[Fwd: Re: [Fwd: [HARDWARE] Just Courious......]]

Robert Norton rjn at execpc.com
Sat Sep 18 14:09:08 EDT 1999


"William J [on the road] Halverson" wrote:

> There is a minimum charge that shippers will ship at; handling an one item order
> worth $30 costs as much as handling a one item order worth $300.
>
> So the group may want to consider - is it easier to sell 750 @ $300, or 7,500 @
> $30?  The answer may surprise you .... we are just talking about recovering  costs
> here ... you can imagine the guy who puts in 4 man months of his time will have
> more 'at risk' if we have to sell 7,500 than 750...
>
> So suppose:  we want to make a go/no go decision on this in two months.  So we
> have to find the designer [probably done], we have to get a bank account to escrow
> the money [no I don't take personal checks or cash ;-) ], and we have to publicize
> this and get the funds together to make it a reality.
>
> People in on this should be serious and feel safe:  they should not be expected to
> send somebody a check; nor should the developer be expected to put 4 months into
> it on a 'wing and a prayer'.
>
> Does the group think this is reasonable?
>
> Any suggestions on a trusted entity/persons to supervise this thing?  [Hmmm...
> would Tim LB get excited about this maybe?]

I think 1000 units at $100 with $20 "profit" per unit would leave $20,000 for
developement costs.  I feel it should be pretty easy to sell 1000 at that price,
especially since a fully unrolled algorithm in an FPGA would crack one code per tick.
That would be 25Mcrack/sec even with the slowest parts!  Also worth considering, with
a fully unrolled algorithm, you don't need any barrel shifting, it is just static
wiring from latch to latch. (Output #2 goes to input #3 of the next stage, and so on.)
That should cut the gate count a bit.

--
To unsubscribe, send 'unsubscribe hardware' to majordomo at lists.distributed.net



More information about the Hardware mailing list