ccsd(t) vs. qcisd(t)

From: <IRIKURA' at \`ENH.NIST.GOV>
Subject: ccsd(t) vs. qcisd(t)
Date: Fri, 04 Mar 1994 11:13:54 -0400 (EDT)

 Is there any difference in chemical performance between the CCSD(T) and
 QCISD(T) methods?  All I've found are a few papers stating that ccsd is
 "better" than qcisd, but the criteria used for judgment had little to
 do with
 real chemical problems.
 QCI appears to lack (at least for now) a diagnostic analogous to Lee's T1 for
 CC. Unfortunately, Gaussian-92 doesn't have T1 implemented.
 Is there a publicly/commercially available program that will do CCSD(T)
 calculations and with the following features (in order of decreasing
 importance)?
 	1.  T1 diagnostic
 	2.  open-shell and closed-shell
 	3.  faster than gaussian92
 	4.  handles effective core potentials
 	5.  direct methods
 	6.  spin-orbit coupling (e.g., using effective potentials)
 	7.  analytic derivatives
 	8.  works on cray, convex, or RS/6000
 	9.  easy to use
 Please post any answers.
 Thank you!
 Karl
 irikura' at \`enh.nist.gov