Talk:Microcom Networking Protocol

The analysis on this page is silly and wrong. It says for example that "when using MNP 3, a user can expect to get about 2600 bit/s throughput from a 2400 bit/s modem."

But this is wrong. It literally says that a user could get 2600 bits per second from a 2400 bps modem. In no way is the user transmitting 2600 bits per second. The user is not sending 2600 bits of data, and even when you add on the modem's overhead the user is not sending 2600 bits of data per second.

The conceptual difficulty encountered by the author of the main article is that a 2400 bps modem does not send 2400 bits of user data. At the usual 8n1 configuration, the modem is sending 1 start bit, 8 data bits, and 1 stop bit. Thus, a 2400 bps modem will result in 240 bytes per second of user data transferred. 240 bytes is 1920 bits of user data transferred per second.

I'm going to assume that the main article is correct when it states that MNP 3 results in 108% efficiency. Thus, a 2400 bps modem using MNP3 would send about 2073 bits of user data per second. However, and this is the key point, the 2400 bps modem is still sending 2400 bits per second.

What the author meant to say, "when using MNP 3, a user of a 2400 bps modem could expect to get throughput equivalent to a standard 2600 bps modem, if such a thing existed." 66.215.10.253 17:26, 14 April 2007 (UTC)


 * I think there should be some history behind this thing.  Not all modems from the year 2010 behave like modems from the year 1990.   Reasonable expectations tend to vary according to past experience projected into the future.   A modem "capable" of 2400 bps, rarely gets 2400 bps, as there are many servers in the middle, some of which respond more to the wishes of the owner of the server, than the wishes (and expectations) of an end receiver/initiator.    The main page of this article could be improved by citing the performance of specific modems in laboratory-perfect conditions.   This is how benchmarks are arrived at.  216.99.219.206 (talk) 20:41, 17 August 2010 (UTC)

Echoplex mention removed
I have removed the mention of "Echoplex". I am not sure why it was inserted where it was, for one thing, but more to the point, "echoplex" is generally another term for "duplex". The source in question appears to have confused this terminology. Maury Markowitz (talk) 12:56, 23 October 2013 (UTC)