Unix in a keyboard-less age

nslay said:
Complexity doesn't necessarily mean it's horribly unreliable or inefficient.

No, complexity doesn't guarantee that your software is less reliable or inefficient, but it has historically proven to be be a fairly reliable indicator.

No matter how efficient you make your code at what it does, equivalent quality code that DOES LESS will consume less resources, be easier to debug and more difficult to exploit.


Put this way: how many product recalls do you see on new bicycles, compared to new cars?

We've been building cars for say, ~100 years now. And yet we still put out new cars with recall-worthy faults, despite most new vehicles going through a development and testing process that is far in excess of what the average bicycle design endures.
 
throAU said:
No, complexity doesn't guarantee that your software is less reliable or inefficient, but it has historically proven to be be a fairly reliable indicator.
Historically, we've been advancing in technology ... and it's still pretty reliable (sometimes even more-so than older and simpler technology). Granted, more parts does make for more points of failure.

No matter how efficient you make your code at what it does, equivalent quality code that DOES LESS will consume less resources, be easier to debug and more difficult to exploit.

The real kicker is that machine learning methods build themselves (small chance for human error), they're algorithmically simple, and typically very efficient. But they're also often black boxes ... small price to pay for smarts.

We've been building cars for say, ~100 years now. And yet we still put out new cars with recall-worthy faults, despite most new vehicles going through a development and testing process that is far in excess of what the average bicycle design endures.

So, what are you suggesting? That we shouldn't try to replace the keyboard and mouse with something better because they're simple? By that logic, we should still be using horse and carriage or even bicycles. Those two are very reliable and extremely simple and the former is pretty practical for longer commutes.

That reminds me, horse and carriage have largely vanished and I anticipate the same for the keyboard in the distant future, much like its simpler predecessors: the punch card, paper tape, switches, etc...

If the keyboard exists 100 years from now, then I'd agree, it might be purely for diagnostic purposes. I'm curious how UNIX will adapt to a keyboard-less environment.
 
Back
Top