[Thumbgps-devel] the serial alternative/radio noise

[Eric S. Raymond]

Dave Taht <dave.taht at gmail.com>:
> I note that in the early stages of any project, I tend to appreciate
> all sorts of blue-sky thinking, rather than tie down to any one goal,
> and try to encourage such in others.

Fair enough.  But, as the person who has sort of dropped into the role
of project manager, it'a my job to ensure that (a) we have a coherent
and achievable set of objectives, and (b) the engineers keep their eye 
on that ball.  (For completeness, I consider the third major challenge of my 
job to be (c) negotiating with vendors and fabricators.)

For me, this is an amazing adventure, really. Me, a defrocked
mathematician, cat-herding a...*hardware* project.  Riiiight.  Good
thing I'm a naturally arrogant person or I'd probably be curled up in
fetal position in a corner somewhere.

Note, I also want to try this as a kind of experiment in just how far
down the post-industrial road we actually are.  I've been reading and
thinking for years about the demassification of production, silicon
foundries, 3D fabbing and other technologies that are supposedly
making semi-custom and individualized products more viable then ever.
The economist in me wants to grok these changes from ground level.

>                       I guess in part what drives me in
> suggesting alternatives (and please note my alternative above was
> intended as a lead-in into the noise question more than an actual
> suggestion), is that I would like to find markets for a 'new device',
> that 'does new stuff', that justifies the manufacturing run, more than
> a 'usb gps with hyper-accurate pps and time output' may, as a
> differentiator.

I understand that urge, but the increment in complexity required for
us to do "new stuff" would be all too likely to complexify the project
out of the zone of feasibility.  Let's do a design for which we *know*
that we have established need and a clear specification first.  If the
team gels and we develop good relationships with our vendors and
fabricators, we can then use that to go back around to things like
randomness sources and environmental sensors.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>