2400 Baud Modems with MNP "Protocol"

[Erik E. &]

In article <2306 at brl-tgr.ARPA> fischer at RAND-UNIX.ARPA writes:
>
>With uucp, the MNP flow control will be incompatible, and thus one will
>have to disable MNP.  
>
>With Kermit, MNP is likely to play havoc particularly where the end-to-end
>flow control needs to be preserved (likely at 2400 baud on systems which
>might become busy), because MNP only appears to support modem to computer
>flow control.
>
>For interactive computer access, if you need control-s or control-q,
>e.g., if you use an editor like emacs ever, then again you might have
>difficulties.

My understanding of MNP was that it was a completely transparent error
correction protocol. How it accomplishes this is not my concern (as a
programmer) because the bytes get there. If my understanding is correct,
it is merely redundant if the protocol being spoken between the two
computers with MNP modems is error-corrected.

If I read you right, you're saying that with MNP, the modem will show
me the character it got in error, AND the correct representation, and
possibly some modem-modem protocol besides. In other words, MNP is NOT
transparent. Is that what you meant to say?  I have a very hard time
believing that, particularly in light of the rave reviews it has gotten
around here from some of the computer center staff.

Can someone post a description of the MNP protocol with particular
attention to the transparency issue raised here?

	Erik E. Fair	ucbvax!fair	fair at ucbarpa.BERKELEY.EDU