Path: utzoo!attcan!uunet!husc6!uwvax!vanvleck!uwmcsd1!ig!agate!ucbvax!UF.MSC.UMN.EDU!fin From: fin@UF.MSC.UMN.EDU (Craig Finseth) Newsgroups: comp.protocols.tcp-ip Subject: Remote Gnumacs servers? Message-ID: <8806231856.AA09224@uf.msc.umn.edu> Date: 23 Jun 88 18:56:39 GMT References: <335@csvaxa.UUCP> Sender: daemon@ucbvax.BERKELEY.EDU Organization: The Internet Lines: 25 I experimeted some time ago with a version of Gnu-Emacs that had been modified to use the LEAP features in an Encore Annex terminal concentrator. The LEAP features offload the terminal operations (echoing, etc.). Although not exactly the same situation as you describe, it did lead me to propose a conjecture (additional testing may turn it into a theorem): Moving processing from a central, shared resource to a distributed, private resource results in lowering the average response time while increasing the standard deviation of the response time. In this particular case, I felt that the increase in standard deviation resulted in crossing a threshold which made performance poor enough that it could not be offset by the decrease in average response. In particular, when you went from typing to editing, the pause was long (compared to echoing) and you had the impression that you editing character (e.g., ^N, ^A) was ignored, causing you to type it again. To say the least, this behavior is annoying (see previous paragraph.) My advice: quit while you're ahead. Craig A. Finseth fin@uc.msc.umn.edu [CAF13] Minnesota Supercomputer Center, Inc. (612) 624-3375