|
Authored by: Wol on Wednesday, April 17 2013 @ 09:31 AM EDT |
There's also a little matter of mindset. Take Linus for example.
Rule 1 of kernel programming is never EVER break user-space. Too many computer
science grads think that correctness is the most important thing. But it isn't
correct if the user's program stops working ... Linus is an engineer, not a
theoretician.
He also thinks that the correct solution to certain problems is to panic and
crash, because any automated attempt to fix it is likely to make matters worse.
Like me with relational databases (oh no, I'm on my hobbyhorse again! :-) but
seriously, how much real world data comes as sets? And how much comes as lists?
And how do you *store* a list in an FNF database again? It's pretty much
accepted that FNF databases are ponderous and slow, and imho most of that is
down to the huge engineering effort required to bash round pegs into the square
hole that is FNF - with all the error-prone faffing about that is required...
The Computer Science mindset is that the maths is correct - too bad the
engineering and science disagree ...
Cheers,
Wol[ Reply to This | Parent | # ]
|
|
|
|
|