All right, but we've had the a.out/ELF transition followed by the libc5/libc6/glibc2 transition. It kind of seems like there will be a new transition waiting when this one evens out.
(and that's why I bought an amp with digital input, plus a simple cheap little soundcard with the C-Media CMI8738 chip on it (SPDIF out). Now all I have to do is install the board and run the wire).
Now only remains this question: If it's better, then how do I upgrade my portable MP3/CD player to Ogg? Does AIWA make a CDC-Ogg yet like the CDC-MP3 they have for in the car? Damn.
"To get a valid comparison, rip the content of the CD as WAV. Then compare the WAV and the OGG, using the same soundcard and the same amp-input. Everything else is totally meaningless."
How 'bout this: rip the CD as WAV, Ogg and De-Ogg to WAV. Then write a couple of CDRs with all songs in the same order as the CD, but with some songs from the original WAV, and others the Ogged&De-Ogged WAV. Then do a blind classification trying to see if ou can tell if it's the original WAV or not by listening to the CDRs. If you are correct 50% of the time, then for you Ogg is lossless for your ears.
In my opinion that's the only way to really test. The lower quality of the soundcard output might hide differences in your suggestion. Plus the blind classification removes any psychological effects of expecting to hear differences or not.
"I actually agree with the concept setting up a lot of machines to reply to the virus with the fix."
The amusing part about that suggestion is that it is probably illegal to do so.
And if it were legal, then crashing the box might work better. Fixing it without the owner doing it will just lead to more lazyness on the part of the owner.
This is yet another example of patents stifling application of innovations. Thanks to this we'll see less AC3 on open source OSs. Where else is Dolby's AC3 library for Linux & BSD?
"It is MUCH more cost effective o spend he extra money on the fastest memory, memory buses, disks and controllers you can get."
Often that is true, but it depends on the application and on the balance point. For example If you're making a aystem with 1GB RAM or more, then the balance is often different. $33 price difference will give you 5% in CPU speed. But the same amount of money spent on RAM over 1GB often won't even make a noticable difference. I'd consider it a waste to build a system of 1GB RAM or more and anything less than the 1.4GHz Athlon. Remember that a faster processor can handle interrupts and context switches at lower latency too, so even if your CPU load is low, latency (response time) will be better with a faster CPU.
And then there are a lot of uses for PCs that are still very much CPU bound (no I'm not talking about setiathome and distributed.net). We have systems with 512MB, no disk and they're 100% CPU bound every time we're waiting on it... We'd wish 25GHz CPUs actually existed yet.
Plus, what is $33 on a CPU compared to going from a 7200RPM UDMA100 to a 10 or 15KPRM SCSI? if you're so afraid to be slow, then spend that $33 on the CPU (and get lotsa RAM too, will often speed you up much more than that slow disk, even at 15KRPM).
Cache is good, but don't go too far hoping that that large L2 on a Xeon will save you in all cases. Sometimes cache is almost completely irrelevant. If you're walking large datasets, then you're completely trashing your cache anyway, then spending the money on getting faster RAM chips is wiser.
Back to my point: It all depends on the balance point and application. Personally, for my own PC use, I'd like it that everybody keeps demanding faster CPUs, because that would speed up development and make it available sooner. Maybe we'll see a split in the future where there are 'normal speed' and 'high speed' CPU's, where the speed difference is an order of magnitude or more. Hey, wasn't it the same in the old days of mainframes/mini's/mirco's versus the home computers?
Plus if you're running jobs that take many hours or sometimes days, then a 5% increase can mean making that deadline to get the bonus for the project, that would easily pay for the $33...
It's just a matter of where do you stop. What if there was a 1.5Ghz for $2500. Would that be worth it? Hmm, a couple of places probably can help you towards trying that too.
The article here was probably written for home users though. Even then it wasn't on the mark, because the main issue that matters for home users is game speed.
And the great thing about is that when they find the safe, they can tinker in it on that tucked away little place where you hid it where nobody is looking anyway, so they won't get caught trying to break the safe.
Longer to find it, often not: You know how they would find the safe? They simply follow you when you go out to check on it. Bye Bye security, it helps nothing.
Often obscurity is not what it seems, the obscurer will think it's obscure, but the safecracker will often quickly find a way to figure it out.
Not to think of even people relocating services to ports above 1024, so that if the daemon can somehow be stopped/crashed after a compromise, the attacker doesn't even need root to bind to the ports to 'emulate' the service and pickup lot of information from people/other systems that thought they were connecting to the legitimate service...
Security through obscurity is quickly a victim of security mistakes (plus the _false_ sense of security is has given the implementer).
If you take your computer security seriously, and actually have your tripwire databases on read-only media, then your standard intrusion recovery method would find all infected binaries, including the compiler, edited configuration files, infected kernel binaries, stealth modules (adore), etc etc.
Of course, if you just mess around and after a breakin assume that patching the backdoor is enough to clean your system... hmm, then I guess you're asking for problems.
A bunch of rootkit backdoors don't even use existing binaries, they just run their own login service, hidden by a special kernel module.
Biggest marker for LCDs: Laptops...
Power will be an issue.
That said, I think that the power consumption of this new type of CRT cannot be compared with the CRTs we know, because the distance to the phosporus is a lot smaller, hence the voltage drop between the anode and cathode will be less, hence the power consumption will be a lot less too. I think it has a good chance to beat LCDs, on preice and power consumption just because it doesn't need the expensive power-sucking flat white backlight.
Think about that.
I hope IBM can put these things on the market in 21" size soon (this year?).
You're right. It's just like Disney, but then for really rich people. So you're rich, have a bunch of houses, two big yachts, lots of fast cars, lots of old cars, all the toys you want and you give generously to charity, then you go to space. Why not?
All right, some might say that that money would better be spent giving more to charity. But I guess that's mostly because space travel is so very extravagant. But then again, wasn't flying and skiing too initially?
"Africans should concentrate on more important things, like food and education."
How blunt. Read again, it said "South African". Now read your history books. South Africa is not Ethiopia. South Africa is a country with it's own unique set of problems but very well a bunch of very strong (read: rich) assets.
It's like putting people from Brazil and the USA in the same 'box' just because they come from the same continent. They are a world apart.
All right, if you want others to do the coding for you, then you shouldn't rely upon the open source community. Then, choosing a commercial tool or choosing a company that provides support for an open source tool are your only options.
Might I ask what it is that you use ssh for that makes ssh such an important program for you that you need support, features, and easy compilaton?
SSH looks like a sufficiently functional secure telnet with port forwarding and secure file transfer to me. I've never found any additional features that I was missing in ssh, no need to touch it. Touching it will only add risk of new security problems.
I just do an easy apt-get for the linuxes, and compile once for each of the (very limited) Unix platforms that I need ssh on, then just distribute binaries. Re-installation is only necessary if there are security fixes.
But it doesn't have the main open source advantage: So you're dependent on the software (OS, duh), and it has a bug, you look at the source, you find the bug and make the fix. Then, if you're in a company you can't use the fixed software, you have to use the buggy original that Microsoft gave you.
The only advantage is that you can find the bug in the MS software easier (maybe) and design a workaround quicker... The advantage to Microsoft is clear: More software will work on what they are selling. The advantage to the user or non-Microsoft company is just not all what it could be. It's just a little improvement in the time it takes to find the workaround.
Slashdot Mirror
Tulip mania? What did I miss?
Sure, and we all know we never change our phone numbers.
I guess that when you migrate across borders, the local phone company will have to enter an exception in their routing tables?
IPV4 anyone? ring a bell?
All right, but we've had the a.out/ELF transition followed by the libc5/libc6/glibc2 transition. It kind of seems like there will be a new transition waiting when this one evens out.
"Eventually everyone will get their stuff straightened out"
/sceptic>
< sceptic>'Eventually'... When is that? While I'm still alive?<
That is correct.
(and that's why I bought an amp with digital input, plus a simple cheap little soundcard with the C-Media CMI8738 chip on it (SPDIF out). Now all I have to do is install the board and run the wire).
Now only remains this question: If it's better, then how do I upgrade my portable MP3/CD player to Ogg? Does AIWA make a CDC-Ogg yet like the CDC-MP3 they have for in the car? Damn.
btw, what is TMTOWTDI?
Will the ticket prices go down too?
I guess not...
"To get a valid comparison, rip the content of the CD as WAV. Then compare the WAV and the OGG, using the same soundcard and the same amp-input. Everything else is totally meaningless."
How 'bout this: rip the CD as WAV, Ogg and De-Ogg to WAV. Then write a couple of CDRs with all songs in the same order as the CD, but with some songs from the original WAV, and others the Ogged&De-Ogged WAV. Then do a blind classification trying to see if ou can tell if it's the original WAV or not by listening to the CDRs. If you are correct 50% of the time, then for you Ogg is lossless for your ears.
In my opinion that's the only way to really test. The lower quality of the soundcard output might hide differences in your suggestion. Plus the blind classification removes any psychological effects of expecting to hear differences or not.
It came back with
"Problem with bid amount
Your bid of $125.00 is less than the minimum bid. Please go back and bid at least $550,000.00. (Please type numbers and the decimal point (.) only.)"
"I actually agree with the concept setting up a lot of machines to reply to the virus with the fix."
The amusing part about that suggestion is that it is probably illegal to do so.
And if it were legal, then crashing the box might work better. Fixing it without the owner doing it will just lead to more lazyness on the part of the owner.
Don't worry, the S in SMTP stands for Simple. Making your own SMTP client is probably easier than figuring out how to use MAPI.
I don't think my mozilla supports that.
Yet another good reason not to use the client part of the Code-Red-virus-spreading-server.
This is yet another example of patents stifling application of innovations. Thanks to this we'll see less AC3 on open source OSs. Where else is Dolby's AC3 library for Linux & BSD?
Hmm, for ssh that is easy to implement. Just delete the entry from ~/.ssh/authorized_keys after each login.
"It is MUCH more cost effective o spend he extra money on the fastest memory, memory buses, disks and controllers you can get."
Often that is true, but it depends on the application and on the balance point. For example If you're making a aystem with 1GB RAM or more, then the balance is often different. $33 price difference will give you 5% in CPU speed. But the same amount of money spent on RAM over 1GB often won't even make a noticable difference. I'd consider it a waste to build a system of 1GB RAM or more and anything less than the 1.4GHz Athlon. Remember that a faster processor can handle interrupts and context switches at lower latency too, so even if your CPU load is low, latency (response time) will be better with a faster CPU.
And then there are a lot of uses for PCs that are still very much CPU bound (no I'm not talking about setiathome and distributed.net). We have systems with 512MB, no disk and they're 100% CPU bound every time we're waiting on it... We'd wish 25GHz CPUs actually existed yet.
Plus, what is $33 on a CPU compared to going from a 7200RPM UDMA100 to a 10 or 15KPRM SCSI? if you're so afraid to be slow, then spend that $33 on the CPU (and get lotsa RAM too, will often speed you up much more than that slow disk, even at 15KRPM).
Cache is good, but don't go too far hoping that that large L2 on a Xeon will save you in all cases. Sometimes cache is almost completely irrelevant. If you're walking large datasets, then you're completely trashing your cache anyway, then spending the money on getting faster RAM chips is wiser.
Back to my point: It all depends on the balance point and application. Personally, for my own PC use, I'd like it that everybody keeps demanding faster CPUs, because that would speed up development and make it available sooner. Maybe we'll see a split in the future where there are 'normal speed' and 'high speed' CPU's, where the speed difference is an order of magnitude or more. Hey, wasn't it the same in the old days of mainframes/mini's/mirco's versus the home computers?
Right on the mark.
Plus if you're running jobs that take many hours or sometimes days, then a 5% increase can mean making that deadline to get the bonus for the project, that would easily pay for the $33...
It's just a matter of where do you stop. What if there was a 1.5Ghz for $2500. Would that be worth it? Hmm, a couple of places probably can help you towards trying that too.
The article here was probably written for home users though. Even then it wasn't on the mark, because the main issue that matters for home users is game speed.
Why not solve your paranoia with one-time passwords? Then they might see what you're doing, but can't steal your logins.
And the great thing about is that when they find the safe, they can tinker in it on that tucked away little place where you hid it where nobody is looking anyway, so they won't get caught trying to break the safe.
Longer to find it, often not: You know how they would find the safe? They simply follow you when you go out to check on it. Bye Bye security, it helps nothing.
Often obscurity is not what it seems, the obscurer will think it's obscure, but the safecracker will often quickly find a way to figure it out.
Not to think of even people relocating services to ports above 1024, so that if the daemon can somehow be stopped/crashed after a compromise, the attacker doesn't even need root to bind to the ports to 'emulate' the service and pickup lot of information from people/other systems that thought they were connecting to the legitimate service...
Security through obscurity is quickly a victim of security mistakes (plus the _false_ sense of security is has given the implementer).
If you take your computer security seriously, and actually have your tripwire databases on read-only media, then your standard intrusion recovery method would find all infected binaries, including the compiler, edited configuration files, infected kernel binaries, stealth modules (adore), etc etc.
Of course, if you just mess around and after a breakin assume that patching the backdoor is enough to clean your system... hmm, then I guess you're asking for problems.
A bunch of rootkit backdoors don't even use existing binaries, they just run their own login service, hidden by a special kernel module.
Other tip besides tripwire: www.snort.org
Not the same continent? All right. But woth are in "America". A person from Brazil is as much an American as a person from the U.S.A.
Biggest marker for LCDs: Laptops... Power will be an issue. That said, I think that the power consumption of this new type of CRT cannot be compared with the CRTs we know, because the distance to the phosporus is a lot smaller, hence the voltage drop between the anode and cathode will be less, hence the power consumption will be a lot less too. I think it has a good chance to beat LCDs, on preice and power consumption just because it doesn't need the expensive power-sucking flat white backlight. Think about that. I hope IBM can put these things on the market in 21" size soon (this year?).
You're right. It's just like Disney, but then for really rich people. So you're rich, have a bunch of houses, two big yachts, lots of fast cars, lots of old cars, all the toys you want and you give generously to charity, then you go to space. Why not?
All right, some might say that that money would better be spent giving more to charity. But I guess that's mostly because space travel is so very extravagant. But then again, wasn't flying and skiing too initially?
"Africans should concentrate on more important things, like food and education."
How blunt. Read again, it said "South African". Now read your history books. South Africa is not Ethiopia. South Africa is a country with it's own unique set of problems but very well a bunch of very strong (read: rich) assets.
It's like putting people from Brazil and the USA in the same 'box' just because they come from the same continent. They are a world apart.
All right, if you want others to do the coding for you, then you shouldn't rely upon the open source community. Then, choosing a commercial tool or choosing a company that provides support for an open source tool are your only options.
Might I ask what it is that you use ssh for that makes ssh such an important program for you that you need support, features, and easy compilaton?
SSH looks like a sufficiently functional secure telnet with port forwarding and secure file transfer to me. I've never found any additional features that I was missing in ssh, no need to touch it. Touching it will only add risk of new security problems.
I just do an easy apt-get for the linuxes, and compile once for each of the (very limited) Unix platforms that I need ssh on, then just distribute binaries. Re-installation is only necessary if there are security fixes.
But it doesn't have the main open source advantage: So you're dependent on the software (OS, duh), and it has a bug, you look at the source, you find the bug and make the fix. Then, if you're in a company you can't use the fixed software, you have to use the buggy original that Microsoft gave you.
The only advantage is that you can find the bug in the MS software easier (maybe) and design a workaround quicker... The advantage to Microsoft is clear: More software will work on what they are selling. The advantage to the user or non-Microsoft company is just not all what it could be. It's just a little improvement in the time it takes to find the workaround.