In response to RFCs 347 and 348, ECHO and DISCARD, MIT-DMCG

sockets 7 and 9.

A few administrative type decisions were made, to which I

would like to solicit comments.

First, if our machine is too full to support another deamon

process, the pro-offered connection will be refused; that is, a

CLS will be sent in response to the RFC on socket 7 or 9.

Second, if no activity takes place for two consecutive minutes

(no characters received on input, or output blocked waiting for

allocation), the connection will be aborted.

Does anyone see the need or desirability for the third

service in this set: a random character generator? Many systems

provide programs like "TTYTST" available over the standard login

connection, but there would be no difficulty in creating a server

connection (like ECHO) offering just this limited service. I'm

sure that it would be useful for new NCPs, as well as statistics

gathering.

[ This RFC was put into machine readable form for entry ]

    [ into the online RFC archives by BBN Corp. under the   ]
    [ direction of Alex McKenzie.                      1/97 ]

[Page 1]

---