Well, let's be a little fair here. Yes, the autobahn is way better than the interstate system. But the U.S. has a lot more ground to cover. There's no excuse for not having better driver education and intelligent traffic laws (and enforcement), though.
The way they reported the results was pretty bad. The left two columns ranked products by false positives, while the right two ranked products by spam caught. It is very difficult to look at this table and get a sense of which products performed the best.
For example, the top product for false positives, BorderWare at 0.04% looks very impressive until you look at the other column and see that it only caught 88%. It's easy to have a low false positive rate when your catch rate is low, too.
At minimum, they should have taken the false positive rate, added it to the percent missed and ranked by that.
Doing so sends BorderWare into the middle of the pack where it belongs, and more likely winners rise to the top. (Postini and MailFrontier).
Pretty shoddy reporting when the end reader has to take your numbers and plug them into a spreadsheet to make any sense out of them.
They could have also weighted the two error rates, but deciding on weights would be pretty subjective. Some might think false positives should be weighted higher, while others might think the opposite. Ranking them without weights would have been an acceptable compromise.
PBParts.com is awesome. About a year and a half ago, my two year old son managed to spend about half an hour unsupervised with my brand new Titanium Powerbook. He popped every single key off the keyboard.
While we managed to replace most of the keys, three of them had damaged scissor mechanisms (little, white plastic pieces that the keycap snaps onto). Both the local Apple dealer, and Apple tech support's answer: replace the keyboard. Do it yourself for about $150 (I don't remember the exact price). Not bad for a brand new laptop that cost over $3000, but still ridiculous when all I wanted was a few tiny pieces of plastic.
Fortunately I decided to ask Google before buying a new keyboard, and found pbparts. They sell just the scissor mechansims for $5 each. My bill: $15 + shipping.
Since Apple doesn't sell these pieces separately, I can only assume that pbparts has people dismantling keyboards in order to sell them one piece at a time.
A lot of hospitals still require people to turn off cell phones, as they allegedly might have an effect on other equipment. Talk about valuing your own life over others...
Ironically enough, one of the problems they can cause is noise in ECG traces.
A movie is art. Offensive or not it is art and it is the vision of the director. Just like real art can not be sensored either should movies.
Okay, it was probably flamebait, but I'll bite.
You're an idiot. If I own a book, it is my right to tear out pages. If I own a photograph, it is my right to draw all over it with permanent marker. If I own a sculpture, it is my right to break it into a thousand pieces with a baseball bat. Artistic vision is irrelevant. No law or precedent gives someone the right to tell me what I can and can't do with my personal property just because it interferes with someone's "creative vision."
People keep dismissing Maildirs because of an erroneous notion that performance (speed and/or inode usage) is somehow the most important consideration. It isn't. Let's look at what the Maildir format gives you.
Maildirs are:
Crash proof: an interrupted delivery cannot cause folder-wide corruption or the delivery of an incomplete message.
Lockless: all Maildir operations (deliver, delete, read, etc) can be performed simultaneously by multiple processes on multiple machines without the need for any sort of file locking.
What does this mean? Reliability. That's why you use Maildirs. It may be slightly slower than some other formats (although I've never noticed a difference) and it certainly consumes more inodes. But it's way more reliable. You never have to worry about someone's mail program crashing and leaving the mail folder in an inconsistent state. Maildirs don't have an inconsistent state. And when you're delivering over NFS you don't have to worry about whether or not file locking is going to work right. Maildir's don't need locking.
In email, reliability is everything. People may grumble a bit if they think their email isn't arriving fast enough. No big deal. But there is nothing more terrifying than a user with corrupted or (gasp) missing email. While using Maildirs won't solve all of your email problems, they are definitely a step in the right direction.
Okay, call me a complete nerd, but the most powerful language IMHO is Scheme. Sure it's full of lots of silly parentheses, but it offers language features you won't find anywhere else (except another functional language, like say Lisp).
What sort of features? How about macros (we're talking R5RS here)? Macros allow you to define new language syntax. For example, if macros were available in C (no, #define doesn't even come close), you could use it to add new language features. Like say, a new type of switch statement that didn't require you to end each case with break.
But what about all the nice features you might find in more conventional languages? You want a nice IDE you say? A GUI toolkit? Objects? You want DrScheme. It offers everything more mainstream languages might offer and more.
Okay, Scheme and it's other Lispy friends might not be the right choice for you, but they're definately worth looking at.
The two janitors were highly annoying. They reminded me of characters from another movie/show that I cannot think of right now. They should have been dropped.
They're both a ripoff of one Simpsons character--the pizza-faced, voice cracking teenager usually used as a fast food employee.
Most autodialers are devices with one or more PRI's coming into them. What's a PRI you ask? It's a 24-channel ISDN line. Think of it as an ISDN T1.
Your typical PRI will have 23 B-channels and 1 D-channel. Each B-channel carries one phone call. The D-channel is the interesting part. It carries signals back and forth about the B-channels.
When the dialer places a call, it sends a signal down the D-channel which says, in effect, call this number on this B-channel. If the dialed number is located in an area that supports ISDN (most places in North America), the dialer will receive status information about the call back across the D-channel, like call complete, busy, out of service, go to Hell, etc.
This is out-of-band signaling. As long as the dialer is getting call status information across the D-channel, it's never even going to listen to the tones that might be sent in-band. The only time this might work is if you live in an area that doesn't support ISDN. In such a case, the autodialer will be forced to use hardware DSP's to decode in-band tones, and it might be fooled if you send it some of your own.
Look at the facts. From a programmer's perspective, Windows is a massive, bizarre and utterly incomprehensible operating system. It has been so carelessly hacked over it's (far too lengthy lifetime) that it has taken on a life of its own. It is no longer understood by its developers, or anyone else. In short, it's already alien.
Plus an advanced alien civilization bent on world domination would probably have a lot of respect for a company like Microsoft. And would be unwilling to release it's technology without an equally advanced and world-dominating NDA.
I switched to FreeBSD two years ago from Linux and have never regretted it. A great deal of convenience can be found in running the same OS on your desktop that you use on your servers. Especially when that OS is as stable and powerful as FreeBSD.
FreeBSD supports a pretty decent array of sound hardware and desktop apps--including the ability to run Linux binaries. It runs Gnome and KDE. There's no reason it shouldn't be valid desktop choice.
The speed differences between various languages is not so important as the speed differences between various technologies. CGI for example is going to be slower in any language and on any web server, than using some sort of in-process API.
Choose a fast web server that allows you to embed scripts inside the server process. Apache gives you lots of choices in this area: mod_perl, mod_python (and mod_snake), PHP or even C. But Apache isn't exactly the fastest web server on the face of the earth (but it should be fast enough). AOLServer is very fast and lets you use Tcl or C (PHP works too, I believe).
The point is that you should use the language you're familiar with and spend more time worrying about the environment the software will execute in. That's where you're going to get the most performance improvements.
For some reason this reminds me of the song, "I know an old lady who swallowed a fly." Just replace fly with moth, change spider to killer bat, and build from there. Needless to say, releasing genetically engineered insects into the wild makes me more than a little nervous.
At the hardware level, PCI compliant is PCI compliant. If a Sun box has a PCI bus then you can plug any PCI card into it. The trick is getting drivers. If you're running Solaris, this means that you're pretty much stuck with PCI cards that say they're for a Sun. I have not seen a lot of PCI video cards marketed to PC users that include Solaris drivers.
Things are cheerier if you run Linux. According to the faq a PCI card will generally work in the UltraSparc version if it works in the i386 version. But that isn't a guarantee. Sometimes a driver writer makes assumptions about the underlying platform that don't hold for other platforms.
You'll probably find that the biggest problem isn't individual PCI cards, but rather getting Linux to run on the thing at all. Although at under $1k this shouldn't be a problem for too long--assuming the Blade architecture is close to other UltraSparc machines, anyway.
I use, in X speak, -misc-fixed-medium-r-semicondensed--13-120-75-75-c -60-iso8859-1 which is aliased in XFree86 as just plain "fixed." It's a clean monospaced font that's easy to read at 1024x768.
I've been using dbjdns for about six months (since before it was called djbdns). I use it to provide primary and secondary nameservers for several thousand domains. It runs flawlessly.
Before djbdns I used Bind. What a piece of garbage. It's slow and bloated. Writing scripts to deal with its data files is a nightmare. And the code is so piss-poor people can't help but find security holes. This is not the first hole in bind and it certainly won't be the last.
Yes, soft updates are cool. Even better, now McKusick is working on snapshots. Snapshots allow you to get a consistent picture of the entire filesystem at one instant in time. This is tremendously useful for backups.
Once snapshots are ready, waiting for fsck on an FFS filesystem will become unnecessary. If a filesystem is dirty, on bootup the system will create a snapshot, and run fsck on that snapshot in the background.
Unlike many technology-related things, the term "latest and greatest," doesn't really apply to security. "Tried and true" is almost always the way to go.
The AES candidates are all relatively new and untested. There may be insecurities that nobody's discovered. Older algorithms, like DES or Blowfish, which have withstood extensive cryptanalysis over the course of many years, tend to be better choices.
Attempting to protect free software using patent law is hypocritical and weakens the free software community's ability to oppose them. It's a bit like telling your little brother to stay out of the cookie jar when you have a mouth full of cookies.
Using patents to fight fire with fire isn't such a hot idea, either. Patent law is a game where the rich invariably come out on top. Big business wins two ways. It has the resources necessary to defend it's own patents, while infringing on a free software project's patents with impunity. Litigation is for big corporations with big legal budgets. It's unlikely that an open-source project would have the resources to properly defend its patents.
The only way to convince Sound Blaster to change its policy is to make a dent in its bottom line. Sure, there are drivers for Linux, but what about the *BSD's?
Do the whole open-source community a favor:
Buy a different card from a friendlier company.
Don't buy an SB Live.
Send Sound Blaster some politefeedback. Tell them you considered buying an SB Live, but decided against it because of the lack of open-source drivers and programming specs. Tell them what sound card you did buy.
If enough people send polite, intelligent feedback, Sound Blaster will have little choice but to release the specs. Until then, please don't buy an SB Live!
Slashdot Mirror
Well, let's be a little fair here. Yes, the autobahn is way better than the interstate system. But the U.S. has a lot more ground to cover. There's no excuse for not having better driver education and intelligent traffic laws (and enforcement), though.
Who cares what it looks like. Is it fun?
At minimum, they should have taken the false positive rate, added it to the percent missed and ranked by that. Doing so sends BorderWare into the middle of the pack where it belongs, and more likely winners rise to the top. (Postini and MailFrontier). Pretty shoddy reporting when the end reader has to take your numbers and plug them into a spreadsheet to make any sense out of them.
They could have also weighted the two error rates, but deciding on weights would be pretty subjective. Some might think false positives should be weighted higher, while others might think the opposite. Ranking them without weights would have been an acceptable compromise.
While we managed to replace most of the keys, three of them had damaged scissor mechanisms (little, white plastic pieces that the keycap snaps onto). Both the local Apple dealer, and Apple tech support's answer: replace the keyboard. Do it yourself for about $150 (I don't remember the exact price). Not bad for a brand new laptop that cost over $3000, but still ridiculous when all I wanted was a few tiny pieces of plastic.
Fortunately I decided to ask Google before buying a new keyboard, and found pbparts. They sell just the scissor mechansims for $5 each. My bill: $15 + shipping.
Since Apple doesn't sell these pieces separately, I can only assume that pbparts has people dismantling keyboards in order to sell them one piece at a time.
http://www.cs.utexas.edu/users/EWD/ewd03xx/EWD340. PDF
Not only have they been slashdotted, it's been done through a link that says, "denial of service."
Ironically enough, one of the problems they can cause is noise in ECG traces.
Okay, it was probably flamebait, but I'll bite.
You're an idiot. If I own a book, it is my right to tear out pages. If I own a photograph, it is my right to draw all over it with permanent marker. If I own a sculpture, it is my right to break it into a thousand pieces with a baseball bat. Artistic vision is irrelevant. No law or precedent gives someone the right to tell me what I can and can't do with my personal property just because it interferes with someone's "creative vision."
Maildirs are:
- Crash proof: an interrupted delivery cannot cause folder-wide corruption or the delivery of an incomplete message.
- Lockless: all Maildir operations (deliver, delete, read, etc) can be performed simultaneously by multiple processes on multiple machines without the need for any sort of file locking.
What does this mean? Reliability. That's why you use Maildirs. It may be slightly slower than some other formats (although I've never noticed a difference) and it certainly consumes more inodes. But it's way more reliable. You never have to worry about someone's mail program crashing and leaving the mail folder in an inconsistent state. Maildirs don't have an inconsistent state. And when you're delivering over NFS you don't have to worry about whether or not file locking is going to work right. Maildir's don't need locking.In email, reliability is everything. People may grumble a bit if they think their email isn't arriving fast enough. No big deal. But there is nothing more terrifying than a user with corrupted or (gasp) missing email. While using Maildirs won't solve all of your email problems, they are definitely a step in the right direction.
What sort of features? How about macros (we're talking R5RS here)? Macros allow you to define new language syntax. For example, if macros were available in C (no, #define doesn't even come close), you could use it to add new language features. Like say, a new type of switch statement that didn't require you to end each case with break.
But what about all the nice features you might find in more conventional languages? You want a nice IDE you say? A GUI toolkit? Objects? You want DrScheme. It offers everything more mainstream languages might offer and more.
Okay, Scheme and it's other Lispy friends might not be the right choice for you, but they're definately worth looking at.
Take a look at the jail(8) and jail(2) manapges.
They're both a ripoff of one Simpsons character--the pizza-faced, voice cracking teenager usually used as a fast food employee.
Your typical PRI will have 23 B-channels and 1 D-channel. Each B-channel carries one phone call. The D-channel is the interesting part. It carries signals back and forth about the B-channels.
When the dialer places a call, it sends a signal down the D-channel which says, in effect, call this number on this B-channel. If the dialed number is located in an area that supports ISDN (most places in North America), the dialer will receive status information about the call back across the D-channel, like call complete, busy, out of service, go to Hell, etc.
This is out-of-band signaling. As long as the dialer is getting call status information across the D-channel, it's never even going to listen to the tones that might be sent in-band. The only time this might work is if you live in an area that doesn't support ISDN. In such a case, the autodialer will be forced to use hardware DSP's to decode in-band tones, and it might be fooled if you send it some of your own.
Plus an advanced alien civilization bent on world domination would probably have a lot of respect for a company like Microsoft. And would be unwilling to release it's technology without an equally advanced and world-dominating NDA.
FreeBSD supports a pretty decent array of sound hardware and desktop apps--including the ability to run Linux binaries. It runs Gnome and KDE. There's no reason it shouldn't be valid desktop choice.
Choose a fast web server that allows you to embed scripts inside the server process. Apache gives you lots of choices in this area: mod_perl, mod_python (and mod_snake), PHP or even C. But Apache isn't exactly the fastest web server on the face of the earth (but it should be fast enough). AOLServer is very fast and lets you use Tcl or C (PHP works too, I believe).
The point is that you should use the language you're familiar with and spend more time worrying about the environment the software will execute in. That's where you're going to get the most performance improvements.
For some reason this reminds me of the song, "I know an old lady who swallowed a fly." Just replace fly with moth, change spider to killer bat, and build from there. Needless to say, releasing genetically engineered insects into the wild makes me more than a little nervous.
Things are cheerier if you run Linux. According to the faq a PCI card will generally work in the UltraSparc version if it works in the i386 version. But that isn't a guarantee. Sometimes a driver writer makes assumptions about the underlying platform that don't hold for other platforms.
You'll probably find that the biggest problem isn't individual PCI cards, but rather getting Linux to run on the thing at all. Although at under $1k this shouldn't be a problem for too long--assuming the Blade architecture is close to other UltraSparc machines, anyway.
It would also be worth looking at NetBSD's UltraSparc port.
I use, in X speak, -misc-fixed-medium-r-semicondensed--13-120-75-75-c -60-iso8859-1 which is aliased in XFree86 as just plain "fixed." It's a clean monospaced font that's easy to read at 1024x768.
Before djbdns I used Bind. What a piece of garbage. It's slow and bloated. Writing scripts to deal with its data files is a nightmare. And the code is so piss-poor people can't help but find security holes. This is not the first hole in bind and it certainly won't be the last.
Once snapshots are ready, waiting for fsck on an FFS filesystem will become unnecessary. If a filesystem is dirty, on bootup the system will create a snapshot, and run fsck on that snapshot in the background.
The AES candidates are all relatively new and untested. There may be insecurities that nobody's discovered. Older algorithms, like DES or Blowfish, which have withstood extensive cryptanalysis over the course of many years, tend to be better choices.
Pat Ekman
NPS Internet Solutions
Using patents to fight fire with fire isn't such a hot idea, either. Patent law is a game where the rich invariably come out on top. Big business wins two ways. It has the resources necessary to defend it's own patents, while infringing on a free software project's patents with impunity. Litigation is for big corporations with big legal budgets. It's unlikely that an open-source project would have the resources to properly defend its patents.
Here's the OpenSSH advisory on the subject.
Do the whole open-source community a favor:
If enough people send polite, intelligent feedback, Sound Blaster will have little choice but to release the specs. Until then, please don't buy an SB Live!