landman@hanami.eng.sun.com (Howard A. Landman x61391) (12/14/89)
I had a few hours to kill at the Sun User's Group meeting yesterday, so I wandered around the vendor booths. Imagine my surprise when I came across a booth demonstrating chemical design software! Here's a brief rundown - this does not imply my endorsement of this product in any way. I only had a few minutes to look at it. Running on a SparcStation 1 it seemed a little slow, and it was pretty clear that they were just going through a canned demo, not actually doing any computation. The "3D" display was not true 3D, just depth-cued. They had the choice of ball and stick, skeleton, or dot-cloud representation (and more) for atoms/bonds. Balls were shaded. The upper limit for graphics is about 1000 atoms per molecule.
nanotech@athos.rutgers.edu (Nanotechnology Newsgroup Nexus) (12/15/89)
[My software chewed this message up and when I reposted the second half it got lost. Here's the whole thing again. --JoSH] I had a few hours to kill at the Sun User's Group meeting yesterday, so I wandered around the vendor booths. Imagine my surprise when I came across a booth demonstrating chemical design software! Here's a brief rundown - this does not imply my endorsement of this product in any way. I only had a few minutes to look at it. Running on a SparcStation 1 it seemed a little slow, and it was pretty clear that they were just going through a canned demo, not actually doing any computation. The "3D" display was not true 3D, just depth-cued. They had the choice of ball and stick, skeleton, or dot-cloud representation (and more) for atoms/bonds. Balls were shaded. The upper limit for graphics is about 1000 atoms per molecule. >From the marketing blurbs, it seems that a portion of this runs on Fujitsu mainframes (similar to IBM 370 architecture and OS). Some parts of the code just prepare JCL (Job Control Language - gack!) to run programs on the mainframe and then submit the jobs. But at least some portion runs on Suns, and I got the impression they're trying to port everything. The Sun version may require a GXP graphics option to work. The reaction information and molecular information databases of ANCHOR are based on MACCS/REACCS, developed by Molecular Design Limited in the USA. ANCHOR includes molecular orbital calculation and energy estimation in addition to the molecule entry/display/editing. I asked about vibrational analysis; they said that was "under development". There are also hooks from ANCHOR into something called ADAPT which does pattern recognition and statistical analysis. >From a questionaire they handed out, I infer that they haven't figured out pricing yet, but that it could be anywhere in the $10K to $100K range. I don't think this includes a display station. At this stage they're more interested in beta sites than paying customers. For more information, the U.S. contact is: Fujitsu Limited 3055 Orchard Drive, B2-6 San Jose, CA 95134-2022 (408) 432-1300 ext. 5167 or in Japan try: Fujitsu Limited Manufacturing Sales Department Marunouchi Center Bldg., 6-1, Marunouchi 1-chome, Chiyoda-ku, Tokyo 100, Japan (03) 216-3211 Howard A. Landman landman%hanami@eng.sun.com