Intel's compiler will work for some but not all of your software. You probably won't get away with compiling the kernel and glibc with it. I understand that there has been progress lately in increasing the compatibility of the Intel compiler and gcc.
The main reason Linux programs are compiled with gcc is that code portability is valued more than it is in the Wintel world. For instance, I have virtually identical user environments on both my home P4 machine and my Powerbook at work. GCC has to produce reasonable code for at least 12 (maybe more) architectures. Intel and MS can optimize the hell out of the IA32/64 only cases. There is ongoing work to re-engineer some parts of GCC that go clear back to the mid eighties. But I seriously doubt a cross compiler will ever produce code as tight for a given arch as a compiler made specifically for that arch. Some code in the compiler is common to all arches. No x86-only optimizations will be made there. There are other parts that will be unique to each arch and I presume they will be doing what they can there.
Even Apple and BSDs are using GCC. It is arguably to most important piece of software in the FOSS world. Thousands of projects have to compile to working binaries on cisc processors, 32/64/higher bits, and big and little endian machines. It is an extremely tough thing to get right and it can't be all things on all arches.
I'm not sure I buy that. At most you are casting small RF shadow on the ground. Unless that shadow falls across a legit subscriber's antenna then it won't do jack to anyone else. That's like saying when I look up at a spotlight that I'm stopping someone else from seeing it.
The "technician" the cable company sent out didn't understand why I wanted to use "all those wires" to put baseband video and audio into our stereo. The coax was "just one wire and easier". True it was marginally easier but I don't see much hope for selling even medium end stuff when the guy who hooks cable up all day doesn't know the least thing about av gear.
I've done things like put "This is really a Gecko browser. www.w3c.org hint hint." in the agent string. As long as that magic "IE" is in there, it'll usually let you in. If not, I'll hit the site a few times to get it in their logs and then change to IE on XP agent string. The User Agent Switcher extension is very handy for this sort of thing.
Now if we could just do something about javascript based detection techniques.....
Even if it were suddenly 50/50 overnight, or 75/25 in favor of FF, sites would still make IE-compliant pages, because nobody can affort to write off 25% of the market.
I'm getting deja vu here. Where have I seen this happen before. Holy shit! Netscape 4 has risen from the dead and walks the earth. IE now is like NS4 then. Various odd things happen with the cache in SSL mode. Alpha channel and CSS support are completely broken. Now everyone has to write ugly hacks to support IE once the standards compliant site is done. What a piece of shit!
So what you're saying is that there is no use mucking about with all these reflectors, LEDs and computers. Just paint the guy pink and turn on a cheap and simple Somebody Else's Problem Field.
Sure AAC is better, sure Ogg is better, but for most folks, even those with huge music collections and very exacting preferences in their audio systems, MP3 is still good enough. Why? Because most people care about the music, not the technology.... Beta was better than VHS but VHS won too.
That's pretty much apples and oranges though. There is a vast difference in the difficulty required to support a codec versus a physical format. Given sufficiently powerful and flexible playback devices, supporting new codecs is trivial. If they can't already, upcoming portables will be able to support any arbitrary codec with a firmware update.
What is needed are portable, easily hackable, and powerful devices. This makes the bitstream format entirely a matter of consumer choice. Of course such devices will support mp3, wma, and maybe even aac by default. Moore's law applies to the portables as well which means such devices are around the corner. Intrepid PDA owners have already had this for awhile.
This brings up a good point; how do you explain to someone who is NOT a geek (and has no interest in being one) about what DRM is and how it will effect them?
By vastly oversimplifying of course! Just tell people that it will almost always make it impossible "to record". When it isn't impossible, it will be very complicated and you'll need a geek such as yourself to make it work.
I'd just go to the bad side and be done with it. I've coded some small things in php because we needed them and weren't going to get them any other way. I pretty much learned what html and SQL I know at the same time. I have no trouble admitting this because I'm an admin who occaisionally codes rather than the other way around.
I've tried to avoid pulling obvious boners but I'm glad none of the machines running these things are exposed to the public net.
Why would it take a long time if X.org is just a fork from XFree?
Debian does a lot patching, cleaning up, and build script twiddling to make their packages run on all architectures. This only recently happened for XFree86 4.3. When they transition to a 4.4 base, they'll go with X.Org same as everybody else.
Oh, sorry, I forgot how NTFS sometimes decides to get bitchy and doesn't let you resize it...
Be sure to defrag it first. I think the gremlins that live in NTFS partitions are happier when they are snuggly wrapped in well packed blocks. Seriously, shut down as many processes as possible before you do defrag and then shut down Windows as quickly as possible when you're done. That procedure helps me more than anything else.
Much like the upcoming Reiser4, I think NTFS isn't sure if it wants to be an object database or a filesystem when it grows up. It sure seems "bitchy" for third party tools to work with though.
Given enough budget, you would be hard pressed to see that I even have a computer. Think stereo systems built into a wall. The best setup I ever saw had the tower on a shelf behind a utility door. It could be as pretty or ugly and not matter, because it's just something else to clutter the decor. Some people put neon in their cases for crying out loud. You can even get them prebuilt that way. Ideally, the only things I want visible are an unobtrusive screen, mouse, keyboard. Gee-whiz industrial design is completely wasted when those are the criteria.
Without an unlimited budget, I hide it under the desk as I said. It doesn't matter what it looks like.
By contrast, there have been around, what, five? standard PC motherboard sizes since the 386. Commodity parts are great for end user prices, but commodity means "same", and it shows in the final product.
Maybe so but why should I care? The tower is something goes under a desk or behind something. Its a computer not an objet de art or something. LCDs are skinny things that should fit on a desk and not draw attention to themselves. Sorry Mr. Industrial Designer but I would rather not see it at all rather than be wowed by it.
If Linux is being worked upon by people whose idea of vision is rhetoric such as "destroy Windows within 10 years" and "Longhorn will be a disaster", then there is no end in sight to the hegemony of Microsoft.
Very little code is being written with the explicit purpose of "destroying MS". Come to think of it, I don't know of any. I see two main reasons the free UNIXes are being developed. One, there is money to made by carving out a piece of the market from established competitor. After all, if MS is making a pile of cash why not siphon off a bit for yourself. This pretty much applies to FOSS pure plays and larger entities like IBM and Novell. Note well that the aim is isn't "destroy MS!", it's "Let's make some money!".
The other reason is what gets called "scratching the itch". There is software available for the need but it costs too much or the vendor can't be trusted. That last is more and more true of MS everyday. Or there isn't anything that precisely addresses the need so you write it yourself. That was certainly true of a few trivial things I've done in PHP.
Culture jam Gi-Joe/Barbie
on
Old Toy Modding?
·
· Score: 3, Funny
In the early nineties, Mattel had an infamous Barbie doll that told girls gems like "Math is hard...giggle.." Some activists broke into a warehouse and swapped boards between the GI-Joes and Barbies so that the Barbies were saying things like "Stop Cobra!".
Oh, and my wife is proof positive it isn't tough to NOT get infected.. 3 years on her own Win98 box and zero infections with no anti-virus suite running on the machine at all.
Then how do you know there are no viruses on the machine? Malware doesn't have to be obvious when it's running.
We're talking about the people who compared VCRs to the Boston Strangler. We're talking about the same industry that has had fear and loathing of every consumer technology since the first wax cylinders. Video rental could have been a reality in the early seventies. They were worried about repeat viewings so the tapes we're gimmicked so they could only be rewound at the video store. "But what if there are more people at the home than the person who rented it?"
You propose reasonableness. The problem is that the industry has been stupid and paranoid for over 90 years and they aren't going to stop now. If there is to be dialog and respect then they have to make the first move.
Lets assume they come up with a watermark that really works. You can't remove it and you can't obscure it. Its ultrasecure and robust yadda yadda.
Suppose me and five of my closest friends purchase the same watermarked tune. The amplitude of each audio stream is reduced to 1/5 and then all are mixed with each other. That could be fun to figure out.
How's this? Me and a buddy purchase the same tune. Feed each into one input of an op-amp and the only thing that will come out is the sum of the watermarks. Analysis fodder. Hmmm. Even better, what if I compare a watermarked tune against a non watermarked version? This is basically a form of stegnography and will be analyzed to death. I believe removal tools will be out within a month of such a scheme coming out.
It seems that most viruses anymore are really just trojans. If you "cleaned" them, there wouldn't be anything left. In any case, ClamAV's primary use seems to be in milters. I use it in a mail filter machine and I have it set to reject infected mail out of hand. If the sender was trying to send something legitimate then he needs to have his machine cleaned. I have problems enough protecting the machines I'm responsible for.
I'm treating this "announcement" strictly as a hypothetical musing. I truly don't know what to make of Sun's motives or intentions. You have to remember these are the people who would have you know that Linux is a good desktop but a lousy server. Sun can say and do so many contridictory things in the course of one day that I think Scott McNealy should tie it all up with dinner at Milliways.
Assuming this is for real, Java needn't become fragmented at all. For one thing, Sun could choose one of the source under glass licenses and call it Open Source. That situation wouldn't be much different than what we have now save the PR bonus/controversy for Sun. It's a pity OSS didn't get their trademark; it would have cut down on that sort of thing.
On the other hand, they could pick the LGPL or even the pure GPL and enforce the Java trademark ruthlessly. No one is going to bundle Joe-Bob's Virtual Machine. Besides, there is a vast body of code that any would be machine has to run. It takes extremely obnoxious behaivor on the part of the maintainers for such forks to even get started. Oh well, the possibility is interesting but doesn't really excite me.
One thing they could do right now is fix their retarded redistribution terms. They need all the mindshare they can get. It is a PITA to have to install Java as a third-party addon in a Linux distro. They just need to fix whatever it is that stops Suse and Redhat from bundling with it their distros. It wouldn't even hurt much if it had to live on some sort of contrib CD.
You really should check XFree86's supported distro page. The only project of any real note using it is NetBSD. Slackware was the last of the major distros to switch to X.Org. Its X.Org that has all of the "street cred" now. Any press assertion to the contrary won't hold any more water than the crap coming from Ken Brown.
Try finding a program you can install and run without administrator priviledges. There are quite a few you can run (i.e. Outlook Express, Microsoft Office), but a lot of ones you can't (TurboTax, The Sims, FAIK every game out there). Windows XP is a step in the right direction, but without the ability to run as a non administrator for a normal user, it's security "features" are worthless.
This makes Windows environments a PISSER to secure in real world environments. Actually, it makes securing it a bad joke. The OS actually provides everything an admin could reasonably need to secure it (I know. I know, application requirements vary.) I have a whole stable of apps on my hands that won't run unless the user is logged in as an admin. Just what the hell am I supposed to do? I have found I can use third party tools to emulate SUID functionality but it still sucks.
Perhaps a whitelist where everything not on the whitelist goes into the "Junk Suspects" box, combined with a Bayesian filter? I don't have the answer, but there's gotta be one.
That is essentially what I do. I have instructed Spamassassin to whitelist friends, family, work, and some mailing lists. This has the additional effect of those mails being autotrained as ham. I also have a training folder that is like your Suspects box. Anything with a positive score winds up there. Every once in a great while, something valid is there but it's usually "innovative" spam that I use to train SA's Bayes filter.
It has interesting behaivor. I'll go a week or two with nothing in the training folder. Crap will show up in it for a few days which then tapers off as I train the filter with them.
I'm interested in DSPAM for a mailserver I maintain but it seems to me that I need quite a bit of user cooperation in the beginning for it to be effective. I only have a few users that are that motivated. I understand the maintainer's arguments about heuristic filters but heuristics do have the virtue of functioning from the beginning. What I would like to see is a Spamassassin with more statistical tests in addition to the Bayes test. Once enough mail is collected to blood the filters, the heuristics can be disabled bit by bit and the statistics tests have their scores raised. That process could even be automated.
Slashdot Mirror
Intel's compiler will work for some but not all of your software. You probably won't get away with compiling the kernel and glibc with it. I understand that there has been progress lately in increasing the compatibility of the Intel compiler and gcc.
The main reason Linux programs are compiled with gcc is that code portability is valued more than it is in the Wintel world. For instance, I have virtually identical user environments on both my home P4 machine and my Powerbook at work. GCC has to produce reasonable code for at least 12 (maybe more) architectures. Intel and MS can optimize the hell out of the IA32/64 only cases. There is ongoing work to re-engineer some parts of GCC that go clear back to the mid eighties. But I seriously doubt a cross compiler will ever produce code as tight for a given arch as a compiler made specifically for that arch. Some code in the compiler is common to all arches. No x86-only optimizations will be made there. There are other parts that will be unique to each arch and I presume they will be doing what they can there.
Even Apple and BSDs are using GCC. It is arguably to most important piece of software in the FOSS world. Thousands of projects have to compile to working binaries on cisc processors, 32/64/higher bits, and big and little endian machines. It is an extremely tough thing to get right and it can't be all things on all arches.
I'm not sure I buy that. At most you are casting small RF shadow on the ground. Unless that shadow falls across a legit subscriber's antenna then it won't do jack to anyone else. That's like saying when I look up at a spotlight that I'm stopping someone else from seeing it.
The "technician" the cable company sent out didn't understand why I wanted to use "all those wires" to put baseband video and audio into our stereo. The coax was "just one wire and easier". True it was marginally easier but I don't see much hope for selling even medium end stuff when the guy who hooks cable up all day doesn't know the least thing about av gear.
I've done things like put "This is really a Gecko browser. www.w3c.org hint hint." in the agent string. As long as that magic "IE" is in there, it'll usually let you in. If not, I'll hit the site a few times to get it in their logs and then change to IE on XP agent string. The User Agent Switcher extension is very handy for this sort of thing.
Now if we could just do something about javascript based detection techniques.....
Even if it were suddenly 50/50 overnight, or 75/25 in favor of FF, sites would still make IE-compliant pages, because nobody can affort to write off 25% of the market.
I'm getting deja vu here. Where have I seen this happen before. Holy shit! Netscape 4 has risen from the dead and walks the earth. IE now is like NS4 then. Various odd things happen with the cache in SSL mode. Alpha channel and CSS support are completely broken. Now everyone has to write ugly hacks to support IE once the standards compliant site is done. What a piece of shit!
So what you're saying is that there is no use mucking about with all these reflectors, LEDs and computers. Just paint the guy pink and turn on a cheap and simple Somebody Else's Problem Field.
Sure AAC is better, sure Ogg is better, but for most folks, even those with huge music collections and very exacting preferences in their audio systems, MP3 is still good enough. Why? Because most people care about the music, not the technology....
Beta was better than VHS but VHS won too.
That's pretty much apples and oranges though. There is a vast difference in the difficulty required to support a codec versus a physical format. Given sufficiently powerful and flexible playback devices, supporting new codecs is trivial. If they can't already, upcoming portables will be able to support any arbitrary codec with a firmware update.
What is needed are portable, easily hackable, and powerful devices. This makes the bitstream format entirely a matter of consumer choice. Of course such devices will support mp3, wma, and maybe even aac by default. Moore's law applies to the portables as well which means such devices are around the corner. Intrepid PDA owners have already had this for awhile.
This brings up a good point; how do you explain to someone who is NOT a geek (and has no interest in being one) about what DRM is and how it will effect them?
By vastly oversimplifying of course! Just tell people that it will almost always make it impossible "to record". When it isn't impossible, it will be very complicated and you'll need a geek such as yourself to make it work.
I'd just go to the bad side and be done with it. I've coded some small things in php because we needed them and weren't going to get them any other way. I pretty much learned what html and SQL I know at the same time. I have no trouble admitting this because I'm an admin who occaisionally codes rather than the other way around.
I've tried to avoid pulling obvious boners but I'm glad none of the machines running these things are exposed to the public net.
Will we finally find out what the secret recipe for Big Mac sauce is?
It's Thousand Island salad dressing.
Why would it take a long time if X.org is just a fork from XFree?
Debian does a lot patching, cleaning up, and build script twiddling to make their packages run on all architectures. This only recently happened for XFree86 4.3. When they transition to a 4.4 base, they'll go with X.Org same as everybody else.
Oh, sorry, I forgot how NTFS sometimes decides to get bitchy and doesn't let you resize it...
Be sure to defrag it first. I think the gremlins that live in NTFS partitions are happier when they are snuggly wrapped in well packed blocks. Seriously, shut down as many processes as possible before you do defrag and then shut down Windows as quickly as possible when you're done. That procedure helps me more than anything else.
Much like the upcoming Reiser4, I think NTFS isn't sure if it wants to be an object database or a filesystem when it grows up. It sure seems "bitchy" for third party tools to work with though.
Given enough budget, you would be hard pressed to see that I even have a computer. Think stereo systems built into a wall. The best setup I ever saw had the tower on a shelf behind a utility door. It could be as pretty or ugly and not matter, because it's just something else to clutter the decor. Some people put neon in their cases for crying out loud. You can even get them prebuilt that way. Ideally, the only things I want visible are an unobtrusive screen, mouse, keyboard. Gee-whiz industrial design is completely wasted when those are the criteria.
Without an unlimited budget, I hide it under the desk as I said. It doesn't matter what it looks like.
By contrast, there have been around, what, five? standard PC motherboard sizes since the 386. Commodity parts are great for end user prices, but commodity means "same", and it shows in the final product.
Maybe so but why should I care? The tower is something goes under a desk or behind something. Its a computer not an objet de art or something. LCDs are skinny things that should fit on a desk and not draw attention to themselves. Sorry Mr. Industrial Designer but I would rather not see it at all rather than be wowed by it.
If Linux is being worked upon by people whose idea of vision is rhetoric such as "destroy Windows within 10 years" and "Longhorn will be a disaster", then there is no end in sight to the hegemony of Microsoft.
Very little code is being written with the explicit purpose of "destroying MS". Come to think of it, I don't know of any. I see two main reasons the free UNIXes are being developed. One, there is money to made by carving out a piece of the market from established competitor. After all, if MS is making a pile of cash why not siphon off a bit for yourself. This pretty much applies to FOSS pure plays and larger entities like IBM and Novell. Note well that the aim is isn't "destroy MS!", it's "Let's make some money!".
The other reason is what gets called "scratching the itch". There is software available for the need but it costs too much or the vendor can't be trusted. That last is more and more true of MS everyday. Or there isn't anything that precisely addresses the need so you write it yourself. That was certainly true of a few trivial things I've done in PHP.
In the early nineties, Mattel had an infamous Barbie doll that told girls gems like "Math is hard. ..giggle.." Some activists broke into a warehouse and swapped boards between the GI-Joes and Barbies so that the Barbies were saying things like "Stop Cobra!".
http://ifaq.wap.org/posters/barbiedir.pdf
Oh, and my wife is proof positive it isn't tough to NOT get infected .. 3 years on her own Win98 box and zero infections with no anti-virus suite running on the machine at all.
Then how do you know there are no viruses on the machine? Malware doesn't have to be obvious when it's running.
We're talking about the people who compared VCRs to the Boston Strangler. We're talking about the same industry that has had fear and loathing of every consumer technology since the first wax cylinders. Video rental could have been a reality in the early seventies. They were worried about repeat viewings so the tapes we're gimmicked so they could only be rewound at the video store. "But what if there are more people at the home than the person who rented it?"
You propose reasonableness. The problem is that the industry has been stupid and paranoid for over 90 years and they aren't going to stop now. If there is to be dialog and respect then they have to make the first move.
Lets assume they come up with a watermark that really works. You can't remove it and you can't obscure it. Its ultrasecure and robust yadda yadda.
Suppose me and five of my closest friends purchase the same watermarked tune. The amplitude of each audio stream is reduced to 1/5 and then all are mixed with each other. That could be fun to figure out.
How's this? Me and a buddy purchase the same tune. Feed each into one input of an op-amp and the only thing that will come out is the sum of the watermarks. Analysis fodder. Hmmm. Even better, what if I compare a watermarked tune against a non watermarked version? This is basically a form of stegnography and will be analyzed to death. I believe removal tools will be out within a month of such a scheme coming out.
It seems that most viruses anymore are really just trojans. If you "cleaned" them, there wouldn't be anything left. In any case, ClamAV's primary use seems to be in milters. I use it in a mail filter machine and I have it set to reject infected mail out of hand. If the sender was trying to send something legitimate then he needs to have his machine cleaned. I have problems enough protecting the machines I'm responsible for.
I'm treating this "announcement" strictly as a hypothetical musing. I truly don't know what to make of Sun's motives or intentions. You have to remember these are the people who would have you know that Linux is a good desktop but a lousy server. Sun can say and do so many contridictory things in the course of one day that I think Scott McNealy should tie it all up with dinner at Milliways.
Assuming this is for real, Java needn't become fragmented at all. For one thing, Sun could choose one of the source under glass licenses and call it Open Source. That situation wouldn't be much different than what we have now save the PR bonus/controversy for Sun. It's a pity OSS didn't get their trademark; it would have cut down on that sort of thing.
On the other hand, they could pick the LGPL or even the pure GPL and enforce the Java trademark ruthlessly. No one is going to bundle Joe-Bob's Virtual Machine. Besides, there is a vast body of code that any would be machine has to run. It takes extremely obnoxious behaivor on the part of the maintainers for such forks to even get started. Oh well, the possibility is interesting but doesn't really excite me.
One thing they could do right now is fix their retarded redistribution terms. They need all the mindshare they can get. It is a PITA to have to install Java as a third-party addon in a Linux distro. They just need to fix whatever it is that stops Suse and Redhat from bundling with it their distros. It wouldn't even hurt much if it had to live on some sort of contrib CD.
You really should check XFree86's supported distro page. The only project of any real note using it is NetBSD. Slackware was the last of the major distros to switch to X.Org. Its X.Org that has all of the "street cred" now. Any press assertion to the contrary won't hold any more water than the crap coming from Ken Brown.
Try finding a program you can install and run without administrator priviledges. There are quite a few you can run (i.e. Outlook Express, Microsoft Office), but a lot of ones you can't (TurboTax, The Sims, FAIK every game out there). Windows XP is a step in the right direction, but without the ability to run as a non administrator for a normal user, it's security "features" are worthless.
This makes Windows environments a PISSER to secure in real world environments. Actually, it makes securing it a bad joke. The OS actually provides everything an admin could reasonably need to secure it (I know. I know, application requirements vary.) I have a whole stable of apps on my hands that won't run unless the user is logged in as an admin. Just what the hell am I supposed to do? I have found I can use third party tools to emulate SUID functionality but it still sucks.
Perhaps a whitelist where everything not on the whitelist goes into the "Junk Suspects" box, combined with a Bayesian filter? I don't have the answer, but there's gotta be one.
That is essentially what I do. I have instructed Spamassassin to whitelist friends, family, work, and some mailing lists. This has the additional effect of those mails being autotrained as ham. I also have a training folder that is like your Suspects box. Anything with a positive score winds up there. Every once in a great while, something valid is there but it's usually "innovative" spam that I use to train SA's Bayes filter.
It has interesting behaivor. I'll go a week or two with nothing in the training folder. Crap will show up in it for a few days which then tapers off as I train the filter with them.
I'm interested in DSPAM for a mailserver I maintain but it seems to me that I need quite a bit of user cooperation in the beginning for it to be effective. I only have a few users that are that motivated. I understand the maintainer's arguments about heuristic filters but heuristics do have the virtue of functioning from the beginning. What I would like to see is a Spamassassin with more statistical tests in addition to the Bayes test. Once enough mail is collected to blood the filters, the heuristics can be disabled bit by bit and the statistics tests have their scores raised. That process could even be automated.
The best "evidence" for Dvorak superiority is a study conducted by Dvorak himself. There isn't much in the way of third party evidence for it.