On Thu, 20 Mar 2003 09:13:47 +1200,
mike(_dot_)pearson(_at_)ssc(_dot_)govt(_dot_)nz writes:
Thanks to those contributors who mentioned this issue. I found the following
1-page summary useful in understanding the background of the concept:
http://members.aol.com/trajcom/private/commons.htm
An important concept though, is that the scarce resource is not [bandwidth],
but [our attention] i.e. technology can give extra Mbs, but not extra
minutes.
Exactly!
At computer prices of $10,000 for 1ghz. [1]
At bandwidth of $100/gb. [2]
At storage of $50/gig. [3]
In an exaggerated situation, I'd rather burn 60ghz*seconds, 60kb of
bandwidth, and 120kb*100 weeks of storage than to deal with a *single*
message. In the first case, I'm out $.02 in resources, in the second,
I'm out 5 seconds. I value 5 seconds of my time as much more valuable
than $.02.
For an unexagerated situation, with bandwidth prices of $5/gig, and
assuming you actually nuke unwanted email weekly, instead of save it
for 2 years, the numbers come to that I'd rather burn 60ghz seconds,
2400kb of bandwidth, 12MB*weeks of storage, for one spam.
Thus, I've still thought that the mantra of 'stealing my bandwidth' is
an argument I cannot understand, because of how weak it is. I have no
idea who made it, or why they focussed on the absolute cheapest
resource that spam consumes. (Could someone illuminate this with some
history?) Had the mantra been 'they steal my time' or 'they steal my
client's time', I'd agree 100%.
The real irony is that that argument seems to be getting even more
weak, as CPU, storage, and bandwith prices continue to decrease, and
human time becomes more valuable.
Ergo, my pet-idea of, at some level, essentially forgetting about
those ever-cheapening resources; bandwidth, CPU, and space and instead
focus on the only resource that is expensive, and getting more
expensive: human attention&time.
Thus, without *really* good arguments, I can't see the point of ideas
that embed complicated policy langauges into SMTP. Nor others that
force a sender to burn an infitestimal amount of CPU; I'd rather burn
an ghz*hour of CPU than waste a minute of my time.
Now, my pet theory of either reducing the cost, in human time to the
recipients of unwanted communication, or increase the cost, in human
time, to the sender. (And do it in a way that avoids breaking mailing
lists.) I throw out CAPTCHA as a starting point.
Scott
[1] Assume 1ghz for $10,000 over a lifetime of 3 years or 100M
seconds, this comes out to $.0001 per second of 1ghz CPU.
[2] Colo seems to be about 300 gigs monthly at $3500, taken from
www.anchor.net.au; australian dollars.
[3] HD is now $1/gig at commodity prices. 50 is conservative. Assume
they last 100 weeks.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg