In just 20 years or so it went from something you could only do with amazingly expensive special hardware to something that is easy for a $1000 computer to handle.
Audio is far from being the only areay where hardware acceleration became unnecessary and disappeared. Intel processors still have instructions for accelerating Cobol decimal number processing (AAA etc.).
20 years from now, may be the same deal with graphics.
No, it will take four years at most. I bet for 2012.
Except for Bonnie++, all of their benchmarks are compute-bound. In other words, they're benchmarking the bundled compiler, not the distribution.
The one exception is Bonnie++, on page 6, which measures raw filesystem performance... and is something that is known to greatly depend on how old and how full a given filesystem is.
The sun is the center of the universe? I though the sun orbited the Milkey Way Galaxy's central black hole?
A scientific theory isn't judged on whether it's ``true''; we leave the concept of ``truth'' to theologians, creationists and other amateurs.
A scientific theory is judged on how useful it is. What Copernicus showed is that by using a model in which the referential is attached to the sun, rather than the earth (as in the earlier Ptolemean model), many computations become easier.
Note that all of these models are useful under some circumstances. When you compute the distance from your home to the butcher's, you disregard the rotation of the earth, and hence use the Ptolemean model. When you compute the date of Easter next, you use the Copernican model. But if you need to compute the position of our Galaxy in a few billion years, you'll likely want a different model.
I think you mean the other way around; Cogent filtered AS1239 paths, blamed Sprint, and offered Sprint customers free circuits. Cogent customers were still visible via XO through Sprint during this mess, but Cogent filtered the traffic.
Right, I think I'm starting to understand. Correct me if I'm wrong.
Over one year ago, cogent got free peering from Sprint, and subsequently filtered out all alternative routes to Sprint. I read this as Cogent blackmailing Sprint: if you disconnect us, there will be no connectivity between your customers and ours.
Last week, Sprint decided not to go with Cogent's blackmail, and disconnected Cogent with no advance notice to their (Sprint's) customers.
Cogent had over a YEAR's notice that they would be cut off.
According to the article, Cogent have been stealing service from Sprint for over a year. However, there is no evidence that Sprint's customers have been informed of the plans to de-peer.
The press release indicates that Sprint have now re-established peering on a temporary basis. To me, that would indicate that Sprint's customers are not amused.
You say Sprint is wrong. Why should Cogent be allowed to steal service?
Why should Sprint be allowed to cut connectivity for their customers, with no advance notice, just because they are unable to resolve a disagreement they have with Cogent?
What proposed legislation has ever fixed an extant problem without making something else, and usually the original problem, worse?
In many European countries, legislation has forced the incumbent telecom operators to allow newcomers to provide the service they were refusing to provide themselves. Probably the most extreme example is that of France, where France Télécom refused to provide ADSL service in order to avoid competing with their profitable ADSL and T1 businesses.
My lay understanding of the situatin is that, once routing tables are changed to reflect the new topography, most everything goes back to normal
My just as lay understanding is that that would be the case were Sprint not actively filtering routes from Cogent.
In other words, assuming I understand the situation right, Sprint are taking active measures to make sure that routes that originate in the Cogent network never reach Sprint's customers.
There's something seriously wrong in a world in which any ISP is allowed to break connectivity for millions of customers without being taken to court for it.
Do we need regulation? As in, if you de-peer without giving 45 days advance notice to your customers and allowing them to cancel their contract at no cost during those 45 days, you get sued into oblivion?
They audit every line of code they ship, including the external stuff they don't write.
Bollocks.
They only audit the base install. The ports tree is almost completely unmaintained.
I am the author of some software that ships in a number of OS distributions. In September 2005, I found a serious security bug in my code. I immediately notified the project's mailing list as well as all of my downstream distributors I was aware of. Debian, SuSE, FreeBSD and others that I forget immediately released an update. OpenBSD left the old, buggy version in the ports tree for three months.
You joke, but what good is the desktop environment to me when I'm playing a game? I liked the days of DOS games much better.
How fast they forget...
Remember the joys of setting up your hardware in every single game? Running GAMECONFIG.EXE to say yes, my SoundBlaster is on IRQ 7, my display can handle 1024x769 in 256 colours, and no, I don't have an AdLib card.
Having a real OS might shave off a few fps, but it allows you to set up your hardware just the once, and have it work in all of your software.
I've browsed through the WiMax standards, and they almost make ATM look elegant. A connection-base technology with no less than three incompatible encapsulations. Disconnected operation is simulated by establishing connections to a back-end server and running bridging software there.
I'm looking forward to the day when multiple implementations of WiMax are available and the interoperability issues start showing up...
Note that DNSCurve and DNSSEC solve different problems. DNSCurve secures that communication channel, where DNSSEC secures the actual data.
I'm definitely no security specialist, but I would reckon that securing the data is more useful, while securing the channel is easier, and hence has a greater chance of actually being deployed.
Compare this to the e-mail situation. PGP secures the actual mail message, while SMTP over SSL secures the communication channel. SMTP over SSL is widely deployed, while PGP is unknown outside of some geek circles.
But the conclusion does bother me: We're basically saying that all software is buggy,
No. What we're saying is that we build layered systems, and that every layer is expected to protect its integrity from the higher layers.
The hardware protects itself from software (no brain-damaged hardware interfaces), the kernel protects itself from userspace (priviledged vs. unpriviledged mode), system userspace protects itself from user userspace (root vs. non-root), userspace protects itself from interpreted network code (sandboxing).
If you read the announcement, you'll find that the major benefit of using a layered architecture such as IP is the ability to use different physical media, depending on the application.
For short-range applications, there are technologies that are both cheaper and more power-efficient than wifi. Off the top of my head, there's Zigbee, Bluetooth and probably lots of others.
And for even lower range applications, nothing beats the cost of an infra-red diode.
don't write your own license, use a well-known license that people already understand;
don't include an advertising clause.
You may be able to convince your supervisor by citing the examples of BSD Unix and X11, which brought fame and money to their creators (the CSG at Berkeley, and project Athena at MIT) while using extremely liberal licenses -- the MIT/X11 license (which is what I use for my research) and the 4-clause BSD license, albeit with the advertising clause not being enforced.
You may also want to cite the following anegdote. Two years ago, I was compiling a Linux LiveCD for our first, second and third year undergrads. One of the pieces of software I wanted to include was a Prolog compiler from a well-known Portuguese university which we use in third-year courses.
Unfortunately, the Prolog implementation was covered by a fairly strict license that would significantly complicate our distribution process. After a few exchanges of e-mail with the copyright holders, they told us that we were welcome to do whatever we wanted, but they'd not change the license for us.
After consulting with our legal department, we decided we could not include the Prolog compiler.
There should be a planned algorithm that kicks in,
This assumes that we know when, in the future, we'll need to insert leap seconds. And we don't.
Leap seconds are introduced in order to compensate for medium-term variations in the earth's rotation speed. We don't have a good understanding of the way the earth rotates -- knowing what UTC time it will be in ten years' time is about as difficult as predicting the weather for next week-end.
Slashdot Mirror
In just 20 years or so it went from something you could only do with amazingly expensive special hardware to something that is easy for a $1000 computer to handle.
Audio is far from being the only areay where hardware acceleration became unnecessary and disappeared. Intel processors still have instructions for accelerating Cobol decimal number processing (AAA etc.).
20 years from now, may be the same deal with graphics.
No, it will take four years at most. I bet for 2012.
Except for Bonnie++, all of their benchmarks are compute-bound. In other words, they're benchmarking the bundled compiler, not the distribution.
The one exception is Bonnie++, on page 6, which measures raw filesystem performance... and is something that is known to greatly depend on how old and how full a given filesystem is.
The sun is the center of the universe? I though the sun orbited the Milkey Way Galaxy's central black hole?
A scientific theory isn't judged on whether it's ``true''; we leave the concept of ``truth'' to theologians, creationists and other amateurs.
A scientific theory is judged on how useful it is. What Copernicus showed is that by using a model in which the referential is attached to the sun, rather than the earth (as in the earlier Ptolemean model), many computations become easier.
Note that all of these models are useful under some circumstances. When you compute the distance from your home to the butcher's, you disregard the rotation of the earth, and hence use the Ptolemean model. When you compute the date of Easter next, you use the Copernican model. But if you need to compute the position of our Galaxy in a few billion years, you'll likely want a different model.
will fans just ignore it
No. They'll start explaining why it's actually an advantage for the user.
I think you mean the other way around; Cogent filtered AS1239 paths, blamed Sprint, and offered Sprint customers free circuits. Cogent customers were still visible via XO through Sprint during this mess, but Cogent filtered the traffic.
Right, I think I'm starting to understand. Correct me if I'm wrong.
Over one year ago, cogent got free peering from Sprint, and subsequently filtered out all alternative routes to Sprint. I read this as Cogent blackmailing Sprint: if you disconnect us, there will be no connectivity between your customers and ours.
Last week, Sprint decided not to go with Cogent's blackmail, and disconnected Cogent with no advance notice to their (Sprint's) customers.
Umm... RTFA,
I have. They need to fix their <title>.
Cogent had over a YEAR's notice that they would be cut off.
According to the article, Cogent have been stealing service from Sprint for over a year. However, there is no evidence that Sprint's customers have been informed of the plans to de-peer.
The press release indicates that Sprint have now re-established peering on a temporary basis. To me, that would indicate that Sprint's customers are not amused.
You say Sprint is wrong. Why should Cogent be allowed to steal service?
Why should Sprint be allowed to cut connectivity for their customers, with no advance notice, just because they are unable to resolve a disagreement they have with Cogent?
What proposed legislation has ever fixed an extant problem without making something else, and usually the original problem, worse?
In many European countries, legislation has forced the incumbent telecom operators to allow newcomers to provide the service they were refusing to provide themselves. Probably the most extreme example is that of France, where France Télécom refused to provide ADSL service in order to avoid competing with their profitable ADSL and T1 businesses.
My lay understanding of the situatin is that, once routing tables are changed to reflect the new topography, most everything goes back to normal
My just as lay understanding is that that would be the case were Sprint not actively filtering routes from Cogent.
In other words, assuming I understand the situation right, Sprint are taking active measures to make sure that routes that originate in the Cogent network never reach Sprint's customers.
Whatever.
There's something seriously wrong in a world in which any ISP is allowed to break connectivity for millions of customers without being taken to court for it.
Do we need regulation? As in, if you de-peer without giving 45 days advance notice to your customers and allowing them to cancel their contract at no cost during those 45 days, you get sued into oblivion?
They audit every line of code they ship, including the external stuff they don't write.
Bollocks.
They only audit the base install. The ports tree is almost completely unmaintained.
I am the author of some software that ships in a number of OS distributions. In September 2005, I found a serious security bug in my code. I immediately notified the project's mailing list as well as all of my downstream distributors I was aware of. Debian, SuSE, FreeBSD and others that I forget immediately released an update. OpenBSD left the old, buggy version in the ports tree for three months.
You joke, but what good is the desktop environment to me when I'm playing a game? I liked the days of DOS games much better.
How fast they forget...
Remember the joys of setting up your hardware in every single game? Running GAMECONFIG.EXE to say yes, my SoundBlaster is on IRQ 7, my display can handle 1024x769 in 256 colours, and no, I don't have an AdLib card.
Having a real OS might shave off a few fps, but it allows you to set up your hardware just the once, and have it work in all of your software.
Ahem, no.
There's more than one programming language called D.
There's Digital Mars D, which is what you describe. And there's Microsoft D, which is almost, but not quite, completely unlike Digital Mars D.
I've browsed through the WiMax standards, and they almost make ATM look elegant. A connection-base technology with no less than three incompatible encapsulations. Disconnected operation is simulated by establishing connections to a back-end server and running bridging software there.
I'm looking forward to the day when multiple implementations of WiMax are available and the interoperability issues start showing up...
dnscurve
Note that DNSCurve and DNSSEC solve different problems. DNSCurve secures that communication channel, where DNSSEC secures the actual data.
I'm definitely no security specialist, but I would reckon that securing the data is more useful, while securing the channel is easier, and hence has a greater chance of actually being deployed.
Compare this to the e-mail situation. PGP secures the actual mail message, while SMTP over SSL secures the communication channel. SMTP over SSL is widely deployed, while PGP is unknown outside of some geek circles.
\
But the conclusion does bother me: We're basically saying that all software is buggy,
No. What we're saying is that we build layered systems, and that every layer is expected to protect its integrity from the higher layers.
The hardware protects itself from software (no brain-damaged hardware interfaces), the kernel protects itself from userspace (priviledged vs. unpriviledged mode), system userspace protects itself from user userspace (root vs. non-root), userspace protects itself from interpreted network code (sandboxing).
Delete those annoying REM statements
It was difficult to write, there's no reason it should be easy to read.
making cheap Wi-fi chipsets is the answer here
If you read the announcement, you'll find that the major benefit of using a layered architecture such as IP is the ability to use different physical media, depending on the application.
For short-range applications, there are technologies that are both cheaper and more power-efficient than wifi. Off the top of my head, there's Zigbee, Bluetooth and probably lots of others.
And for even lower range applications, nothing beats the cost of an infra-red diode.
Two points to keep in mind:
You may be able to convince your supervisor by citing the examples of BSD Unix and X11, which brought fame and money to their creators (the CSG at Berkeley, and project Athena at MIT) while using extremely liberal licenses -- the MIT/X11 license (which is what I use for my research) and the 4-clause BSD license, albeit with the advertising clause not being enforced.
You may also want to cite the following anegdote. Two years ago, I was compiling a Linux LiveCD for our first, second and third year undergrads. One of the pieces of software I wanted to include was a Prolog compiler from a well-known Portuguese university which we use in third-year courses.
Unfortunately, the Prolog implementation was covered by a fairly strict license that would significantly complicate our distribution process. After a few exchanges of e-mail with the copyright holders, they told us that we were welcome to do whatever we wanted, but they'd not change the license for us.
After consulting with our legal department, we decided we could not include the Prolog compiler.
[Creationism] is specific to the judeochristian religion.
No, it isn't. There's Islamic Creationism, Hindu Creationism, ancient Greek Pagan Creationism, and doubtless all sorts of other brands of Creationism.
Many other religions believe that the universe was created in a different way.
Yes, but their claims still constitute forms of Creationism.
A proper Compared pseudo-science class should cover a number of different brands of creationism, and show how they contradict each other.
Many of those laptops that can be configured have "FreeDos" as an option for the OS.
The same is true of Dell, but in many cases you won't find that option on the website -- you need to ask your friendly Dell representative.
I am the proud owner of a Dell-branded FreeDOS CD.
What directory layout is that? GoboLinux?
Mac OS X. You may have hear about it, it's a hybrid of CMU Mach and FreeBSD recombined with NeXTstep.
Nah. It just gets bitchy. "Blah blah has modification time in the future!"
Now, an NFS server that's running SLOW, OTOH....
Yep, you're right, I got confused.
I don't give a shit about 10 minutes either way.
It depends on the application. Having one's NFS file server just a second fast will break most Makefiles.
There should be a planned algorithm that kicks in,
This assumes that we know when, in the future, we'll need to insert leap seconds. And we don't.
Leap seconds are introduced in order to compensate for medium-term variations in the earth's rotation speed. We don't have a good understanding of the way the earth rotates -- knowing what UTC time it will be in ten years' time is about as difficult as predicting the weather for next week-end.