Re: what's a callback?
From: Steve at fivetrees (steve_at_NOSPAMTAfivetrees.com)
Date: 12/25/04
- Next message: John Woodgate: "Re: Is zero even or odd?"
- Previous message: mike: "Re: diy vga signal from ram"
- In reply to: Frank Bemelman: "Re: what's a callback?"
- Next in thread: Anthony Fremont: "Re: what's a callback?"
- Reply: Anthony Fremont: "Re: what's a callback?"
- Messages sorted by: [ date ] [ thread ]
Date: Sat, 25 Dec 2004 00:50:39 -0000
"Frank Bemelman" <f.bemelmanq@xs4all.invalid.nl> wrote in message
news:41cc7617$0$6219$e4fe514c@news.xs4all.nl...
> You should not compare your embedded stuff with de desktop computer
> running windows. Those are two entirely different species. Your 68K
> does only one thing - easy. If I plug your software in whatever 68K
> board, nothing happens, nada, zip. These are projects that can be
> done by just one or two persons, that's why I said it is kindergarten
> stuff compared to windows.
Hmmm.
I'm not sure there needs to be any distinction between the two: "good"
design remains good, regardless of platform, or degree of complexity.
Granny and her boot time are part of the requirements. Bloat and reliability
are down to the implementation. The two are - and should be kept - separate.
Different mindsets and skillsets.
A "good" programming team should be able to take a well-designed requirement
(for e.g. a vapourware, intuitive OS or application) and turn it (using best
practice, with maximum synchronicity/determinism) into a piece of code which
does what's required, with no surprises and no side-effects. And no bloat.
What's so hard about that? We do it all the time in the embedded arena.
Complexity is *not* the issue.
I deal mainly in bare-metal bulletproof (but plenty complex enough - mainly
cooperative multitasking) code. The lessons I've learned over the years
translate well to the desktop (given a good enough spec). Mainly they have
to do with *managing* complexity, not letting it manage me.
This is a soapbox of mine. I hear "complexity" bandied about as an excuse
for poor code all the time, and it *really* pisses me off. It's a sham, a
smokescreen. Grrrr.
> Imagine you had to deliver some kind of open system, where 3rd parties
> plug in their boards & drivers in your box. Would you like that? Can
> you still guarantee your product doesn't crash?
OpenBSD seems to be a pretty good example of an OS that's as reliable as it
gets. I *heart* OpenBSD. (One of my commercial webhosting servers had a
record uptime of 457 days, despite being under hacker attack 24/7 - only
interrupted by a fairly major system upgrade ;).)
Summary: it doesn't have to be this way. There is still a market out there
for a reliable, well-written, reliable, easy-to-use, and mainly *reliable*
desktop OS. Windows ain't it, and neither are *nix nor the Mac (although the
latter are probably/possibly closest). We've had decent multi-level MMUs for
years - a rogue app needn't pull the whole system down. Bah humbug.
Meanwhile, may I wish you all a very merry and restful Christmas!
Steve
http://www.fivetrees.com
- Next message: John Woodgate: "Re: Is zero even or odd?"
- Previous message: mike: "Re: diy vga signal from ram"
- In reply to: Frank Bemelman: "Re: what's a callback?"
- Next in thread: Anthony Fremont: "Re: what's a callback?"
- Reply: Anthony Fremont: "Re: what's a callback?"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
