[AT] Re[2]: Delay in messages

Henry Miller hank at millerfarm.com
Thu Feb 2 19:10:22 PST 2006


On Wednesday 01 February 2006 20:21, Spencer Yost wrote:
> I almost didn't send this, but I now realize the now is the time to speak
> plainly.   I am sending this in reply to Dean's message but it is intended

I was hoping that if I ignored it, it would go away.   It isn't, so it is time 
for me to remininse about the bad old days of the net (I'm only 31 and I 
remember this)

Once upon a time some crazy computer scientists got the idea of connecting 
their computers together.   There was no internet (well there was, but only a 
few universities were allowed to connect, and they didn't allow toys like the 
machines that Unix ran on to connect to their network, so we will ignore it) 
then, but many companies did have offices scattered around the country, and 
their computers had to talk to the main office once in a while.  Generally 
this did this in the middle of the night when long distance rates where 
cheaper (but still expensive compared to today).  

During the day the phones lines the computers used were not used.   So the 
computer people in these companies (at first without the knowledge of the 
company) allowed their friends to connect.   Their friends could then send a 
message that would be stored until night when it was forwared to 
headquarters.   Headquarters would forward it to a branch office, which would 
forward it to a friend. 

Friends meant computer scientsits at other mid sized companies - computers 
were still expensive machines in large rooms with glass walls (so executives 
could look at their expensive investment).  Normal people could not dream of 
owning such a thing - a small computer could cost more than the average 
person would earn in a lifetime.

A typical message could take 12 days to arrive, but it was free.   AT&T was 
still a monopoly so long distance phone calls cost a lot of money.   A stamp 
was only about $.15 (or less), and arrived faster,but you couldn't use a 
computer to send normal mail.

Normal mail had one other advantage of these early days: it was more likely to 
arrive.   In those days a email wasn't sent to the final destination, you had 
to specify the entire path it would take.  Joe at acmecorp (no .com!) was a 
different person depending on if you were sending from CDC or DEC, as there 
were several different acmecorps around the country, some connected to CDC 
and some to DEC.  (In fact most connected to DEC, but this wasn't a problem 
so long as only one connected to any particular branch office of DEC.) 

So if you sent an email you first had to consult the latest map of the network 
to figure out how to get it there.  Executives often caught onto the cost and 
didn't allow their computers to send mail.   Then everyone had to find a 
different connection, and then re-create the maps and send them to everyone.

This loss was a big deal, so the programmers worked hard to figure out ways to 
prevent it.   However there was nothing they could do about the speed, so 
they didn't try.   Eventially the experience of their email was brought to 
the internet.   (internet mail generally is different, but USENET still 
operates from the same herritage)

So to all those whining about slow email: you expectations are wrong.   Email 
was designed to take 12 days to arrive.   There is nothing that prevents it 
from taking longer, on the electronic whim of some computer somewhere.  It is 
just luck that your email often arrrives faster. GET OVER IT!


P.S. I took some artistic liberty in the descriptions above, though the facts 
are close enough.   If you want a historical account see a historian.  



More information about the AT mailing list