Paul, at 13:27 on Jan 28 1996, you wrote to Bill Grimsley...

PE> 57600 will not give me anything that 38400 won't. 

BG> Of course it will - it'll allow you to transfer compressed data at full 
BG> speed with 33600 connects (I've already logged speeds in excess of 4000 cps 
BG> here).

PE> 119%?  I'd like to see that. 

There you go again.  Take off the blinkers, Paul.  Not only did I get 4000
cps with Russell Brooks' Courier V.34+, he just happened to be here to see
it when it happened.  Oh dear, an independant witness - sorry about that.
:)

PE> Did you do your stats on a large (500k) file to get proper stats?

Yep, 95SJONG2.ZIP from David, 1.35Mb in size.  Big enough for you?

PE> Also, 38400 should be less strain on my computer I have been told.

BG> That's crap Paul.  It's true on XTs perhaps, but not with the faster 486s, 
BG> which will happily MT with port speeds of even 115200.  Ask Rod.

PE> He agreed that it was more strain? 

Now you're just being intentionally silly.

PE> But can still cope anyway.  I agree, I doubt that it would put >1% extra 
PE> strain on my computer. The main reason was I didn't think there was any 
PE> point, even for 1%.

I've already told you that with your port locked at 38400, the fastest it
will allow data through is at 3840cps (in reality, it's a tad less than
that), so with a 33k6 connect, you immediately lose ~200 cps from your top
speed.  IOW, you'll NEVER see 4000 cps unless your port is locked at 57600
or higher.

Is that clear enough for you yet?

BG> Note also the delay in the command, which was put in for a reason (although 
BG> I forget exactly what that reason was now).  :)

PE> A bug in the USR Courier I would assume.

BG> No, it applied equally to all of my Sportsters.

PE> A generic USR bug then.  What a pity you didn't document it.

Nope, it applied to several other HS modems as well.  I think that Binkley
is sending data too fast for the modems to assimilate.  Even el-cheapo
comms apps have an adjustable modem-command delay (defaults to 10mS in
Qmodem).

Regards, Bill

--- Msgedsq/2 3.20