Re: [exim] DKIM, Mailing-lists and signing lengths

Somewhat OFF TOPIC - but not 100%, as DKIM has functional predecessors here w/r
trying to vet the integrity of a communication...

===

Martin A. Brooks wrote:
> On Tue, June 15, 2010 19:55, W B Hacker wrote:
>> Best we could do to equate to an IRC's
>> nailed-up CLR's and a specific TTY's 'WRU'.
>
> Internet Relay Chat's....

International Record Carrier

- A government (and CCITT / ITU) recognized / licensed / regulated, carrier that
was required to be able to substantiate the end-to-end integrity of traffic it
carried. 'Back in the day' this involved knowing precisely over what
space-division (read hard-wired copper) pairs the signal had traveled and/or
keeping copies of 'torn tape' telegraph/teletype traffic for typically seven years.

> Common Language Runtime's....

Circuit Layout Records

Taken as a group, these were the evidence as to what pairs a private network
route traversed.

> Tele-Typewriter's....

got it in one..

> Western Reserve University....
>

Who Are You (half-vassed phoneticized to WRU).

A TTY had a programmable electromechanical device that recognized the 'WRU'
query and responded with it's pre-programmed identity.


> I will freely admit to TLA-fu fail, could someone please decode that
> sentence for me? :)
>

The earliest 'data' circuits were telegraph, heliograph, and semaphore - even
signal fires and smoke. Surprisingly effecting even in the bronze age.

Not just Morse code, but many others, and nearly always also often encoded and
or encrypted. A two-digit number might signify 'arrived safely' in 4WPM days
[1]. Baudot then ASCII TTY were the last major uses of these as they segued
into what we recognize as computer data communications today [2].

Telegraph, TWX and Telex largely moved over 'nailed-up' or dedicated circuits,
later switched, frequency-division, or time-division multiplex - all long before
TCP/IP or any other form of packet. All this was controlled by the IRC and
comparable national carriers - often the government itself (GPO and C&W for the
UK) - and no others.

So it was not hard with the above combination for Party A to haul Party B into a
court of law and - with the 'WRU' response and the aid of records subpoenaed
from the 'Record Carrier' prove that a given message had been received. And the
content unaltered.

Government, big business, shipping, and Banking were - and remain - the major
user groups in need of that 'feature' - largely because there isn't much point
in fighting it if one knows ahead of time that it is a slam-dunk to prove.

smtp came into being when packets were already the prime mover.

In a sense the 'IRC' function has been replaced with the need to obtain
certifiable logs of AT LEAST two MTA, and usually two tail-connection ISP as
well. Even those worthless w/o certification of routing at the point in time by
multiple backbone providers.

PTR RR and HELO with FQDN ++ WHOIS et al provide the replacement for more parts
of that 'attempt' to verify a communication.

All still very much harder to do with open-to-all-comers TCP/IP - and harder to
protect against spoofing or man-in-the-middle, even if/as/when using a more
inherently paranoid service than smtp (X.400 for example).

Note that your Bank, Credit-card issuer, Insurance company etc all want to
communicate primarily/only via an online 'secure message center' and will use
smtp - if at all - only to let you know they have received your input, or have
an answer waiting.

Needless to say - bank to bank EFT's don't travel via smtp. Nor necessarily even
TCP/IP.

All historical - but humankind tends to build on the familiar.

smtp didn't set out to be a secure or guaranteed communication any more than ham
radio did. Better than a message in a bottle - yes. But still 'best efforts' only.

The challenge is that the vast majority of users have found smtp good enough,
for long enough, and so fast, cheap, and cheerful that they tend to ignore all
that and expect more than was planned. Hell, most wouldn't even recognize the
term 'smtp'. They think it all 'just happens somehow'.

And 'we' (sysadmins) need to deliver to that expectation.

ELSE go fit a new bumper to the pickup truck.

.... which I am off to do now...

;-)


Regards,

Bill

(formerly at global cable address: 'WBHACKER' with an ASR-33)


[1] See 'Girdle Round the Earth' Hugh Barty-King nd the circa 1898 global
telegraph network(s). Denmark's Great Northern is the other major player.

[2] AT&T focused THEIR Unix on data manipulation - storage, search, recovery of
patents and the like. Over on the 'left coast' Berkeley Systems Development
focused 'their' Unix on data *movement* instead. Ergo we owe Fast File Systems
and IP to the left coast. In fact, we owe a very large part of it all to one
specific domestic couple - just two individials.