[667] in Commercialization & Privatization of the Internet
Size of the NREN Market
daemon@ATHENA.MIT.EDU (Stephen Wolff)
Wed May 8 13:34:20 1991
To: tmn!cook@uunet.uu.net
Cc: com-priv@psi.com
Date: Wed, 08 May 91 13:23:03 EDT
From: Stephen Wolff <steve@cise.nsf.gov>
I share Craig Partidge's malaise with your figures for current users; several
sources suggest 1-3 million is closer than your 700k.
If you're going to categorize the universe, then as Bill Feller said in
Elementary Prob, the categories should be "mutually exclusive and
collectively exhaustive." Your categories 1 & 2 don't exhaust all of post
secondary society; how about changing #1 to "Post secondary people with
intensive computing needs" and #2 to "Other post secondary people," on the
grounds that not all those with "intensive computing needs" are scientists
(e.g., someone doing numerology on Dante's collected works - poor Charles
Singleton made a career of that without [much] benefit of even the late '60s
computers available to him), and (for #2) that there are non-social-
scientists and non-humanists with non-intensive computing needs.
Then I guess category 3 becomes "Everyone else," unless there is some segment
of society you are trying to exclude.
It's a peculiar enough classification that I'm led to wonder what motivated
it. I assume you're not about to make a "greatest good for the greatest
number" argument.
If one user in ten is "intensive" then your figure for #1 is probably ok.
Category 2 is probably low by a factor of two to five, since we need to get
to a final figure of 1-3 million and I agree category 3 is probably under a
half million at present (cf. Craig's discourse).
We really part company on the percentages you've chosen, except perhaps for
your <1% choice for #3. Why do you think that we've reached 20-30% of the
"intensive" users? Remember my tractor-and-Buick analogy; I think it likely
that the most UNsophisticated users (the ones we've gotten to less than 1%
of) may well be launching *very* complex and compute-intensive applications -
whose complexity will of course be utterly hidden by a friendly puppy dog
user interface, and whose users might well scoff at the notion that anything
very complicated is going on. (But consider what a simple PRNDL hides: have
you ever seen a cutaway of a GM turbo-hydro automatic tranny?)
And I'm totally flummoxed by the idea we've got "5 to 10% of the total" in
#2. Where'd that come from?
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Seems to me you're taking a current user population of uncertain size and
applying mutliplying factors of questionable lineage to arrive at a
"potential market" estimate that has neither bottom nor breeding to recommend
it.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Why not just estimate directly? The student+staff+faculty population in post
secondary schools is known. Ditto for K-12. There are estimates of numbers
of folk engaged in comercial/industrial research, and of those embarked on
new business ventures. Somewhere someone knows how many Government (local,
state, and Federal) info providers there are. Then just add in your favorite
reasonable percentage of the remaining populace-at-large who may want access
to the previously-counted people and resources for educational or research
purposes.
