You said:
 PR> IMO this thread seems to point out how little most of us understand
 PR> about the virtually plug and play device called a modem.
The "virtually" can be a stretch sometimes.  :)
 PR> As I understand it there are 8 bits to each character.  If one's
 PR> thruput is running at about 1000 characters per second (cps), this
 PR> equates to 8000 bits per second (bps).  1500 cps equates to 12,000
You're right and wrong all at the same time.  
For our purposes, a single character of asynchronous serial data is 10 bits 
long - 1 start bit, 8 data bits, 1 stop bit.  (We'll ignore the various other 
permtations.)  Given that, a plain-jane 2400 bps modem can do a max of 240 
cps.  The same relationship is true of a 14.4k or 28.8k, when ignoring error 
correction and compression.
This is where it gets fun.  With MNP, v.42 and HST (which are all basically 
similar), the asynchronous data is converted to a synchronous data stream.  
As a by-product of this conversion, the start and stop bits are dropped.  
(This is the part where you were wrong and right.)  The net result is a 20% 
throughput benefit, so your 14400 modem with error correction at full tilt is 
theoretically capable of 1728 cps.  In real life this seldom happens, because 
neither phone lines nor software cps counters are perfect.  Add to that 
whatever error correction overhead is introduced by the transfer protocol and 
real life expectations drop a little bit more.
MNP-5 and v42bis are compression schemes that operate on top of MNP-4 and 
v.42 error correction respectively.  MNP-5 is capable of 4:1 on highly 
compressable files, but is actually detramental to .ZIP and other 
already-compressed files.  v.42bis is capable of 8:1 (if fully implemented) 
on highly compressable files and is ordinarily smart enough not to try to 
compress already comressed data.  The caveat there is that not all modem 
manufacturers make full use of the v.34 specification which allows for an 8k 
v.42bis dictionary.  Hayes modems do.  Most cheaper cookie-cutter modems only 
use 2k, which limits them to around 4:1.  
To answer your question about your connect speed, it is not unreasonable to 
expect to see around 1600 cps on a .ZIP download given average conditions.  
The fact that you are getting 1300 or so can point to a couple of things.  
Phone line conditions are foremost.  It could be that you connect ok but 
something causes the line to degrade and the modems step back to 12000 to 
keep the line alive.  While v.32 does provide for fallback and fall-forward, 
many modems don't do fall forward.  Meaning that they will slow down if 
conditions get bad, but won't speed up again if they improve.  It could be 
that the modems are staying at 14400 but having to resend packets in the 
error correction because of line hits.  It could be something as simple as a 
flaky cord between the modem and the wall jack.  If you're using the modem on 
a house extension, the fact that there are other phones bridging the line can 
affect the line impedance, which is very influential in line conditions.  How 
much phone wiring is in your house can affect it as well - long lengths of 
wire are capacitive and will shift impedance some.  Could also be crosstalk 
in the lines.  Try a new phone cord, try unplugging all the phones, try 
hooking a cord from the modem to the test jack on the customer interface.  If 
you see improvement, then you're on to something.  If not, it may well be you 
have old copper in your neighborhood.  If you're out in the boonies, you may 
well be on one of those rural multiplex carrier lines that are used to put 
more lines on  less wire.  
Those "plug and play" modems can be sort of a black art sometimes.
-dw
--- 
---------------