This is good. One of the major problems with graphic design systems, both CAD and animation, is that it's only possible to select one thing at a time. Many operations involve two objects, and you're forced to some sequential select-and-manipulate interface. This gets you past that.
Many high-end animation systems will accept multiple input devices, from MIDI keyboards to knob boxes to articulated skeletons. At the low end, we have the scroll wheel, which was a big improvement. Finally, you could do two things at once without shifting modes. This is a further step in that direction.
Video editors on deadline are going to want this thing. The obvious application is a replacement for the Avid NewsCutter, which sells into an environment which is not very price-sensitive.
This was the concept that propelled Autodesk into working on virtual reality in the 1980s. Drawing in 3D on a screen required (and still requires) a complex interface; gestures in 3D looked like a way to make it a freehand job.
Didn't work. Humans can't draw with any degree of precision in 3D free space. Clay sculpting works because of tactile feedback; it's not a pure eye/hand thing. And drawing in 3D free space gets really tiring within minutes. Technically, you can get the hardware and software to work. But it's too hard on the users.
There was no lack of legislative authority to deal with Katrina. The real problem was that 1) the FEMA director slot was being filled by a jerk, and 2) the FEMA director wasn't allowed to exercise his statutory authority.
Under the Disaster Relief and Emergency Assistance Act (42 U.S.C. 5170b(c)), the Director of FEMA had the statutory authority to task other federal agencies to provide specific assistance to states overwhelmed by disaster. But when Homeland Security was established, FEMA was put under DHS and deemphasized. So, while Brown had the statutory authority to call up DoD and tell them to get moving, he wasn't allowed by his boss, Chertoff, to do it.
And Chertoff didn't issue that authority to Brown unti 36 hours after Katrina hit.
Clinton's former FEMA director, Witt, spoke after Katrina of how they did things when he was in charge: "We had all the resources of the federal government. We didn't have to ask anyone to activate the Department of Defense. I did that. I called up the President." Reagan's FEMA director, a former general, also had and used that authority. Brown wasn't even allowed to call up the President.
Part of the problem was that Brown, personally, was a nobody. His previous career peak was chief of horse show judges for the Arabian Horse Association. Clinton's and Reagan's FEMA directors had track records; Clinton's was respected by everyone, and Reagan's was a former general. Both of them would have had no problem calling up people at DoD and telling them to start loading up, with official orders to follow as necessary.
Brown could have gotten things moving even without the authority if he'd had a clue. He could have called up people at the Pentagon and subordinate commands and said "This is the FEMA director. You've seen CNN; New Orleans is under water. You're going to be tasked to go there and do... as soon as we get the paperwork done in Washington. Meanwhile, I'm giving you a heads-up; get loaded up and ready to move." The military understands back-channel stuff like that. In the military, if there's big trouble coming, subordinate commanders are expected to crank up and get ready to roll, with or without orders, so if and when the go order comes, things happen fast.
Before an attack can install something like "Blue Pill", it has to be running in kernel mode. At that point, it already has full control of the machine. The only question is what to do with that control. Installing a hypervisor underneath the OS is kind of neat, but there are lots of other things to do.
What this does demonstrate is that after-the-fact malware detectors are a dead end.
There's a great comment in the article:
The solution (includes) checking all the possible "dynamic hooking places" in kernel data sections.
(This) is actually impossible to achieve 100 percent as nobody knows all those dynamic hooking places, but we could at least start building a list of them. I believe the number of the hooking places is a finite number for every given operating system.
In other words, there is only a finite number of "ways" to write Type II malware of any specific kind (e.g. a keystroke logger).
Now that's a big part of the problem - Microsoft's use of "dynamic hooking", or places where user code can insert callbacks which privileged code might access, is so messed up that security researchers can't even find all the places where it is allowed.
"Dynamic hooking" is really a lame method of interprocess communication left over from the DOS version of Windows. It should never have made it into NT/W2000/XP/etc.
There's less of a temptation to do this in open source operating systems, since, if you really need to legitimately add a feature, you can put it in the source, rather than tapping into some binary. The Linux netfilter/ipchains mechanism offers a "dynamic hooking" attack vector into the kernel, though, so Linux isn't immune to attacks of this type.
AMD is converting Fab 30 in Dresden from 90um and 200mm wafers to Fab 38, with 65nm and 300mm wafers. This should come on line in 2007. Longer term, AMD is building a new fab in upstate New York for 32nm features on 300mm wafers. That should come on line in 2010.
Down at the user level, this means that first shipments of AMD CPUs made with 65nm technology should appear in December of 2006.
Coming soon to Dell Dimension desktops.
Actually, Stallman is a better legal drafter and policymaker than a programmer. The GPL is his great work, and that's what he'll be remembered for. He changed the world of software with it.
He's known as a programmer mostly for reimplementing EMACS; the original version, in TECO macros, was primarily by Guy Steele. Stallman's version of EMACS was bloated; JOVE (Johnathan's Own Version of EMACS), written by a high school student in the 1980s, had most of the useful functionality of EMACS in 128K of code. Stallman's frantic efforts at Lisp Machines Incorporated to keep up with Symbolics didn't lead anywhere; the whole LISP machine world was a dead end. Stallman's HURD operating system effort was a decade-long disaster.
Stallman did write the original GCC, but eventually it was replaced by the EGCS compilers, which are now the "official" GCC suite of compilers.
Almost all of Stallman's programs were reimplementations of something already written by someone else.
But the GPL was original and brilliant. That's his legacy.
Six years out of school and "programming on and off" seems strange. What kind of programming do you want to do? GUI stuff? Graphics? Games? Algorithms? Databases? Real-time?
One way to get better as a programmer is to do maintenance programming on code written by someone better than you. Learning to understand someone's thinking by reading their code can be a worthwhile exercise. It's also useful to be able to write in someone else's style.
Right now, something worth getting good at is understanding how to write highly parallel programs that are reliable. Write something that has lots of intercommunicating threads and be confident the locking is correct. There aren't that many people who consistently get that right. You have a CS degree, so you have the theory for that. Put it into practice. The world is full of underutilized multiprocessors. Learning how to write safe concurrent code will definitely make you a better programmer. (It will also make you realize how bad most mainstream programming languages are for this.)
On the language front, today I'd say that you should be good at either Java or C++ (C# if you're in Microsoft land), and either Perl or Python (VB if you're in Microsoft land). One strongly typed language that goes fast, and one weakly typed interpreter. Basic familiarity with the HTML/PHP/Javascript world is useful, but don't spend all your time on the details of that - that's for low-level programmers with two years of experience. Also, learn how and when to use a relational database, at least at the MySQL call level.
It's not clear from the Adia diamonds whether these are grown like semiconductor wafers or made in high-pressure presses. Gemesys has a Florida plant making gemstones in high-pressure presses. They finally caved in to deBeers and laser-engraves their stones with some ID information. The FTC caved in to the diamond industry and insists they be called "cultured diamonds". They're distinguishable from natural diamonds by their absorbtion spectrum, and deBeers has a tester for this
Grown synthetics were still experimental when Wired wrote their article, but that's the more promising process. Those, in theory, can be indistinguishable from natural ones.
The diamond industry had painted itself into a corner with the concept that the most valuable diamonds are "flawless". You do not want to be in that marketing position when going up against the technology that makes semiconductor wafers. Look for PR about how real diamonds have "natural flaws".
Tied to this is the "Kimberly Process", the agreement supposedly intended to restrict the flow of conflict diamonds. This requires source documentation to travel along with diamonds as they pass through the distribution chain. Previously, diamonds were generic; nobody cared where they came from. The Kimberly Process has the effect of making it much harder to insert large quantities of synthetic diamonds into the distribution system.
Incidentally, most industrial diamonds have been synthetic for years. Annual synthetic production is around 600 metric tons, most of it in the form of abrasive grits for cutting wheels and such. When you need to cut a slot in concrete pavement, you use a diamond cutting wheel.
If you accept payments through a web site without disclosing who you are, you're a criminal.
That's the law in California, or if you sell in Calfornia.
California Business and Professions code section 17538 provides that:
(d) A vendor conducting business through the Internet or any other electronic means of communication shall do all of the following when the transaction involves a buyer located in this state:
(1) Before accepting any payment or processing any debit or credit charge or funds transfer, the vendor shall disclose to the buyer in writing or by electronic means of communication, such as e-mail or an on-screen notice, the vendor's return and refund policy, the legal name under which the business is conducted and, except as provided in paragraph (3), the complete street address from which the business is actually conducted.
(2) If the disclosure of the vendor's legal name and address information required by this subdivision is made by on-screen notice, all of the following shall apply:
(A) The disclosure of the legal name and address information shall appear on any of the following: (i) the first screen displayed when the vendor's electronic site is accessed, (ii) on the screen on which goods or services are first offered, (iii) on the screen on which a buyer may place the order for goods or services, (iv) on the screen on which the buyer may enter payment information, such as a credit card account number, or (v) for nonbrowser-based technologies, in a manner that gives the user a reasonable opportunity to review that information. The communication of that disclosure shall not be structured to be smaller or less legible than the text of the offer of the goods or services....
(g) Any violation of the provisions of this section is a misdemeanor punishable by imprisonment in the county jail not exceeding six months, by a fine not exceeding one thousand dollars ($1,000), or by both that imprisonment and fine.
Yes, running a business anonymously is a criminal offense. You can incorporate, if you like, but you can't just use a psuedonym.
There should be a mechanism for encrypting web data that doesn't rely on paying a third party for a service only tangentially related to encryption.
There is. You can still get cheap SSL certificates. But if you're accepting payments, plan on getting one that clearly identifies who you are.
If you accept payments through a web site without disclosing who you are, you're a criminal. (California Business and Professions code section 17538, other provisions in other jurisdictions.) And soon, browsers are going to put up a big red flag that will make your customers go away.
Has anyone actually been able to find the specification for "high assurance" certificates? Apparently this is being closely held. The spec comes from something called the "CA Browser Forum", which is invitation-only and doesn't seem to have a web site. A standard was supposed to be issued in August, but apparently agreement wasn't reached until a meeting in September. There are many press releases, but no hard data.
So that's why it's not in Mozilla.
It's actually a good idea. Early in the history of SSL, getting a certificate required presenting appropriate business identification info to the certificate issuer. The problem is that some issuers (GoDaddy comes to mind) started issuing "domain only" SSL certificates; the only verification is that the domain can get email. Then, instead of revoking GoDaddy's root certificate for this, the other cert issuers copied GoDaddy's approach. Now anybody can get a meaningless certificate with a meaningless Relying Party Agreement.
The way it's supposed to work is that the certificate issuer bears financial responsibility for misidentification of the certificate owner. Some certificates from Verisign have a Relying Party Agreement that does provide a financial guarantee to the party relying on the certificate - $100 for a class 1 cert, $5000 for a class 2 cert, and $100,000 for a class 3 cert. Most of the other issuers have relying party agreements which promise nothing and deliver less.
So what's happening is that, soon, you'll be able to tell the difference between the crap certificates and the good ones. Before you buy. The idea is that if you put your credit card into a site that showed a green toolbar in IE, and it wasn't really the company it should have been, you can collect from the certificate issuer. This puts certificate issuers on the hook for phishing losses.
Unfortunately, the rules and the Relying Party Agreements for the new certificates haven't yet appeared, so we can't tell if the rules are tough enough to make this work. Since they're being drafted by the certificate issuers, there will probably be some loophole that lets them off the hook.
If Microsoft actually had a microkernel (and NT 3.51 was close), this would be reasonable. At NT 4, a vast amount of crap from Windows 95 was put in the NT kernel for "compatibility", which is how we got to the mess we have today.
QNX has a closed microkenel. It does timers, memory management, CPU dispatching, and interprocess communication. That's it. All drivers, networking, file systems, etc. are outside the kernel as user processes. Nothing other than the stock kernel runs in kernel space. So the kernel changes very little over the years, leading to very good reliability. (It can even be put in ROM, and often is on embedded systems). Sometimes system components outside the kernel break; they crash, an ordinary process core dump is taken, and it can be examined in an ordinary debugger. Drivers are such components. So there's a clear separation between the stuff you don't need to change, and the stuff you might want to change. Even driver developers don't need to see inside the kernel; interaction with the kernel is through the regular user program APIs, with a few extra calls for privileged drivers allowed to map physical memory and access device hardware.
Much the same is true of IBM's VM. So this can be done right.
But with Microsoft's bloated kernel, locking down the kernel has much stronger implications. Many functions are in the kernel because Microsoft wants to control them. Far more is there for legacy reasons. (My favorite is the decompresser for.RLE images, which has an exploitable bug.) There's also too much in there because it seems to be needed during the boot process. (One big lesson of OS design: use a boot loader that lets you load both the kernel and various support programs and libraries before the kernel starts. This reduces the temptation to put everything needed at startup in the kernel. Linux now has this, but got it late, so too much went into the Linux kernel too.)
Another issue is the DRM problem, or how to protect DRM code without putting it in the kernel. Microsoft doesn't seem to have a good solution for that. The right answer is probably rings of protection, like Multics or VMS, but Microsoft didn't go that route.
Symantec used to sell compilers, developer tools, and even some user applications like ThinkTank, an early outliner. Microsoft pushed them out of the tools field on Windows; Symantec had a more portable alternative to MFC, and Microsoft didn't like that. Outliners disappeared as a standalone product category; Word now does that. All that's left is the anti-virus business. Now that, too, looks like it's toast.
Actually, the OS vendor should be doing the security system. The primary function of an operating system is security and resource management; everything else could potentially be an application. Only because of Microsoft's appallingly bad security does the anti-virus industry even exist.
While Wales is talking about this, the Internet Archive is quietly doing it and has been for several years now. There's a sizable book scanning effort underway, in cooperation with various libraries and Yahoo. Automatic book scanning equipment has been developed and is in use. The images of the pages are stored, and the text extracted with OCR for searching and other uses. They have over 33,000 books on line now.
Loading existing works into a wiki system isn't particularly useful, anyway. Making them editable probably won't make them better.
The move to PowerPC was Apple's big mistake. That was the point at which Apple market share dropped, and it never came back. Even today, Apple has much lower market share than it did the day the PowerPC machines were announced. The argument for going with the PowerPC was that IBM was going to make Macs. Yes, that was the whole point of the deal. Didn't happen, but that was Apple's big plan. And that bad move happened under Jobs.
In fact, when the PowerPC 601 came out, Motorola was shipping the 68060, which outperformed the early PowerPC chips. The 68000 line could have been developed further; there was nothing in the architecture that limited it. But when Apple dropped it, that was the end of the demand for high-end 68000 parts.
The PowerPC transition killed many existing apps. The engineering community dumped the Mac at the PowerPC transition; existing CAD applications like AutoCAD were not ported to PowerPC, and most of the printed circuit board design applications were dropped at that point, too. So Apple lost a whole market segment, and one willing to pay for big screens and good graphics.
Copeland was actually a good operating system. The problem was that applications had to be revised for it, and Microsoft didn't want to bother. Apple no longer had the clout with developers it had had back at the System 7 transition, where all apps had to be revised. But Apple hadn't realized internally that it could no longer order developers around; the developers had the option of going to Windows. So backwards compatibility had become more important.
Copeland (the original "MacOS 8") actually shipped to some developers. It was almost ready to go. Acquiring NeXT delayed the release of a new OS by several years; it took much longer to get NeXT code onto the Apple platform than Jobs said it would. But it saved Jobs' ass financially; he was heavily invested in NeXT, which was headed for bankruptcy.
As for design, one of the coolest Macs ever was the 20th Anniversary Mac, the first Mac with an LCD panel. In 1997, way ahead of everyone else. That was before Jobs took over and "Steved" the product, because it wasn't his.
The iMac clamshell looked like the Lear-Seigler ADM 3A from 1977, which was a very popular low-end terminal in its day. It wasn't an original concept.
Jobs' big contribution was to suck up to Gates and thus keep Microsoft Office on the Mac That's what saved Apple.
So that's what it looks like with the Reality Distortion Field turned off.
Marginally irrelevant, but good news on spam: Update on Jeanson James Ancheta, botnet spammer. The short version: he's now Federal inmate number 32392-112 at the California City Correctional Institution.
Like I said before, YouTube has the same legal problems Napster did.
If they're not yet doing so, I'd expect the RIAA to start running a song recognition program against YouTube content. That will catch all those videos with commercial music attached.
Spammers register pump-and-dump spam domains for use in spam runs. These domains are commonly discarded after a few days. The tactic is commonplace but the the arrest of alleged botmaster Jeanson James Ancheta, 20, of Downey, California, on 3 November has been accompanied by a radical shift in the landscape.
"Up to recently, the graphs were all fairly smooth, with the stats showing that 12 days was about the maximum lifetime for this type of domain, while 30 per cent only lasted a day or under, and 10 per cent only lasted three hours or under," Shipp said. "This kind of activity just disappeared completely from the radar on 2 November."
Location: CALIFORNIA CITY CORRECTIONAL INSTITUTION
California City Prison: "This medium security desert prison opened in 2000, and is a stunning sight, either by day when its monolithic forms stand out on the desert pavement like ancient Egyptian architecture, or by night when floodlights bathe the gleaming facility in an orange glow which can be seen from as much as 30 miles away."
This should be reported, in very clear terms, to "enforcement@sec.gov". Or on the SEC's online form. Or to the SEC Division of Enforcement, 100 F Street, N.E. Washington, D.C. 20549. Because it's a felony being committed in support of a pump-and-dump stock scam.
The stock being hyped is "TTEN", which has very low volume. The SEC can find out who was trading it just before the spam run started. That's how to find the people behind this. They can follow the money.
So put together a comprehensive package listing all known stocks being hyped by this thing and the dates the spam began, and ship it off to the SEC. The SEC and FinCen (the U.S. Treasury Financial Crimes Enforcement Network) have the data mining tools to look at the stock transactions and find the people behind this. The SEC has gone after pump-and-dump spammers many times before, and they usually get them.
"Team 2 is developing nanomaterials that are capable of mechanical actuation and dynamic stiffness. As part of the soldier's battlesuit, these adaptive multifunctional materials will improve soldier survivability.
Mechanical actuators embedded as part of a soldier's uniform will allow a transformation from a flexible and compliant material to a non-compliant material that becomes armor, thus protecting the soldier by distributing impact. Soft switchable clothing can also be transformed into a reconfigurable cast that stabilizes an injury such as a broken leg. Contracting materials can be made to apply direct pressure to a wound, function as a tourniquet, or even perform CPR when needed. Mechanical actuators can also be used as exo-muscles for augmentation of a soldier's physical strength or agility and as wound compresses."
The leading teams this time are Stanford/Volkswagen, CMU/General Motors, and Oshkosh Truck. The prize doesn't matter to either. And all three already got $1 million in Government money. Each.
It's a much tougher job this time. Driving in traffic, backing out of tight spots, parking and unparking. I'll be surprised if anyone wins the first year, because nobody knows quite what to expect. Year two, someone will win.
I've been expecting this for a while, ever since the transistor count of the GPU passed that of the CPU. Actually, I thought it would happen sooner. It's certainly time. Putting more transistors into a single CPU doesn't help any more, which is why we now have "multicore" machines. So it's time to put more of the computer into a single part.
NVidia already makes the nForce line, the "everything but the CPU" part, with graphics, Ethernet, disk interface, etc. If they stick a CPU in there, they have a whole computer.
Chip designers can license x86 implementations; they don't have to be redesigned from scratch. This isn't going to be a tough job for NVidia.
What we're headed for is the one-chip "value PC", the one that sits on every corporate desk. That's where the best price/performance is.
Of course this happened. YouTube is the next Napster. Same centralized hosting of uploaded content, same business model, same excuses, same legal problems. YouTube is in a worse legal position than Napster. Napster just hosted the index. YouTube hosts the actual content.
YouTube could well be shut down by an injunction. That's what happened to Napster.
"Napster is enjoined from copying or assisting or enabling or contributing to the copy or duplication of all copyrighted songs and musical compositions of which the plaintiffs hold rights." -- U.S. District Judge Marilyn Hall Patel.
As for it being the responsibility of the copyright holder to find the material,
"Napster wrote the software; it's up to them to write software that will remove from users the ability to copy copyrighted material," -- Judge Patel
YouTube, like Napster, is a contributory infringer.
"The district court determined that plaintiffs had demonstrated they would
likely succeed in establishing that Napster has a direct financial interest in the
infringing activity. We agree. Financial
benefit exists where the availability of infringing material "acts as a 'draw' for
customers." -- 9th Circuit Court of Appeals.
I was amazed that Google bought YouTube. It was obvious they were buying into a huge litigation problem.
Actually, the Mac IIci went together very well. That was a vertical-assembly machine; everything went in with a straight-down move. Including the power supply, which was just pushed in vertically, and engaged alignment guides on the case and connectors on the motherboard.
Then Apple offshored manufacturing, and design for automated assembly mattered less.
Slashdot Mirror
This is good. One of the major problems with graphic design systems, both CAD and animation, is that it's only possible to select one thing at a time. Many operations involve two objects, and you're forced to some sequential select-and-manipulate interface. This gets you past that.
Many high-end animation systems will accept multiple input devices, from MIDI keyboards to knob boxes to articulated skeletons. At the low end, we have the scroll wheel, which was a big improvement. Finally, you could do two things at once without shifting modes. This is a further step in that direction.
Video editors on deadline are going to want this thing. The obvious application is a replacement for the Avid NewsCutter, which sells into an environment which is not very price-sensitive.
vaulted (sic) hero, John Wayne
John Wayne was not a hero. He was an actor who played heroic roles.
He did not serve in the military during WWII; he got a draft deferment.
Great actor, yes. Inspiration to many, yes. Hero, no.
This was the concept that propelled Autodesk into working on virtual reality in the 1980s. Drawing in 3D on a screen required (and still requires) a complex interface; gestures in 3D looked like a way to make it a freehand job.
Didn't work. Humans can't draw with any degree of precision in 3D free space. Clay sculpting works because of tactile feedback; it's not a pure eye/hand thing. And drawing in 3D free space gets really tiring within minutes. Technically, you can get the hardware and software to work. But it's too hard on the users.
So virtual reality CAD R&D was dropped.
There was no lack of legislative authority to deal with Katrina. The real problem was that 1) the FEMA director slot was being filled by a jerk, and 2) the FEMA director wasn't allowed to exercise his statutory authority.
Under the Disaster Relief and Emergency Assistance Act (42 U.S.C. 5170b(c)), the Director of FEMA had the statutory authority to task other federal agencies to provide specific assistance to states overwhelmed by disaster. But when Homeland Security was established, FEMA was put under DHS and deemphasized. So, while Brown had the statutory authority to call up DoD and tell them to get moving, he wasn't allowed by his boss, Chertoff, to do it. And Chertoff didn't issue that authority to Brown unti 36 hours after Katrina hit.
Clinton's former FEMA director, Witt, spoke after Katrina of how they did things when he was in charge: "We had all the resources of the federal government. We didn't have to ask anyone to activate the Department of Defense. I did that. I called up the President." Reagan's FEMA director, a former general, also had and used that authority. Brown wasn't even allowed to call up the President.
Part of the problem was that Brown, personally, was a nobody. His previous career peak was chief of horse show judges for the Arabian Horse Association. Clinton's and Reagan's FEMA directors had track records; Clinton's was respected by everyone, and Reagan's was a former general. Both of them would have had no problem calling up people at DoD and telling them to start loading up, with official orders to follow as necessary.
Brown could have gotten things moving even without the authority if he'd had a clue. He could have called up people at the Pentagon and subordinate commands and said "This is the FEMA director. You've seen CNN; New Orleans is under water. You're going to be tasked to go there and do ... as soon as we get the paperwork done in Washington. Meanwhile, I'm giving you a heads-up; get loaded up and ready to move." The military understands back-channel stuff like that. In the military, if there's big trouble coming, subordinate commanders are expected to crank up and get ready to roll, with or without orders, so if and when the go order comes, things happen fast.
Before an attack can install something like "Blue Pill", it has to be running in kernel mode. At that point, it already has full control of the machine. The only question is what to do with that control. Installing a hypervisor underneath the OS is kind of neat, but there are lots of other things to do.
What this does demonstrate is that after-the-fact malware detectors are a dead end.
There's a great comment in the article:
The solution (includes) checking all the possible "dynamic hooking places" in kernel data sections.
(This) is actually impossible to achieve 100 percent as nobody knows all those dynamic hooking places, but we could at least start building a list of them. I believe the number of the hooking places is a finite number for every given operating system.
In other words, there is only a finite number of "ways" to write Type II malware of any specific kind (e.g. a keystroke logger).
Now that's a big part of the problem - Microsoft's use of "dynamic hooking", or places where user code can insert callbacks which privileged code might access, is so messed up that security researchers can't even find all the places where it is allowed. "Dynamic hooking" is really a lame method of interprocess communication left over from the DOS version of Windows. It should never have made it into NT/W2000/XP/etc.
There's less of a temptation to do this in open source operating systems, since, if you really need to legitimately add a feature, you can put it in the source, rather than tapping into some binary. The Linux netfilter/ipchains mechanism offers a "dynamic hooking" attack vector into the kernel, though, so Linux isn't immune to attacks of this type.
AMD is converting Fab 30 in Dresden from 90um and 200mm wafers to Fab 38, with 65nm and 300mm wafers. This should come on line in 2007. Longer term, AMD is building a new fab in upstate New York for 32nm features on 300mm wafers. That should come on line in 2010.
Meanwhile, AMD's main fab, Fab 36 in Dresden, is starting to produce 65nm features on 200mm wafers. AMD is also outsourcing some production to a 65nm fab in Singapore.
Down at the user level, this means that first shipments of AMD CPUs made with 65nm technology should appear in December of 2006. Coming soon to Dell Dimension desktops.
Actually, Stallman is a better legal drafter and policymaker than a programmer. The GPL is his great work, and that's what he'll be remembered for. He changed the world of software with it.
He's known as a programmer mostly for reimplementing EMACS; the original version, in TECO macros, was primarily by Guy Steele. Stallman's version of EMACS was bloated; JOVE (Johnathan's Own Version of EMACS), written by a high school student in the 1980s, had most of the useful functionality of EMACS in 128K of code. Stallman's frantic efforts at Lisp Machines Incorporated to keep up with Symbolics didn't lead anywhere; the whole LISP machine world was a dead end. Stallman's HURD operating system effort was a decade-long disaster.
Stallman did write the original GCC, but eventually it was replaced by the EGCS compilers, which are now the "official" GCC suite of compilers.
Almost all of Stallman's programs were reimplementations of something already written by someone else. But the GPL was original and brilliant. That's his legacy.
Six years out of school and "programming on and off" seems strange. What kind of programming do you want to do? GUI stuff? Graphics? Games? Algorithms? Databases? Real-time?
One way to get better as a programmer is to do maintenance programming on code written by someone better than you. Learning to understand someone's thinking by reading their code can be a worthwhile exercise. It's also useful to be able to write in someone else's style.
Right now, something worth getting good at is understanding how to write highly parallel programs that are reliable. Write something that has lots of intercommunicating threads and be confident the locking is correct. There aren't that many people who consistently get that right. You have a CS degree, so you have the theory for that. Put it into practice. The world is full of underutilized multiprocessors. Learning how to write safe concurrent code will definitely make you a better programmer. (It will also make you realize how bad most mainstream programming languages are for this.)
On the language front, today I'd say that you should be good at either Java or C++ (C# if you're in Microsoft land), and either Perl or Python (VB if you're in Microsoft land). One strongly typed language that goes fast, and one weakly typed interpreter. Basic familiarity with the HTML/PHP/Javascript world is useful, but don't spend all your time on the details of that - that's for low-level programmers with two years of experience. Also, learn how and when to use a relational database, at least at the MySQL call level.
It's not clear from the Adia diamonds whether these are grown like semiconductor wafers or made in high-pressure presses. Gemesys has a Florida plant making gemstones in high-pressure presses. They finally caved in to deBeers and laser-engraves their stones with some ID information. The FTC caved in to the diamond industry and insists they be called "cultured diamonds". They're distinguishable from natural diamonds by their absorbtion spectrum, and deBeers has a tester for this
Grown synthetics were still experimental when Wired wrote their article, but that's the more promising process. Those, in theory, can be indistinguishable from natural ones.
The diamond industry had painted itself into a corner with the concept that the most valuable diamonds are "flawless". You do not want to be in that marketing position when going up against the technology that makes semiconductor wafers. Look for PR about how real diamonds have "natural flaws".
Tied to this is the "Kimberly Process", the agreement supposedly intended to restrict the flow of conflict diamonds. This requires source documentation to travel along with diamonds as they pass through the distribution chain. Previously, diamonds were generic; nobody cared where they came from. The Kimberly Process has the effect of making it much harder to insert large quantities of synthetic diamonds into the distribution system.
Incidentally, most industrial diamonds have been synthetic for years. Annual synthetic production is around 600 metric tons, most of it in the form of abrasive grits for cutting wheels and such. When you need to cut a slot in concrete pavement, you use a diamond cutting wheel.
If you accept payments through a web site without disclosing who you are, you're a criminal.
That's the law in California, or if you sell in Calfornia. California Business and Professions code section 17538 provides that:
(d) A vendor conducting business through the Internet or any other electronic means of communication shall do all of the following when the transaction involves a buyer located in this state:
(1) Before accepting any payment or processing any debit or credit charge or funds transfer, the vendor shall disclose to the buyer in writing or by electronic means of communication, such as e-mail or an on-screen notice, the vendor's return and refund policy, the legal name under which the business is conducted and, except as provided in paragraph (3), the complete street address from which the business is actually conducted. ...
(2) If the disclosure of the vendor's legal name and address information required by this subdivision is made by on-screen notice, all of the following shall apply: (A) The disclosure of the legal name and address information shall appear on any of the following: (i) the first screen displayed when the vendor's electronic site is accessed, (ii) on the screen on which goods or services are first offered, (iii) on the screen on which a buyer may place the order for goods or services, (iv) on the screen on which the buyer may enter payment information, such as a credit card account number, or (v) for nonbrowser-based technologies, in a manner that gives the user a reasonable opportunity to review that information. The communication of that disclosure shall not be structured to be smaller or less legible than the text of the offer of the goods or services.
(g) Any violation of the provisions of this section is a misdemeanor punishable by imprisonment in the county jail not exceeding six months, by a fine not exceeding one thousand dollars ($1,000), or by both that imprisonment and fine.
Yes, running a business anonymously is a criminal offense. You can incorporate, if you like, but you can't just use a psuedonym.
There should be a mechanism for encrypting web data that doesn't rely on paying a third party for a service only tangentially related to encryption.
There is. You can still get cheap SSL certificates. But if you're accepting payments, plan on getting one that clearly identifies who you are.
If you accept payments through a web site without disclosing who you are, you're a criminal. (California Business and Professions code section 17538, other provisions in other jurisdictions.) And soon, browsers are going to put up a big red flag that will make your customers go away.
Has anyone actually been able to find the specification for "high assurance" certificates? Apparently this is being closely held. The spec comes from something called the "CA Browser Forum", which is invitation-only and doesn't seem to have a web site. A standard was supposed to be issued in August, but apparently agreement wasn't reached until a meeting in September. There are many press releases, but no hard data.
So that's why it's not in Mozilla.
It's actually a good idea. Early in the history of SSL, getting a certificate required presenting appropriate business identification info to the certificate issuer. The problem is that some issuers (GoDaddy comes to mind) started issuing "domain only" SSL certificates; the only verification is that the domain can get email. Then, instead of revoking GoDaddy's root certificate for this, the other cert issuers copied GoDaddy's approach. Now anybody can get a meaningless certificate with a meaningless Relying Party Agreement.
The way it's supposed to work is that the certificate issuer bears financial responsibility for misidentification of the certificate owner. Some certificates from Verisign have a Relying Party Agreement that does provide a financial guarantee to the party relying on the certificate - $100 for a class 1 cert, $5000 for a class 2 cert, and $100,000 for a class 3 cert. Most of the other issuers have relying party agreements which promise nothing and deliver less.
So what's happening is that, soon, you'll be able to tell the difference between the crap certificates and the good ones. Before you buy. The idea is that if you put your credit card into a site that showed a green toolbar in IE, and it wasn't really the company it should have been, you can collect from the certificate issuer. This puts certificate issuers on the hook for phishing losses.
Unfortunately, the rules and the Relying Party Agreements for the new certificates haven't yet appeared, so we can't tell if the rules are tough enough to make this work. Since they're being drafted by the certificate issuers, there will probably be some loophole that lets them off the hook.
If Microsoft actually had a microkernel (and NT 3.51 was close), this would be reasonable. At NT 4, a vast amount of crap from Windows 95 was put in the NT kernel for "compatibility", which is how we got to the mess we have today.
QNX has a closed microkenel. It does timers, memory management, CPU dispatching, and interprocess communication. That's it. All drivers, networking, file systems, etc. are outside the kernel as user processes. Nothing other than the stock kernel runs in kernel space. So the kernel changes very little over the years, leading to very good reliability. (It can even be put in ROM, and often is on embedded systems). Sometimes system components outside the kernel break; they crash, an ordinary process core dump is taken, and it can be examined in an ordinary debugger. Drivers are such components. So there's a clear separation between the stuff you don't need to change, and the stuff you might want to change. Even driver developers don't need to see inside the kernel; interaction with the kernel is through the regular user program APIs, with a few extra calls for privileged drivers allowed to map physical memory and access device hardware. Much the same is true of IBM's VM. So this can be done right.
But with Microsoft's bloated kernel, locking down the kernel has much stronger implications. Many functions are in the kernel because Microsoft wants to control them. Far more is there for legacy reasons. (My favorite is the decompresser for .RLE images, which has an exploitable bug.) There's also too much in there because it seems to be needed during the boot process. (One big lesson of OS design: use a boot loader that lets you load both the kernel and various support programs and libraries before the kernel starts. This reduces the temptation to put everything needed at startup in the kernel. Linux now has this, but got it late, so too much went into the Linux kernel too.)
Another issue is the DRM problem, or how to protect DRM code without putting it in the kernel. Microsoft doesn't seem to have a good solution for that. The right answer is probably rings of protection, like Multics or VMS, but Microsoft didn't go that route.
Symantec used to sell compilers, developer tools, and even some user applications like ThinkTank, an early outliner. Microsoft pushed them out of the tools field on Windows; Symantec had a more portable alternative to MFC, and Microsoft didn't like that. Outliners disappeared as a standalone product category; Word now does that. All that's left is the anti-virus business. Now that, too, looks like it's toast.
Actually, the OS vendor should be doing the security system. The primary function of an operating system is security and resource management; everything else could potentially be an application. Only because of Microsoft's appallingly bad security does the anti-virus industry even exist.
While Wales is talking about this, the Internet Archive is quietly doing it and has been for several years now. There's a sizable book scanning effort underway, in cooperation with various libraries and Yahoo. Automatic book scanning equipment has been developed and is in use. The images of the pages are stored, and the text extracted with OCR for searching and other uses. They have over 33,000 books on line now.
Loading existing works into a wiki system isn't particularly useful, anyway. Making them editable probably won't make them better.
The move to PowerPC was Apple's big mistake. That was the point at which Apple market share dropped, and it never came back. Even today, Apple has much lower market share than it did the day the PowerPC machines were announced. The argument for going with the PowerPC was that IBM was going to make Macs. Yes, that was the whole point of the deal. Didn't happen, but that was Apple's big plan. And that bad move happened under Jobs.
In fact, when the PowerPC 601 came out, Motorola was shipping the 68060, which outperformed the early PowerPC chips. The 68000 line could have been developed further; there was nothing in the architecture that limited it. But when Apple dropped it, that was the end of the demand for high-end 68000 parts.
The PowerPC transition killed many existing apps. The engineering community dumped the Mac at the PowerPC transition; existing CAD applications like AutoCAD were not ported to PowerPC, and most of the printed circuit board design applications were dropped at that point, too. So Apple lost a whole market segment, and one willing to pay for big screens and good graphics.
Copeland was actually a good operating system. The problem was that applications had to be revised for it, and Microsoft didn't want to bother. Apple no longer had the clout with developers it had had back at the System 7 transition, where all apps had to be revised. But Apple hadn't realized internally that it could no longer order developers around; the developers had the option of going to Windows. So backwards compatibility had become more important.
Copeland (the original "MacOS 8") actually shipped to some developers. It was almost ready to go. Acquiring NeXT delayed the release of a new OS by several years; it took much longer to get NeXT code onto the Apple platform than Jobs said it would. But it saved Jobs' ass financially; he was heavily invested in NeXT, which was headed for bankruptcy.
As for design, one of the coolest Macs ever was the 20th Anniversary Mac, the first Mac with an LCD panel. In 1997, way ahead of everyone else. That was before Jobs took over and "Steved" the product, because it wasn't his.
The iMac clamshell looked like the Lear-Seigler ADM 3A from 1977, which was a very popular low-end terminal in its day. It wasn't an original concept.
Jobs' big contribution was to suck up to Gates and thus keep Microsoft Office on the Mac That's what saved Apple.
So that's what it looks like with the Reality Distortion Field turned off.
Marginally irrelevant, but good news on spam: Update on Jeanson James Ancheta, botnet spammer. The short version: he's now Federal inmate number 32392-112 at the California City Correctional Institution.
Like I said before, YouTube has the same legal problems Napster did.
If they're not yet doing so, I'd expect the RIAA to start running a song recognition program against YouTube content. That will catch all those videos with commercial music attached.
Let's take a look at the career of last year's big pump-and-dump spammer:
"Computer Virus Broker Arrested for Selling Armies of Infected Computers to Hackers and Spammers
"Pump-and-dump spam domains go silent after botnet closure"
Spammers register pump-and-dump spam domains for use in spam runs. These domains are commonly discarded after a few days. The tactic is commonplace but the the arrest of alleged botmaster Jeanson James Ancheta, 20, of Downey, California, on 3 November has been accompanied by a radical shift in the landscape. "Up to recently, the graphs were all fairly smooth, with the stats showing that 12 days was about the maximum lifetime for this type of domain, while 30 per cent only lasted a day or under, and 10 per cent only lasted three hours or under," Shipp said. "This kind of activity just disappeared completely from the radar on 2 November."
Following up:
"Botnet Creator Pleads Guilty, Faces 25 Years"
Federal Bureau of Prisons Inmate Locator
California City Prison: "This medium security desert prison opened in 2000, and is a stunning sight, either by day when its monolithic forms stand out on the desert pavement like ancient Egyptian architecture, or by night when floodlights bathe the gleaming facility in an orange glow which can be seen from as much as 30 miles away."
Next spammer, please.
This should be reported, in very clear terms, to "enforcement@sec.gov". Or on the SEC's online form. Or to the SEC Division of Enforcement, 100 F Street, N.E. Washington, D.C. 20549. Because it's a felony being committed in support of a pump-and-dump stock scam.
The stock being hyped is "TTEN", which has very low volume. The SEC can find out who was trading it just before the spam run started. That's how to find the people behind this. They can follow the money.
So put together a comprehensive package listing all known stocks being hyped by this thing and the dates the spam began, and ship it off to the SEC. The SEC and FinCen (the U.S. Treasury Financial Crimes Enforcement Network) have the data mining tools to look at the stock transactions and find the people behind this. The SEC has gone after pump-and-dump spammers many times before, and they usually get them.
There's an effort at MIT to develop Mechanically Active Materials.
"Team 2 is developing nanomaterials that are capable of mechanical actuation and dynamic stiffness. As part of the soldier's battlesuit, these adaptive multifunctional materials will improve soldier survivability.
Mechanical actuators embedded as part of a soldier's uniform will allow a transformation from a flexible and compliant material to a non-compliant material that becomes armor, thus protecting the soldier by distributing impact. Soft switchable clothing can also be transformed into a reconfigurable cast that stabilizes an injury such as a broken leg. Contracting materials can be made to apply direct pressure to a wound, function as a tourniquet, or even perform CPR when needed. Mechanical actuators can also be used as exo-muscles for augmentation of a soldier's physical strength or agility and as wound compresses."
There are some polymer gels that change shape when an electric field is applied. And shape-memory alloys, which have been a solution looking for a problem for decades, have already been used for a dress with a changing hemline.
All those little motors and strings are a theatrical effect; that approach is too fragile and bulky.
The leading teams this time are Stanford/Volkswagen, CMU/General Motors, and Oshkosh Truck. The prize doesn't matter to either. And all three already got $1 million in Government money. Each.
It's a much tougher job this time. Driving in traffic, backing out of tight spots, parking and unparking. I'll be surprised if anyone wins the first year, because nobody knows quite what to expect. Year two, someone will win.
I've been expecting this for a while, ever since the transistor count of the GPU passed that of the CPU. Actually, I thought it would happen sooner. It's certainly time. Putting more transistors into a single CPU doesn't help any more, which is why we now have "multicore" machines. So it's time to put more of the computer into a single part.
NVidia already makes the nForce line, the "everything but the CPU" part, with graphics, Ethernet, disk interface, etc. If they stick a CPU in there, they have a whole computer.
Chip designers can license x86 implementations; they don't have to be redesigned from scratch. This isn't going to be a tough job for NVidia.
What we're headed for is the one-chip "value PC", the one that sits on every corporate desk. That's where the best price/performance is.
Of course this happened. YouTube is the next Napster. Same centralized hosting of uploaded content, same business model, same excuses, same legal problems. YouTube is in a worse legal position than Napster. Napster just hosted the index. YouTube hosts the actual content.
YouTube could well be shut down by an injunction. That's what happened to Napster. "Napster is enjoined from copying or assisting or enabling or contributing to the copy or duplication of all copyrighted songs and musical compositions of which the plaintiffs hold rights." -- U.S. District Judge Marilyn Hall Patel.
As for it being the responsibility of the copyright holder to find the material, "Napster wrote the software; it's up to them to write software that will remove from users the ability to copy copyrighted material," -- Judge Patel
YouTube, like Napster, is a contributory infringer. "The district court determined that plaintiffs had demonstrated they would likely succeed in establishing that Napster has a direct financial interest in the infringing activity. We agree. Financial benefit exists where the availability of infringing material "acts as a 'draw' for customers." -- 9th Circuit Court of Appeals.
I was amazed that Google bought YouTube. It was obvious they were buying into a huge litigation problem.
Actually, the Mac IIci went together very well. That was a vertical-assembly machine; everything went in with a straight-down move. Including the power supply, which was just pushed in vertically, and engaged alignment guides on the case and connectors on the motherboard.
Then Apple offshored manufacturing, and design for automated assembly mattered less.