Which system for 3D graphics (FRODO/CHAIN/O)

Mark Israel misrael at csi.uottawa.ca
Sat Jan 30 01:06:00 EST 1993

In article <1993Jan26.120945.1 at ssrl01.slac.stanford.edu>, hudel at ssrl01.slac.stanford.edu writes:

> I am looking for comments on "which way to go" for 3d graphics.  We have the
> option of buying:
>		(1) an ESV 10 (moderate educ. discount)
>		(2) an SGI Elan (good educ. discount)

   I'd go with the SGI.  You have to look beyond the current machine you're buying.
One day you may have a family of machines, and you'll want them to be compatible.
Both brands are satisfactory (my ex-boss, Mike James at the University of Alberta,
really likes his SGIs; Dr. Derewenda down the hall really likes his E&S); but SGI
does seem to lead in performance/price ratio.

> We will be using mostly FRODO/CHAIN (density fitting)

   What's the current status of CHAIN?  Last I heard, its future was in doubt 
because of a dispute between Flo Quiocho and Jack Sack.  Does it have bones yet?

> with an occasional "O"-session (modelling).

   You could use "O" for fitting as well.  That's what people at the NRC here 
in Ottawa do (on SGIs).

> We would like to have full-screen stereo.

   No problem:  all the programs support that.  But Mike's people use 
side-by-side split stereo even on the machine that has full-screen stereo 
installed.  Does "O" have split stereo yet?  A couple of years ago it
didn't (which discouraged us from using it).

> (1) How do SGI FRODO/O compare with their ESV counterparts?

   Which Frodo are you using on the E&S?  Rice or CCP4?

   O and CHAIN were devised on the E&S and then imitated on the SGI, so
they'll be inferior on the SGI (but maybe only slightly).  They're certainly

   Arthur Chirino at CalTech and I have extensively hacked Christian Cambillau's
FRODO/TOM (for the SGI) to include some of the best features from O and CHAIN.  
Cambillau used to charge for TOM, but has now given us permission to distribute 
it free of charge; so it will shortly be available by anonymous ftp.  TOM is the 
only map-fitting program for the SGI for which you can currently get source code.
(The source code is a horrible mess, but if you're desperate, you *can* fix
any problem yourself, whereas with O, CHAIN, or TURBO you would have to wait
for someone else to fix it for you.)

   Then there's BronXtal, a program they're developing at the Albert Einstein
College of Medicine (again, only for the SGI).

> (2) Do you know of anyone porting FRODO/O to the DEC Alpha?

   O is written in PHIGS, so it should be portable to any modern graphics
machine.  But Alwyn Jones is keeping the source code under wraps, so if
he doesn't know about a port to the DEC, there isn't one.

misrael at csi.uottawa.ca			Mark Israel
Sunbeams brightly play, where Fancy's fair pavilion once is pight.

More information about the Xtal-log mailing list