Felger said:
No. Wrt 'thin clients', you're more like the crazy uncle who escaped
from the locked room in the attic. C'mon, Robert, just how many
.chips denizens are prepared to trust *all* their personal data (now
on their hard disk) to the not-so-tender mercies of Amalgamated ISP
Heavily Incorporated at the far end of a serial line? Besides
yourself, I mean. ;-) ;-)
I wouldn't venture to guess. I haven't figured this crowd out yet.
It's actually not something I'd be willing to guess about in general
were there any of my own money on the table. Somebody trying a cold
start on this would be facing some serious up-front costs: get enough
test subjects to try it, then find out what they really think about it.
Google has the advantage that they can just reel people in by slowly
expanding the boundaries of what's available. How much reeling would it
take before any substantial number of people were ready for diskless
operation? I don't know, but, unlike you, I don't see any show-stoppers.
I see two very hard problems, but I don't think they are show-stoppers.
One problem is that you need a way to establish enough bulletproof
trust between the diskless client and the managing server to allow
actual remote management in a situation where EFI presents the very real
possibility of the equivalent of a BIOS break-in. If trusted computing
is going to work for, say, corporations, it has to solve that problem.
The second problem is that you need to find a way for the remote storage
to act like a safety deposit box to which the bank doesn't have ready
access. I'm assuming that existing corporate remote storage services
must solve that problem in some way.
..chips denizens keep *all* their personal data on hard disk? You're
kidding.
A properly-managed remote service should be able to provide better
security and privacy than most people actually have today. Actually
convincing potential customers of that might, I admit, be a challenge,
but that's what focus groups are for.
RM