Thought Game
If computer hardware stopped getting better today, then what would software be like in one hundred years?
'Joel' on Thu, 16 Feb 2006 12:33:25 GMT, sez: wow, that's a really good question ... hmm
'Joe' on Thu, 16 Feb 2006 12:35:41 GMT, sez: In Cringely's Triumph of the Nerds on PBS, Steve Wazniak has said that he can't wait for "Moore's Law" to finally play itself out. Once the hardware is constant and software can't rely on the hardware to make it seem faster in N months/years, then software will finally start to mature.
My opinion, if hardware froze at today's levels, then software would continue to suck. Trouble is that there's no way that hardware will ever freeze as such, because like cars and buildings and any other form of hardware, engineers know there's always room to improve and patents give that incentive to keep trying. Even when moore's law freezes the speed of chips, then they'll learn to just throw more chips at a problem - you can always add more memory and more processors, provided your firmware engineers can write robust software to control it all.
'Haacked' on Thu, 16 Feb 2006 14:19:07 GMT, sez: 100 years is a LOOOOONG time. Look at the innovations we've made in last 10 years.
If hardware remained constant, I think we'd see software become much better at parallelization. Agent oriented software would flourish.
'leon' on Thu, 16 Feb 2006 20:03:42 GMT, sez: i hadn't thought about parallelization. i was thinking that that kind of thing was frozen too, i guess.
remember the old old days when people had 16K to play with and they managed to use all sorts of tricks to squash amazing amounts of stuff into that space, and still have highly responsive software.
if the same level of optimisation was performed on today's machines, i think we'd have many magnitudes of improvement in our software.
when you think about what those old machines did with so little, and you look at what we do with so much, it's almost embarrassing. we're drowning in extra memory and extra hard drive space and hardly realise it!
(my brother came up with this question years ago)
'Farmer Jeb' on Fri, 17 Feb 2006 00:58:51 GMT, sez: (Did I?)
'lb' on Fri, 17 Feb 2006 01:05:35 GMT, sez: >(Did I?)
not you, my other brother.
no, of course you. Yeh, i remember you said it once, when we were talking about star trek or something.
how if computers didn't keep getting bigger then software could actually settle down for a minute and start to have real standards. and things like patches and settings would stop being a voodoo science and just be uneccessary.
really usable computers would be just ordinary, taken for granted, much like turning on a tap and expecting water to come out.
oh, you musta said a lot o things, the way i remember it.
'mike' on Fri, 17 Feb 2006 05:28:41 GMT, sez: >they managed to use all sorts of tricks to squash amazing
>amounts of stuff into that space
Yeah, but don't forget that a lot of those tricks for the 640K days (not so much was accomplished with 16K) were pretty dodgy, and it wasn't so unheard of for programs to hammer each other. Assuming you could even run more than one at a time, which mostly you couldn't. (Of course, the 8080 architecture practically demanded that people to come up with weirdness like the LIM standard for accessing extended memory. "Oops, did we put video memory smack into reserved address space in the base 1MB addressable? Sorry about that.")
'aaron' on Wed, 15 Mar 2006 19:02:00 GMT, sez: you know...I remember when I got my first palm pilot. Professional. USR logo. (you remember US Robotics, don't you?)
$500. 1 megabyte of RAM.
And, if pudits were to be believed, more computing power then we used to send a man to the moon....
So if that's true, we should be able to send a man to mars with my nifty new 312mhz Zire 72. Posh on NASA for saying it's difficult..
Right?
|