DW> Well, I believe it is a well known fact... actually I believe most modem
DW> that supports V34/V34+ usually recommand 115,800 (or something close)...
DW> where under ideal condition it will push an amazing amount of data past
DW> the POTS... but hey, it can't hurt...
DW> Newer USR Courier Internal modem can support DTE up to 230 Kbps... too 
ad
DW> I can't crank my Courier External up like that... thou I CAN crank that
DW> Hayes serial card up to around 900Kbps... :) Then again, I don't think my
DW> modem will like it... :)
Yes, but the much higher DTE rate is only required when you are
using compression. I was transferring zipfiles. In fact for one
series of tests I turned compression off completely (&K0).
My arument is more to do with the DTE required to use LAPM
effectively at 31,200 bps, not V.42bis.
BTW, have you looked at the affect of compression on Internet
usage? With the Courier you can access the modem with a standard
comms program after an Internet session, issue ATI6, and see
how many characters and octets were received in the
current/previous session (not cleared until the next connect, I
believe).
I did a 4-hour session on the Internet yesterday at 28,800 bps
DCE. I set the DTE to 115,200 bps under Win 3.11 with Netscape
Navigator V3.01 on my 486/50 with 16 MB and a 16550A UART.
V.42bis specs were 2048/32. During the session I was careful not
to do any file downloads and spent most of the time exploring the
New Scientist web site (still need to spend more time there). The
data would be a mixture of non-compressible icons and other
graphic files and heaps of text. The total chars received was
2.25M while 1M of octets were received for a compression ratio of
2.25:1. As website go, www.newScientist.com, is probably more
text-based then most, so I believe, in most cases you would
achieve a lower compression ratio.
My standard DTE of 57,600 bps would probably have held this back
somewhat. Here, 115K was the shot. I doubt that 230K would have
improved things at all. What shocked me was that I'd had 1
overrun during the session. From memory I had the FIFO trigger
level set to 4. This is the first time I've tried 115K so I may
have to play around a bit more with the trigger setting. I don't
normally communicate under Windows (I'm OS/2 based) so this area
is a bit stange to me.
___
 X SLMR 2.1a X 
--- Maximus/2 3.01
---------------