is TASC an analytic engineering firm in Reading, Mass. I collaborate with Alex Pentland, thus my useof MIT computing resources. Hope this helps. Jay Hersh ======== I add one story: in "Land's Polaroid" a 3-D movie was shown to one of MGM (I think it was Sam Goldwyn). He was not impressed. Land later found out G had a glass eye. Blinn looks at many of his images with the color suppressed to simulate black and white. A significant part of the male community is color blind, especially so, many scientists. --e. nobuo miya, NASA Ames Research Center, eugene@orville.nas.nasa.gov {uunet,mailrus,other gateways}!ames!eugene