Very welcome comment on the "third way" from Brazil there !
(And sure I had been too generalising about NGOs[*].)
And now there seems to be even a (hmm, almost) ready-made model from
NZ too !
I wonder about one thing regarding the "load" or processing charge of
the server to the "cluster" (of SurvPCs): isn't this mostly determined
by the volume of "graphical" processing ? And could scarcer use of
graphical elements considerably increase the server's ability to serve
for *more* clients with the same "outside" line ? While at the same
time this would maintain even the most "primitive" SurvPCs to survive,
as clients, without excluding more sophisticated applications on others ?
I find HealtNet exemplary in that respect: It was built up on the base
(and at the time) of XT/AT286 class of clients, and strictly oriented
to text-mode use - which did not exclude the clients to get those
(high-resolution!) graphic images of, say, microbes they needed but
strictly avoided to move wallpapers and office "furniture"[**] around
the world over scarce and high-priced connections (all-satellite in the
beginnings). By now it has developed into a full-blown WWW-based
network but still maintains those basic criteria: which means that
"poor" medical outpost in the Andes or in the Kongo basin which never
would have the equipment nor could follow he hasty "innovation" pace of
browser versions in N.Y. offices, still have the best possible
standard of information=*content* available, at level with what you
have in central regions of the "Northern" hemisphere (or in the
parachuted container compounds of oil companies, or other "northern"
business isles like the Hiltons or World Trade Centres in the middle
of "southern" 5-million-towns with less telco lines than a small
residential quarter in a suburb of Hamburg).
When Eko triggered off this avalanche of a thread it appeared to me
that this "cybercafe" model could as well be seen as a "microweb"
which would transpose that functionality of HealthNet to a small,
local or "neighbourhood" scale. While electrification did advance
considerably, telco "penetration" lacks behind bitterly; and no matter
what tariff structure, telecomm will remain priced far beyond
affordability of many households (one of the reasons why radio an TV
were instant successes). But just one (more) line to the handful that
go to a "bush village", and such a neighbourhood server, say, in the
school house, could make for a tremendous difference. The more so if
it's used in the same efficient way as the HealthNet use of
connectivity.
(BTW Eko, what do they charge you for electricity at that rather weak
end of your line ? Here in Belgium we're sitting almost at the highest
end of the tariff scale in all EUrope, with approx. 0.15 US$ per kWh,
thanks to the heavy nuclear power overhead and the private monopoly
of the national power generator Co.; same for telco: cheapest=night
time local call, which ist the rate for IAP dial-in numbers too, is
1.60 US$ per hour, daytime/office hours it's 3.10 US$ !)
The model could even be viable in highly developed=highly telco
metered environments - our house has four telco lines (to the four
households), on three of which are "netters" (the fourth has just
moved in, I don't know yet). With modest use, each is at least good
for 30 US$ telco fees for net use per month (I'm near 60$ p.m. for my
part already). With one additional line (and counting the overhead for
that as well as the current the server draws) for the "house ISP" we
could be at par - and have our normal lines free for voice -, but most
probably we would be better off, as two of the three work much online;
each minute two of us or more would connect to the outside line at the
same time, the fee is divided. Taking the spread of online use in
account and some hardware overhead (another, "reserve" '486 or low-end
Pentium, one to be exchanged each year), such a "neighbourhood ISP"
could be viable economically from five to six clients onwards.
And would be so much more so under condition of scarce ouside lines.
// Heimo Claasen // <[EMAIL PROTECTED]> // Brussels 1999-09-03
HomePage of ReRead - and much to read ==> http://www.inti.be/hammer
-----
[*] There are differences like day an night - people who have "grown
up" with FIDO connections know well what they do and how to use
connectivity efficiently, especially when they have passed some time
somewhere in the African bush. But I'm surrounded here with the lobby
offices of all sorts of NGO networks which got their office equipment
and setups lock, stock and barrel from the marketeers - and the people
working with it have no training, no experience at all. Thus the
central office of the major network of EU development NGOs blasts
their memos to "partner" NGOs in the South as MS-Word-8 "docs" -
incompatible with anything else - and thus force them to throw the
project money out of the Window$ to "upgrade". Another guy who runs
communications for, of all, a network for rural development in the
Sahel regions (all iMacs for them here in Brussels) simply does not
know, and does not see, that each and any one of their memos, reports
and notes is sent off to Timbuktu threefold, as (a) MIME-QP, (b) HTML-
formatted, (c) Base-64 encoded M$-Word-"document".
[**] the images are even called like that in the code of the BBC's web
pages which are crammed with them. On some of the pages, I counted
the ratio between "furniture" and content to 80 to 1.
(The BBC had somewhat sober, but straightforward web sites - that of the
World Service especially usefull - until a year ago when they
"outsourced" their web "authoring" to a consultancy/contractor.)
-hc
To unsubscribe from SURVPC send a message to [EMAIL PROTECTED] with
unsubscribe SURVPC in the body of the message.
Also, trim this footer from any quoted replies.
