--On Saturday, 11 December, 1999 00:01 -0800 "John W.
Noerenberg" <[EMAIL PROTECTED]> wrote:
> At 5:39 PM -0800 12/10/99, Ross Finlayson wrote:
> >
> >So I wonder, what do people think is a reasonable suggested
> maximum email >message size these days (for messages sent over
> the public Internet, not >just within a private intranet)?
>
> I'm not sure you can express this in absolute terms. If a
> message is so large delivery of it impedes the ability of you
> neighbors to use the net, it's too large. But I have no idea
> how to apply a criteria like that in a realistic way. Once a
> message has been handed to the first smtp-receiver,
>...
Ross,
The only thing that might need to be added to John's answer is
that you should be aware, if you are not, that a protocol
extension was added some years ago that permits a receiving SMTP
to specify the maximum size message it is willing to receive,
thereby eliminating the historically-nasty problem of having to
wait for transmission of a very long message before rejecting it
(see RFC1653). As John's note implies, this, like everything
else, works more efficiently in the "no relay" case, or the case
where the only relay is a gateway or firewall that, itself,
enforces the size limitations, than it does with long relay
chains.
Especially at this time of year, many sites try to equip their
receiving MTAs, and, sometimes as a matter of courtesy to
others, their sending ones, with content-specific filters to
prevent large mailings of messages that contain large image,
audio, or video attachments with holiday greetings. I
personally would not encourage the practice --having automated
intermediate systems scan messages for content to heurstically
determine what should be rejected has always impressed me as
dangerous-- but it is certainly understandable in some
circumstances and does have the advantage of trying to focus on
one of the real issues for many sites, which is not transmission
size per se.
john
