[Mailman-Developers] Re: [Mailman-Users] Allowing users to join without specifying pas swords

Chuq Von Rospach chuqui@plaidworks.com
Sun, 17 Jun 2001 08:29:50 -0700

On Sunday, June 17, 2001, at 06:21 AM, alex wetmore wrote:

> Please don't make this assumption.  It is true for the commonly used
> Unix MTAs, but it is not true for all MTAs.

you're misreading what I was doing here -- I'm looking at this based on 
how it goes over the wire, not how it's delivered to the MTA. So it 
doesn't matter if your MTA repackages it going over the wire or not -- 
I'm assuming it doesn't for simplicity, but if it does, you simply have 
to change the values to take that into consideration (and, for what it's 
worth, I did point out that postfix CAN do this, although I didn't 
clearly tie that back to this, because I thought the message was too 
long and opaque already...)

> Your method for figuring out bandwidth usage is interesting, and I
> think I'll do something similar for the recipient base and message
> sizes on my system.

It's just a rough attempt to get in the ballpark, but I think the 
numbers are going to be fairly good for the general SMTP protocol.

> 10k is much larger than my average message size,

true. It was convenient (especially for me, since I needed that data for 
my big emarketing machine anyway, and our messages are 10-14K and 
35-45K). If you go to smaller messages, the advantages of the 
buddying-up drops (for a 1K message, instead of  N * 11K + M * .1K, it's 
N*2K + M * .1K) and the protocol overhead becomes more important. For 
larger messages, the advantage grows completely. For 1-2K messages, you 
might see the advantage drop to 30% or less, I haven't done the math.

This also ignores the MTA's ability to cache connections, by the way. 
But that's really a random process and impossible to model this way.

> I would argue that it should default to OFF

I'm not surprised. It's Barry's call, but I think the customized URL is 
useful enough we want people to use it unless they have to turn it off, 
we don't want to have to try to convince the people who install stuff 
and leave everything defaulted to turn it on.

Chuq Von Rospach, Internet Gnome <http://www.chuqui.com>
[<chuqui@plaidworks.com> = <me@chuqui.com> = <chuq@apple.com>]
Yes, yes, I've finally finished my home page. Lucky you.

It's not the pace of life that concerns me, it's the sudden stop at the