Are you talking about the speech program published by Quiksilver? That's not what I am referring to, I believe that someone managed to make the Z80 itself produce notes (since most older computers make some kind of humming noise depending on what instructions they are executing).
We need to make the most of GPS while we can. For the amount of money the Europeans are considering, they could employ thousands of people to walk around the globe, write down GPS coordinates at each location, and stick them onto the ground on Post-It notes (tm).
Hurry - we don't have much time to finish this task before the US government switches off the GPS system!
The BBC Micro had a built-in UHF-PAL output so you could view the display on your television. But it didn't include sound; that came out of the computer's built-in speaker. If you did turn the volume up on your TV you'd just hear noise, presumably it was caused as a byproduct of the picture generation and nobody bothered to filter it out. I remember generating lots of different buzzes and sqeaks from the television by poking random bytes into the video controller's registers and turning up the TV volume.
I also read that the Sinclair ZX80 - a machine with no sound hardware whatsoever - could be made to play notes. There was a program listing in some magazine (ZX User, I think) to do it. Not having a ZX80 (the predecessor to the ZX81) I couldn't try it out, and I never checked whether the magazine's date was April.
It always used to be that.org was for non-profits, but the rules have gotten relaxed over the years (AFAIK). If you made a domain strictly reserved for non-profit organizations, then corporations would not be able to sue for ownership of (for example) gateway.org, if a nonprofit group with that name legitimiately owns it.
I'm talking about how things _should_ be, not how they are:-(.
If company names are unique only at state level, then I'd suggest.com.tx.us and so on. The idea is to make a defined space where company names determine domain ownership, so that companies will not be able to claim ownership of other domains (like.org and.net) for the same reason. Make it clear that foo.org has no particular connection with Foo, Inc. whereas foo.com.$STATE.us does, and you should eliminate at least some of the lawyeritis that clogs the current TLDs.
We don't need any pain-in-the-ass buerocracy that will determine if we're truly deserving of a specific TLD either.
I agree, but some people will try to inflict PITA bureaucracy whether you ask for it or not. Witness the arbitration system for.com (guinessbeersucks.com, gateway.com and so on).
My suggestion was aimed at keeping these people in a contained space. They can have their legally regulated.tm and.com domains - where the rules are clear and explicit, and not as arbitrary as the current system - and everyone else can use.fcfs for first come, first served.
Dammit, why couldn't they take the opportunity to do it *right* for a change?.com.us domains should be given only to a real corporation with that name..tm.us should be administered by the USPTO and subdomains given strictly on the basis of trademark ownership. Conversely, trademark considerations should not impinge on the other subdomains - as long as it is clear that this is the case, so nobody gets misled.
You could also have.org.us strictly for non-profits (or maybe.charity.us for legally recognized tax-exempt charities),.fcfs.us for strictly first-come-first-served assignment, and so on.
Still, one positive feature of the new setup is that there won't be artificial scarcity created underneath the.us domain. Someone can buy.co.us and happily hand out the levels below that. At least until the lawyers get their hands into it...
(Possible new business for Sealand: lawyer-proof.sealand domain names. If you can get yourself assigned a country code that is.)
Maybe you can't make people do stuff by changing their dreams, but you can certainly influence important decisions. For example:
"Recent reports suggested that Mullah Omar, facing almost certain defeat, had agreed to surrender Kandahar. But yesterday Ahmad Karzai, whose brother Hamid has been negotiating with the Taliban for the surrender of the city, said Mullah Omar had changed his mind because he had had a prophetic dream in which he remained in power. 'I have had a dream in which I am in charge for as long as I live,' Mr Karzai quoted Mullah Omar as saying."
I sense some serious black-helicopter possibilities here...
A lot of the static checking made possible by Cyclone can be done for ordinary C with lclint, which lets you add annotations to C source code to express things like 'this pointer may not be null', 'this is the only pointer to the object' and so on. You write these assertions as special comments, for example/*@notnull@*/. These are checked by lclint but (of course) ignored by a C compiler so you compile as normal.
(If you weaken the checking done, lclint can also act as a traditional 'lint' program.)
Also C++ provides a lot of the Cyclone features, not all of them, but it certainly has a stronger type system than C. I'd like to see something which combines all three: an lclint-type program that lets you annotate C++ code to provide the extra checks that Cyclone (and lclint) have over C++.
Re:Bad screenshots for showing anti-aliasing
on
KDE 3.0 Screenshots
·
· Score: 1
It would help if people made screenshots of low-colour (16 or 256) desktops. In most cases, you can still demonstrate things just as well (unless you're showing off some new theme) but download time is slashed.
Re:too bad it was going to be a big leap forward
on
XBox Released
·
· Score: 3, Funny
From the article:
it's not clear the Bungie name will draw in console players. Especially considering its "hero," known only as the Master Chief -- a nameless ultracommando in a bulky power suit.
Now if they changed that to Master Chef - and had Lloyd Grossman flying around in a shiny suit - they might have a surefire hit.
That's what I said - the number of setuid programs should be reduced, slashed even. It is very bad practice for Mandrake to install games suid root, I hope they have stopped doing that. So I can appreciate that this is a reason to avoid Mandrake on a server, unless they've cleaned up their act. (In fact it sounds like a good reason to avoid Mandrake altogether - if I ever install it again I'll have to do a brief security check afterwards.)
But that's not what the original messages were talking about. They criticized disk usage and installing lots of packages as if that were in itself a risk on a server. It is not. Provided there are no setuid root binaries or daemons getting started, it doesn't hurt to have a good selection of packages installed. And it's not as if the disk space costs anything.
Why does it matter to have 'crap' installed? It's not as if the disk space is expensive.
On a server (or on any machine in fact) you want to minimize the number of services running, suid binaries, and so on. But why should there be any problem having ordinary, non-suid files lying around the filesystem?
If your system security is adversely affected by having a copy of same-gnome installed then you have bigger problems than worrying about 'bloat'.
I think a penny a page is a great idea. It would provide a real incentive for me to click through the pages and pages of comments on stories like this one. Even trolls and frist posts would get a chance of being read, if I could get a small payment for each one.
When it comes to Jon Katz, however, I'd probably demand a bit more payment for my time.
Three times in the past, Intel has tried to move away from the x86 architecture to a new, more modern one. The
iAPX 432, the i860, and the i960 were all moves in that direction. All three were dismal flops.
Intel own Alpha. And StrongARM. Why don't they try those?
One way might be to configure the slow RAM as a ramdisk and use it as swap.
I considered that, but it's a lot of overhead pagefaulting and going through the VM system and ramdisk driver every time a page is needed from outside the eight fast megabytes. It works fine to run processes using the memory above 8M, it's just slower. So I'd like Linux to somehow 'prefer' using the lower pages of memory and maybe rearrange things occasionally so that the more frequently used pages are kept in the fast RAM.
I'm currently wondering about NUMA - or something close to it. I'm running Linux on a couple of machines where the memory is of differing speeds: a fast eight megabytes and then the rest of the RAM is a lot slower. Can existing Linux kernels handle that sensibly?
Perhaps the possibility (if still a while off) of ternary computing devices will encourage some people to avoid meaningless powers of two when picking arbitrary limits. I'm talking about things like the maximum length of a filename: no point making it 64 characters, when 50 or 100 would work just as well and not cause the reader of the code to worry about whether the power-of-two-ness was significant.
Of course for buffer sizes you sometimes want the space to fit into an exact multiple of pages in memory. But a structure definition that has one int, one char and an int[128] is really pointless. Just make it a round number in base ten!
I believe that Alan Cox prefers GNOME to KDE. Therefore I have a cunning plan for an Alan Cox Trap. I will set up a machine with a specially patched kernel to use Rik's VM when GNOME runs, and Andrea's VM when running KDE. Mwhahaha....
Slashdot Mirror
Are you talking about the speech program published by Quiksilver? That's not what I am referring to, I believe that someone managed to make the Z80 itself produce notes (since most older computers make some kind of humming noise depending on what instructions they are executing).
We need to make the most of GPS while we can. For the amount of money the Europeans are considering, they could employ thousands of people to walk around the globe, write down GPS coordinates at each location, and stick them onto the ground on Post-It notes (tm).
Hurry - we don't have much time to finish this task before the US government switches off the GPS system!
The BBC Micro had a built-in UHF-PAL output so you could view the display on your television. But it didn't include sound; that came out of the computer's built-in speaker. If you did turn the volume up on your TV you'd just hear noise, presumably it was caused as a byproduct of the picture generation and nobody bothered to filter it out. I remember generating lots of different buzzes and sqeaks from the television by poking random bytes into the video controller's registers and turning up the TV volume.
I also read that the Sinclair ZX80 - a machine with no sound hardware whatsoever - could be made to play notes. There was a program listing in some magazine (ZX User, I think) to do it. Not having a ZX80 (the predecessor to the ZX81) I couldn't try it out, and I never checked whether the magazine's date was April.
It always used to be that .org was for non-profits, but the rules have gotten relaxed over the years (AFAIK). If you made a domain strictly reserved for non-profit organizations, then corporations would not be able to sue for ownership of (for example) gateway.org, if a nonprofit group with that name legitimiately owns it.
:-(.
I'm talking about how things _should_ be, not how they are
If company names are unique only at state level, then I'd suggest .com.tx.us and so on. The idea is to make a defined space where company names determine domain ownership, so that companies will not be able to claim ownership of other domains (like .org and .net) for the same reason. Make it clear that foo.org has no particular connection with Foo, Inc. whereas foo.com.$STATE.us does, and you should eliminate at least some of the lawyeritis that clogs the current TLDs.
I agree, but some people will try to inflict PITA bureaucracy whether you ask for it or not. Witness the arbitration system for .com (guinessbeersucks.com, gateway.com and so on).
My suggestion was aimed at keeping these people in a contained space. They can have their legally regulated .tm and .com domains - where the rules are clear and explicit, and not as arbitrary as the current system - and everyone else can use .fcfs for first come, first served.
I was thinking of .tm subdomains corresponding to trademark categories - so olympia.food.tm.us, olympia.travel.tm.us, and so on.
.com.us, I was thinking of a company registered with that exact name. Company names have to be unique, right?
For
Dammit, why couldn't they take the opportunity to do it *right* for a change? .com.us domains should be given only to a real corporation with that name. .tm.us should be administered by the USPTO and subdomains given strictly on the basis of trademark ownership. Conversely, trademark considerations should not impinge on the other subdomains - as long as it is clear that this is the case, so nobody gets misled.
.org.us strictly for non-profits (or maybe .charity.us for legally recognized tax-exempt charities), .fcfs.us for strictly first-come-first-served assignment, and so on.
.us domain. Someone can buy .co.us and happily hand out the levels below that. At least until the lawyers get their hands into it...
.sealand domain names. If you can get yourself assigned a country code that is.)
You could also have
Still, one positive feature of the new setup is that there won't be artificial scarcity created underneath the
(Possible new business for Sealand: lawyer-proof
Maybe you can't make people do stuff by changing their dreams, but you can certainly influence important decisions. For example:
I sense some serious black-helicopter possibilities here...
Presumably the same could be done for <meta name="keywords"> in HTML.
A lot of the static checking made possible by Cyclone can be done for ordinary C with lclint, which lets you add annotations to C source code to express things like 'this pointer may not be null', 'this is the only pointer to the object' and so on. You write these assertions as special comments, for example /*@notnull@*/. These are checked by lclint but (of course) ignored by a C compiler so you compile as normal.
(If you weaken the checking done, lclint can also act as a traditional 'lint' program.)
Also C++ provides a lot of the Cyclone features, not all of them, but it certainly has a stronger type system than C. I'd like to see something which combines all three: an lclint-type program that lets you annotate C++ code to provide the extra checks that Cyclone (and lclint) have over C++.
It would help if people made screenshots of low-colour (16 or 256) desktops. In most cases, you can still demonstrate things just as well (unless you're showing off some new theme) but download time is slashed.
Now if they changed that to Master Chef - and had Lloyd Grossman flying around in a shiny suit - they might have a surefire hit.
That's what I said - the number of setuid programs should be reduced, slashed even. It is very bad practice for Mandrake to install games suid root, I hope they have stopped doing that. So I can appreciate that this is a reason to avoid Mandrake on a server, unless they've cleaned up their act. (In fact it sounds like a good reason to avoid Mandrake altogether - if I ever install it again I'll have to do a brief security check afterwards.)
But that's not what the original messages were talking about. They criticized disk usage and installing lots of packages as if that were in itself a risk on a server. It is not. Provided there are no setuid root binaries or daemons getting started, it doesn't hurt to have a good selection of packages installed. And it's not as if the disk space costs anything.
Why does it matter to have 'crap' installed? It's not as if the disk space is expensive.
On a server (or on any machine in fact) you want to minimize the number of services running, suid binaries, and so on. But why should there be any problem having ordinary, non-suid files lying around the filesystem?
If your system security is adversely affected by having a copy of same-gnome installed then you have bigger problems than worrying about 'bloat'.
I think a penny a page is a great idea. It would provide a real incentive for me to click through the pages and pages of comments on stories like this one. Even trolls and frist posts would get a chance of being read, if I could get a small payment for each one.
When it comes to Jon Katz, however, I'd probably demand a bit more payment for my time.
Intel own Alpha. And StrongARM. Why don't they try those?
When I glanced at the headline, I thought it said 'KDE 3 Wins awards'. I know KDE development is fast, but still it was a bit surprising...
Do you think I should buy a .biz domain for my surname?
bbc.com was Boston Business Consulting. I wonder how much money they got...
I'm currently wondering about NUMA - or something close to it. I'm running Linux on a couple of machines where the memory is of differing speeds: a fast eight megabytes and then the rest of the RAM is a lot slower. Can existing Linux kernels handle that sensibly?
Perhaps the possibility (if still a while off) of ternary computing devices will encourage some people to avoid meaningless powers of two when picking arbitrary limits. I'm talking about things like the maximum length of a filename: no point making it 64 characters, when 50 or 100 would work just as well and not cause the reader of the code to worry about whether the power-of-two-ness was significant.
Of course for buffer sizes you sometimes want the space to fit into an exact multiple of pages in memory. But a structure definition that has one int, one char and an int[128] is really pointless. Just make it a round number in base ten!
Suppose Linux did fork...
If Linus's version is called linux, what is Alan's version called?
I believe that Alan Cox prefers GNOME to KDE. Therefore I have a cunning plan for an Alan Cox Trap. I will set up a machine with a specially patched kernel to use Rik's VM when GNOME runs, and Andrea's VM when running KDE. Mwhahaha....