HYPERCHEM, CACHE, GAUSSIAN92, CEREBELLA, ROCKET-SCIENCE

 Netters:  Re:HYPERCHEM
 It's a great package in my opinion.  It's not perfect in my view, but for the
 PC enviroment it does more than most others I am familiar with.  I would
 compare it against George Purvis' Mac CACHE package at Tektronix. Two very good
 packages!  I would like to see a bit more flexibility in some of HYPERCHEM's
 approach to some of its features.  I share the concern related to conjugate
 gradient optimisation for semi-empirical, and in cartesian coordinates.  When
 such an approach is taken for floppy molecules such as dyes, a lot of cpu time
 gets used!  Internal coordinates using a Newton-Raphson approach  is more
 powerful than anything we have used for the past 10 years. Baker's EF routine,
 particularly in MOPAC-93, is hard to beat in our experience on PC's and work
 stations.
 Now in good fun, I have to take a friendly poke about the recent comment about
 what IOP(4/20=7) means in Gaussian: 1) Most places have source. 2) All places
 SHOULD have documentation.  I realise things are getting to "If I have to
 read
 the documentation, the program's not well written,"  but you gotta do it!
 IOP(4/20=N) is described therein, though not explicitly for N=6,7,8. The
 documenation is a bit behind, as it usually reflects the text at the beginning
 of each block of Links for the source....so Rocket-Science is not required, just
 a little logic.
 John McKelvey
 Eastman Kodak