Re: Remember "Bit-Slice" Chips ?

On 2024-12-15, The Natural Philosopher <tnp@invalid.invalid> wrote:

On 15/12/2024 04:57, 186282@ud0s4.net wrote:

"seems close enough"
   is NOT good enough. Planes, spacecraft, bridges, huge
   buildings, medical implants - GOTTA refine with the
   hard-core/hard-math tools.

>
I think you would be aghast at how "seems good enough" guides most
engineering design.
>
No-one accurately measures every single component that goes into a design.
>
At best they do a full test on the final product.
>
There is always room for the black swan unit where all the tolerances
were exactly the wrong way.
>
In general it is cheaper to simply scrap that one, or if it escapes into
the wild, give the customer a replacement.
>
The development algorithm of the racing Cosworth V8 was "remove metal
till it breaks, then put that bit back again".
>
And we can only calculate what we thought of. Some failure modes are
completely unexpected.
>
Some of the most durable civil engineering was done by Victorian
engineers who were not able to do the calculations. Their conservative
over-enginering resulted in structures that stand good even to day.
>
Admittedly their failures are long gone :-( (Tay bridge, any one?)

Don't have chapter and verse on a source, but I understand some
sources say the reason Roman civil engineering created such
durable structures was because the engineer who designed each
arch was required to stand underneath it while the scaffolding
was removed.  The engineers who didn't over-engineer their arches
didn't design many of them.

The thought of doing long division in Roman numerals fills me
with dread--let along manually doing finite element analysis.

--
Robert Riches
spamtrap42@jacob21819.net
(Yes, that is one of my email addresses.)