Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site bonnie.UUCP
Path: utzoo!watmath!clyde!bonnie!tk
From: tk@bonnie.UUCP (Tom Kirk)
Newsgroups: net.lang.lisp
Subject: flavors vs. flavors
Message-ID: <718@bonnie.UUCP>
Date: Sat, 15-Feb-86 19:18:52 EST
Article-I.D.: bonnie.718
Posted: Sat Feb 15 19:18:52 1986
Date-Received: Sun, 16-Feb-86 07:38:01 EST
Reply-To: tk@bonnie.UUCP (Tom Kirk)
Organization: AT&T Bell Laboratories, Whippany NJ
Lines: 28

Can anyone comment on experience with using the Univ. or Maryland
implementation of flavors in their variation of Franz Opus 38.91
versus the MIT flavors implementation distributed with Opus 38.92?
I have a choice of which to use, so I'm interested in finding support
for favoring one over the other. Specific considerations include:

  1) Efficiency. I'm aware that the UofM flavors has support in the
     franz kernel for method selection. Is the overhead of method
     selection large enough (in general) that this an important
     improvement? 

  2) Implementation. MIT flavors are implemented using closures;
     Maryland's flavors were implemented using hunks. So what?
     Am I missing anything?

  3) Generality. the MIT implementation seem to be a little more
     general, though I expect I wouldn't miss many of its features if
     they weren't there. On the other hand...

Any experience with either (or both) of these implementations that you
can share would be appreciated. I'm working with both before I decide
on one or the other, so if you have any information/opinions that
would help me with the decision I'd be glad to hear them. Thanks.

--------------------
Tom Kirk
ihnp4!{moss|bonnie|clyde}!tk
AT&T-BL (201)386-6044