This is BACKUPMXsf, a simple store-and-forward MTA written in C. It is implemented as two modules, which communicate via files in a spool area. The interface between the two is simply a list of SMTP commands to execute, rather than some arbitrary format control file. The "store" program accepts all mail offered to it and saves the mail in its spool diretory. The "forward" program sends all the mail in the directory to a remote MTA specified on its command line. "Forward" has no logic for retries or error recovery; if there is an error in a batch, it exits, and relies on the invoking layer to cause it to re-execute. Typically the invoking layer would be a shell script. When mail is forwarded, as many items are sent in one batch as can be; each is separated by a RSET followed by a HELO as received at the "store" program. The stored email data has one "Received:" line added to it. This is *very simple* code, while attempting to cover a functional implementation of SMTP which should be RFC compliant and work successfully with any other RFC-compatible servers. Note that the meaning of compatible allows for rejecting extensions to SMTP such as EHLO. The code does not do *any* checks such as for third-party relaying, for blacklisted IPs, or for spam or virus content. It is up to the destination MTA to do that, or to add it to the code yourself. The expected use of this code is as a lightweight store-and-forward MTA for use in a backup MX site. Note that this could be done with an off-the-shelf mailer such as sendmail or postfix, suitably configured. The main advantage of using this code is that the configuration is almost nil. Because the code is quite simple and easier to understand than, say, sendmail, it should be easy to modify for use in custom projects. For example as a spamtrap or honeypot or a greylist dummy mailer or to waste the time of blacklisted sites. It is expected that customised applications like this will probably use only the "store" program, but not the "forward" code. The code has been written to avoid buffer overflow exploits but does make the assumption that it is running in a controlled environment, i.e. not on a multi-user system with hostile users. It should be reasonably secure from outside attacks. It does not do any clever recovery from SMTP errors. The application for which it was written assumes that the MTA that mail will be forwarded to will either not be up, or will be up and accepting mail. This is an alpha release and you might expect to have to do some tweaking. Please feedback improvements to Graham Toal