Leonard Erickson wrote in a message to Roy J. Tellason:

 LE> Alas, the molds required are fairly high precision blocks of solid 
 LE> steel. That's required by the pressures and temps involved. Which
 LE> makes them expensive.

 RJT> Are solid steel molds really required for one-off or low-volume runs, 
 RJT> though? 

 LE> They are required for injection molding. When running the molds are
 LE> at a few hundred degrees and under a *lot* of pressure (hundreds of
 LE> psi). 

 LE> And because the expense of making them, you don't *do* low volume,
 LE> much less "one off" runs.

I was just wondering if there was some other way to do it...

 LE> There's no method that's cheap for low volume use.

 RJT> I wonder,  but then I don't know enough about manufacturing processes,
 RJT> which all seem to be geared toward high-volume anyway. 

 LE> And, of course if they ever get molecular assemblers/fabricators to
 LE> work, all bets are off. No, this isn't "nano-technology". At least
 LE> not in the normal sense. These would be "normal" sized devices that
 LE> build objects one atom or molecule at a time.

 RJT> The people who make chips and such stuff are bumping into limits all
 RJT> the time,  and figuring out ways around it. 

 LE> Oh, we *could* build stuff up a molecule at a time with tech we
 LE> could build easily enough. But it'd take a few hundred years to
 LE> buiuld anything big enough to handle. :-)

Not necessarily.  The answer there is to go highly parallel,  and that
depends on some "fabricator" that can be made in high volume and
cheaply.  Maybe a tailored organism?  The lines are getting blurry, 
anyhow...

---