home *** CD-ROM | disk | FTP | other *** search
- X-Gateway-Source-Info: INTERNET
- Path: sparky!uunet!usc!elroy.jpl.nasa.gov!ames!network.ucsd.edu!mvb.saic.com!tgv.com!info-multinet
- Date: 30 JUL 92 15:06:49 GMT
- Newsgroups: vmsnet.networks.tcp-ip.multinet
- X-Return-path: <info-multinet-relay@TGV.COM>
- X-RFC822-From: Ned Freed <NED@INNOSOFT.COM>
- From: Ned Freed <NED@INNOSOFT.COM>
- Subject: Re: RFC 1339 support: Remote Mail Checking Protocol.
- X-VMS-To: IN%"skapur@ccmail.sunysb.edu"
- X-VMS-Cc: IN%"info-multinet@tgv.com"
- Organization: The INFO-MULTINET Community
- Message-ID: <238063CF30JUL92150649@TGV.COM>
- Nntp-Posting-Host: Mvb.Saic.Com
- Lines: 45
-
- > Most users are more interested in the presence or absence of new mail and for
- > them a simple check of the VMSMAIL profile should be good enough. The time
- > stamps are more or less optional and I do not see any way of implementing them
- > in VMS. The section "Server Implementation Notes" in RFC 1339 provides for a
- > way of not implementing time stamps.
-
- Sorry, in my opinion the notion of a simple check of the information in the
- profile file is a nonstarter. The idea behind this protocol is to see if any
- new mail messages have arrived _since you last checked in and read your mail_.
- The only information in the profile file is whether or not there are any
- messages you haven't read. This is a very different beast.
-
- > A client makes a request and sends the username. The server asks for the
- > password (unless the ipaddress/port/maildrop triple is stored) and checks
- > the SYSUAF for password and existence of account. If the account does
- > not exist, zeros are sent back. If the account exists, the VMSMAIL profile is
- > checked for new mail. If new mail exists or the number of records in the
- > user's MAIL.MAI is more than one, an appropriate response is send back for New
- > mail or Old Mail.
-
- My count of new mail messages hasn't been zero even once in the past five
- years. In my experience most users who get mail even intermittently have
- nonzero mail counts.
-
- Another problem is that certain kinds of delivery to VMS MAIL don't increment
- the new mail count. So it is possible for mail to arrive without the the
- profile file knowing about it.
-
- This protocol is very carefully specified to avoid use of unread message counts
- and instead uses time stamps to obtain its information. I cannot help but think
- that there are good reasons for this (I've mentioned a bunch of the ones that
- apply on VMS), and that an implementation that uses message counts to obtain
- its results is violating the intent of the protocol.
-
- > The current POP protocol has too much overhead which includes a process
- > creation and much more work on the part of the server.
-
- No argument. I think this protocol is a good idea. However, I think that
- without some meaningful support in both the delivery and user agents there is
- no way to return useful information from VMS systems. Please note that I don't
- think this is necessarily an impossible thing to do; it just requires some
- coordinated development work.
-
- Ned
-