This is a very real issue for numeric work. People who do engineering simulation applications often run compiler validation test suites, to make sure the different versions of compilers get the same answer.
It's not at all uncommon to find bugs in compiler optimizations that way.
This isn't that big a deal. The "slander of title" claim wasn't going anywhere. SCO still has a contract dispute with Novell pending, The odds of SCO winning that case just dropped, but SCO's claims there are not totally bogus. It's a tough contract case. Clearly SCO bought some rights from Novell, but the language of exactly who owns what is ambiguous.
More significantly, SCO did get an extension in the IBM case. Trial has been pushed back to November 2005, and discovery has been extended.
So we have another year and a half of FUD ahead.
There's still Red Hat vs. SCO, SCO vs. AutoZone, and SCO vs. Damlier-Chrysler. By now, corporate Linux users have figured out that there's no reason to pay SCO money until SCO wins all of these cases. Hence the $11,000 total revenue from SCOsource.
The problem the "nuclear weapons establishment" faces is that most of the bomb designers have retired. All the major nuclear weapons R&D was completed before 1970, so everybody with actual experience is gone.
How do you recruit new ones? The job requires a PhD in a specialized area of physics. It requires someone who's willing to devote their life to doing something many people disapprove of, and which is a dead end job as well.
There's an additional problem that's not well known. Existing US bomb designs, especially fusion bomb designs, are a bit too clever. Back when bomb design was an active field, when top people went to Los Alamos and Lawerence Livermore, there was a tendency to overdesign. Pushing for the highest possible yields from the smallest warheads provided useful work for many bright people. But trying to reach that level of cleverness with today's second and third tier people is hard. Especially when you can't test.
Hence the "National Ignition Facility". Part of the idea is to make bomb design without testing easier. And part of the idea is simply to keep physicists busy, so they're around when needed.
There's a paper from 2000 showing the "1 G testbed", which is another air-hockey-puck robot. It doesn't fly; it skates around on a flat surface. That's the "testbed."
According to the 2000 paper, the next step was supposed to be a 1.75x scale model of the 6 degree of freedom flyer, which was to be tested by flying it in the KC-135 aircraft used to train people to operate in zero G. We don't see that mentioned again.
As far as I can tell, the Wired article's reference to "a version of the Personal Satellite Assistant, or PSA, that's fully mobile, with a sensor suite that's nearly space-ready" refers to the 2D hockey-puck robot.
NASA's web site has pictures of non-functioning prototypes, a mockup, and computer graphics. Usually, anything NASA is hyping is accompanied by high-resolution color pictures. Is this thing for real?
In response to the earnings report, the stock has dropped 9% today.
The stock has been around $5 ± 0.50 for weeks. You can't short a NASDAQ listed stock when the price is below $5, so the bears drop out at that point. Trading volume has been down for weeks. SCO's concern with being listed on some additional exchanges may be that they have different rules on short positions, allowing shorting at lower price points.
The full numbers from this quarter aren't out yet, so it's hard to figure out when they run out of cash. It looks like they have three to six quarters of cash left at their current burn rate, but that's a rough estimate. Unless they get new financing, it looks like they won't make it to the IBM trial date in 2005.
At this point, the remaining stockholders who think SCO has a case against IBM might well ask why SCO wants additional delay. SCO can't afford much more delay.
This is a well-known problem in film CG work. The effects industry has already been through this and is coming out the other side. This was a hot topic at SIGGRAPH a few years back.
For the popular version, read Keanu Reeves bitching about what it's like to be an actor who's been thoroughly digitized.
"From the directors, it's like 'Mwa-ha-haaa, now I can control you; I have you!'... And it's designed so you really can't distinguish between the real actor and ones and zeroes. Closeups on faces and emotions coming through, some really remarkable scenes that the audience will never know there's nobody in it.
There's a moment in the very beginning of the movie when Carrie-Anne [Moss]' character dives through the window. You see these two guns go by the camera, and the camera moves past her face. She goes out the window and down, and there is nothing real about that. Nothing. No guns, no girl, no window.
Games are running a few years behind that level of realism. They'll get there.
System administration is a lot like stationary engineering. Stationary engineering, running the mechanical infrastructure of a complex building or plant, was a cool job around 1900 or so. Today, it's a unionized trade. The US has about 120,000 unionized stationary engineers.
It's much like system administration. There are control rooms, and equipment to be monitored and fixed. There's certification, required by law for some types of plants. New equipment comes and goes, and you have to keep up.
Technically, I agree. What we really need are more attributes for the INPUT tag, to allow more validation. Right now, you can't even express things an IBM 3270 dumb terminal could do, like "5 digit field, must be filled".
Today, we'd probably use regular expressions.
Something like this, for a social security number.
The main reason remaining for paper publication in high-priced journals is prestige, in the academic "publish or perish" sense. Many academic journals expect authors and readers to pay them. They don't pay reviewers. Often, they don't even pay editors. Then they have subscription prices upwards of $1000 per year, so only libraries subscribe.
Even big-name journals like Nature seem to be in decline. When Nature publishes articles that aren't about the biological sciences, they range from weak to totally bogus.
A friend who writes for mass-market magazines was once talking to me about journal publication. When I described "page fees", which the author, or the author's institution, pays, she said "That's a vanity press". She's right.
An academic journal is really just a blog with tough editors. Deal with it.
This sounds like an ad for an FPGA. But they don't want to make it too flexible:
The architecture, therefore, must be protected, because that is what establishes how you communicate. So, yes, we have to lock down the architecture so that it cannot be randomly disrupted. Otherwise, imagine that somebody came along and altered the core architecture along with the core instruction set. It would invalidate all the prior intellectual property and destroy the ecosystem.
FPGAs are useful, but very few applications reprogram them while in use. I doubt that's what they intend.
Machines have been built into which you could load application-specific microcode. The DEC VAX 11/780 had that feature, although nobody ever used it. Shader programs in GPUs are probably the most successful example of this idea.
They'll probably export some internal tuning parameters, like the instruction/data ratio for the cache or the depth of instruction prefetch, produce some benchmarks showing that if you tune those performance goes up 20% or so, and hype that into "reconfigurability".
Web site design today needs to eliminate JavaScript, as more people turn it off. It's important that your e-commerce site be able to process a sale without JavaScript. If it can't, you're losing customers.
Turn off JavaScript and try to buy something from your site. If you can't, you have a problem. Yes, you. Not your customer. You, the web designer.
The
SCO® Group, Inc. (Nasdaq: SCOX) will host its second quarter 2004 financial results conference call on Thursday, June 10, 2004, at 9:00 a.m. (MDT), or 11:00 a.m. (EST).
If you would like to participate in the live call you may dial 800-795-1259 or 785-832-0326; confirmation code: 431766.
Listen via webcast here.
RealAudio or Windows Media Player, of course.
We need a few good volunteers in Silicon Valley. No pay, some risk, long hours, we cover all the expenses. We're close to a working vehicle, as can be seen from the pictures on our web site.
Someday, somebody from Netgear is going to have to explain that to a judge and jury. And it's not going to go over well. Once might be considered ordinary negligence. But the second time moves it into the "gross negligence" category: "an act or omission in reckless disregard of the consequences affecting the life or property of another."
There are law firms that handle class actions for negligence.
That firm has already won against Microsoft in another case. They're currently sueing AOL, AT&T, Nextel, and Lucent over various consumer-related claims. So they clearly handle cases like this.
So if you're a victim of Microsoft's negligence in making systems that can easily be converted to attack zombies, click here to contact that law firm. The most effective victims would be those who run Linux, because they're not subject to Microsoft's EULA. For them, it's a pure negligence issue. A Linux-based ISP or hosting service would be the poster child for such an action. They're being hammered on, they didn't sign any Microsoft EULA, and they're clearly suffering sizable damages due to Microsoft's negligence.
There was a more serious effort to do this during the Cold War. There was a set of "how to" microfiche stored in major fallout shelters, so that lost technology could be recovered.
Contain images of child abuse, anywhere in the world.
Contain adult material that potentially breaches the Obscene Publications Act in the UK.
Contain criminally racist material in the UK.
This last is a major issue. It's similar to the "hate speech" issue on college campuses. It would be a great excuse for, say, blocking Aljazeera.
(They have cool anti-American cartoons, in Flash.
Some of them are anti-white-people.)
A cache model in which there is one cache for
instructions and another cache for data is called a
"Harvard-style" cache. This is the model assumed by
the PowerPC Architecture, e.g., in the descriptions of
the Cache Management instructions in Section 3.2,
"Cache Management Instructions" on page 16. Alternative
cache models may be implemented (e.g., a
"combined cache" model, in which a single cache is
used for both instructions and data, or a model in
which there are several levels of caches), but they
support the programming model implied by a
Harvard-style cache.
The processor is not required to maintain copies of
storage locations in the instruction cache consistent
with modifications to those storage locations (e.g.,
modifications caused by Store instructions).
That's the point here. The PowerPC does not guarantee that self-modifying code will work. There's a long discussion of how to do self-modifying code in section 1.8.1 of the architecture manual. This is mainly for debugger support.
Given this instruction/data separation, it's not surprising that
the PowerPC architecture defines a no-execute bit
(it's bit 61 of each page table entry). Unfortunately, not all PowerPC models implement it.
Slashdot Mirror
This is a very real issue for numeric work. People who do engineering simulation applications often run compiler validation test suites, to make sure the different versions of compilers get the same answer. It's not at all uncommon to find bugs in compiler optimizations that way.
More significantly, SCO did get an extension in the IBM case. Trial has been pushed back to November 2005, and discovery has been extended. So we have another year and a half of FUD ahead.
There's still Red Hat vs. SCO, SCO vs. AutoZone, and SCO vs. Damlier-Chrysler. By now, corporate Linux users have figured out that there's no reason to pay SCO money until SCO wins all of these cases. Hence the $11,000 total revenue from SCOsource.
If you want to hear what you want, it's going to cost you.
Yeah, you can listen to indie bands. Click here to listen to bands that suck.
How do you recruit new ones? The job requires a PhD in a specialized area of physics. It requires someone who's willing to devote their life to doing something many people disapprove of, and which is a dead end job as well.
There's an additional problem that's not well known. Existing US bomb designs, especially fusion bomb designs, are a bit too clever. Back when bomb design was an active field, when top people went to Los Alamos and Lawerence Livermore, there was a tendency to overdesign. Pushing for the highest possible yields from the smallest warheads provided useful work for many bright people. But trying to reach that level of cleverness with today's second and third tier people is hard. Especially when you can't test.
Hence the "National Ignition Facility". Part of the idea is to make bomb design without testing easier. And part of the idea is simply to keep physicists busy, so they're around when needed.
Physicists are part of the stockpile, you see.
There's a paper from 2000 showing the "1 G testbed", which is another air-hockey-puck robot. It doesn't fly; it skates around on a flat surface. That's the "testbed."
According to the 2000 paper, the next step was supposed to be a 1.75x scale model of the 6 degree of freedom flyer, which was to be tested by flying it in the KC-135 aircraft used to train people to operate in zero G. We don't see that mentioned again.
As far as I can tell, the Wired article's reference to "a version of the Personal Satellite Assistant, or PSA, that's fully mobile, with a sensor suite that's nearly space-ready" refers to the 2D hockey-puck robot.
NASA's web site has pictures of non-functioning prototypes, a mockup, and computer graphics. Usually, anything NASA is hyping is accompanied by high-resolution color pictures. Is this thing for real?
The stock has been around $5 ± 0.50 for weeks. You can't short a NASDAQ listed stock when the price is below $5, so the bears drop out at that point. Trading volume has been down for weeks. SCO's concern with being listed on some additional exchanges may be that they have different rules on short positions, allowing shorting at lower price points.
The full numbers from this quarter aren't out yet, so it's hard to figure out when they run out of cash. It looks like they have three to six quarters of cash left at their current burn rate, but that's a rough estimate. Unless they get new financing, it looks like they won't make it to the IBM trial date in 2005.
At this point, the remaining stockholders who think SCO has a case against IBM might well ask why SCO wants additional delay. SCO can't afford much more delay.
For the popular version, read Keanu Reeves bitching about what it's like to be an actor who's been thoroughly digitized.
There's a moment in the very beginning of the movie when Carrie-Anne [Moss]' character dives through the window. You see these two guns go by the camera, and the camera moves past her face. She goes out the window and down, and there is nothing real about that. Nothing. No guns, no girl, no window.
Games are running a few years behind that level of realism. They'll get there.
The $5 effect comes from the NASDAQ rule that you can't short stocks below $5. The bears drop out at this point.
It's much like system administration. There are control rooms, and equipment to be monitored and fixed. There's certification, required by law for some types of plants. New equipment comes and goes, and you have to keep up.
That's your future, sysadmins.
Today, we'd probably use regular expressions. Something like this, for a social security number.
-
<INPUT name="social" type="text" format="\d\d\d-\d\d-\d\d\d\d\">
That has plenty of power, enough for things like phone number formats.Even big-name journals like Nature seem to be in decline. When Nature publishes articles that aren't about the biological sciences, they range from weak to totally bogus.
A friend who writes for mass-market magazines was once talking to me about journal publication. When I described "page fees", which the author, or the author's institution, pays, she said "That's a vanity press". She's right.
An academic journal is really just a blog with tough editors. Deal with it.
FPGAs are useful, but very few applications reprogram them while in use. I doubt that's what they intend.
Machines have been built into which you could load application-specific microcode. The DEC VAX 11/780 had that feature, although nobody ever used it. Shader programs in GPUs are probably the most successful example of this idea.
They'll probably export some internal tuning parameters, like the instruction/data ratio for the cache or the depth of instruction prefetch, produce some benchmarks showing that if you tune those performance goes up 20% or so, and hype that into "reconfigurability".
Turn off JavaScript and try to buy something from your site. If you can't, you have a problem. Yes, you. Not your customer. You, the web designer.
-
The
SCO® Group, Inc. (Nasdaq: SCOX) will host its second quarter 2004 financial results conference call on Thursday, June 10, 2004, at 9:00 a.m. (MDT), or 11:00 a.m. (EST).
Listen via webcast here. RealAudio or Windows Media Player, of course.If you would like to participate in the live call you may dial 800-795-1259 or 785-832-0326; confirmation code: 431766.
We need a few good volunteers in Silicon Valley. No pay, some risk, long hours, we cover all the expenses. We're close to a working vehicle, as can be seen from the pictures on our web site.
As usual, I offer a $250 award for information regarding the identity of the person responsible. We can be reached at (650) 326-3529.
Nantero isn't publicly held, though, so this isn't a stock hype.
Someday, somebody from Netgear is going to have to explain that to a judge and jury. And it's not going to go over well. Once might be considered ordinary negligence. But the second time moves it into the "gross negligence" category: "an act or omission in reckless disregard of the consequences affecting the life or property of another."
So if you're a victim of Microsoft's negligence in making systems that can easily be converted to attack zombies, click here to contact that law firm. The most effective victims would be those who run Linux, because they're not subject to Microsoft's EULA. For them, it's a pure negligence issue. A Linux-based ISP or hosting service would be the poster child for such an action. They're being hammered on, they didn't sign any Microsoft EULA, and they're clearly suffering sizable damages due to Microsoft's negligence.
It's time for this to become a major legal issue.
I expect that Google will in time give drastically lower weight to easily-modified pages like "blogs" and "wikis". They're not that hard to recognize.
I'd love to have a copy of those fiche.
Either Google or the Internet Archive would be happy to archive that data for the City of Baltimore and keep it available for public reference.
This last is a major issue. It's similar to the "hate speech" issue on college campuses. It would be a great excuse for, say, blocking Aljazeera. (They have cool anti-American cartoons, in Flash. Some of them are anti-white-people.)
That's the point here. The PowerPC does not guarantee that self-modifying code will work. There's a long discussion of how to do self-modifying code in section 1.8.1 of the architecture manual. This is mainly for debugger support.
Given this instruction/data separation, it's not surprising that the PowerPC architecture defines a no-execute bit (it's bit 61 of each page table entry). Unfortunately, not all PowerPC models implement it.