TECH:NOMAD

Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Sat, 31 May 1997 14:00:56 +0200 (MET DST)


On Sat, 31 May 1997, Felix Ungman wrote:

> > [ how much can be done on a laptop? ]
> It's not a technological problem. You can get >200 MHz laptops with
> enough memory to do almost anything you would with a desktop
> computer. Internet is usually not a problem, 56K modem or perhaps

Bad thing with them, you can work just a few hours in untethered mode.
Admittedly, this is rarely required. I'd still like a solar augmented fuel
cell, for religion reasons. Roaming is getting better, but one'd still
need Iridium to be a "go-anywhere" technomad. Iridium is going to be
expensive, though, I wonder about what Teledesic's rates might be (nudge,
nudge).

I am really looking forward to wearables, with really adequate headup,
large storage and sophisticated GUIs. Using cellphones has become socially
acceptable very rapidly, I wonder however how long it takes until people
with headsets, heavy hipbelts and black monocles, with a micro CCD
embedded, while babbling to invisible presences, won't raise eyebrows
anymore. I think Linux to be an excellent platform (most of the MIT
wearables used it), albeit speech input was pretty nonexistant half a year
ago. As PC hardware has the embedded microcontroller market now pervaded,
cheap low-juice hardware should be quite easy to get. Any proprietary
distributed embedded solution would be superiour, but clearly noviable due
to nonexistant software base. I am still waiting for any major outfit
having the guts to challenge the PC cul de sac, by build a consumer maspar
DSP box. This sure'd be a reason for throwing a major party.

Those who are signed up with a CSP are going to need an automatic
emergency pager anyway, so'd they'd be using biotelemetry, and a
cellmodem, already. Consider this a CFI (call for ideas) for just such a
pager; which can be incrementally morphed into a wearable. I know Hara Ra
is interested in building just a gadget, prior to that we the local Kraut
cryos had also plans which were progressed pretty far, but which died due
to insufficient funding. I'd still like this to result in a viable design,
whether commercial, or in the public domain.

> ISDN will get you far enough. As a mobile knowledge worker, you
> don't need any reference books, most of the information is on the
> Internet or on CD-roms.

True. I am going to scan and OCR the bulk of my library in the course of
next years, though. Some of the relevant stuff isn't on the net yet, and
some never will.

ciao,
'gene

P.S. I'm reading up on some peculiar crackpot topics just now. I have now
firmly identified Bohm as the originator of the quantum mind conspiracy.
I have found early Kauffman in "Kinetic Logic -- A Boolean Approach to the
Analysis of Complex Regulatory Systems", a 1977 Lecture Notes in
Biomathematics Proceedings in Brussels. "Vision, Instuction and Action"
by David Chapman is a great read on how far brittle AI is going to get us
(this is an expert system playing an arcade game. It uses vision, but it
does not ANN nor GA). The "Art of Molecular Dynamics" by D.C. Rapaport is
a good in-depth introduction for those interested in modeling nanotech. I
am reading an ancient Russian book on design of spacecraft, the book on
mail from Marshall T. Rose's Internet trilogy. I also came across early De
Garis, his LIZZY (GA evolving neural control) sounds very cool. As an
uploader, I am worried about suspected CA-like computation properties of
microtubulin, though this has fair chance of being bogus, having been
published in the infamous "Rethinking Neural Networks: Quantum Fields and
Biological Data".