Subj : Another suggestion....
To   : Time Warrior
From : Deuce
Date : Sun May 29 2005 07:27 pm

  Re: Another suggestion....
  By: Time Warrior to Deuce on Sun May 29 2005 16:41:00

 >  > Again, this is another issue that you flat out can't do with TCP.  It wou
 >  > possible to do something that almost works correctly most of the time... 
 >  > I'm not the kind of guy who likes solutions like that.
 > 
 >  > I can't sepak for DM, but I strongly suspect he feels the same way.
 > 
 > Thats why seeing as a BBS software CAN do that kind of tcp <=> baud
 > translation (CONNECT 300000 or whatever the default login says) and programs
 > such as com/ip and netmodem32 (open source now from what i've heard) that
 > have the ability to "lock baudrate" on a virtual com port -- I know BBS
 > Softwares are the most capable thing to handle some sort of @CHECKSPEED@ typ
 > code in an ansi.
 > 
 > If it were anything else but BBS Software, i'd be enclined to agree.

Actually, the "CONNECT 300000" or whatever is a fabrication to keep old
software happy.  The so-called "lock baudrate" is the same type of feature. 
There is absolutely no way using TCP to get a specific bandwidth.  As I said,
you could do something that almost works most of the time, but it's actually
flat out impossible to do it "correctly".

The issues are caused by TCP itself.  If you want a long technical explanation,
I could bore you with one, but it bascially boils down to the fact that (Like
persistant HTTP logins) it's not possible to do it correctly so I personally
feel it's better to not do it at all.

---
 þ Synchronet þ ``Penguins make tasty snacks''

.