Path: utzoo!attcan!uunet!munnari!muller
From: muller@munnari.oz (Paul Muller)
Newsgroups: comp.dcom.modems
Subject: Re: TB+ jerky transmission
Summary: Why not have a dynamic 'low water mark'?
Message-ID: <2647@munnari.oz>
Date: 24 Jan 89 17:22:11 GMT
References: <2760@skivs.UUCP> <1470@neoucom.UUCP>
Lines: 13

Why not do a bit of quick math to work out what a suitable low water mark
should be, dynamically? That would allow you to do something like say editing
where you may hit one or two keys, then get a pageful of data (scroll page),
there you would like a low 'low water mark', or where you are typing a lot 
and little else, where a higher 'low water mark' could be used. I can't see
why it should be much harder to do than implement a suitable S register, it
would also save much netmoise over what an optimum setting for that particular
S register should be (just ask any MS-DOSer what the best buffer size is while
in a crowded room full of other MS-DOS junkies, get out of the way quickly!)
. The Trailblazer had a 68000 in it last I looked, must have some spare juice
to run a little math......
Paul
"If you can just tell me what your opinion looks like...." Me