They're not 0-day, LJ is just really slow in keeping up with these things. I noticed one recently fixed XSS bug was fixed in Horde/IMP over a year ago.
The new cookie system hasn't actually been implemented yet. They've only implemented the first part, which involves changing the subdomain system. The new cookie system will involve cookies for each subdomain and various redirects to ensure that you can only compromise individual journal access at best. Even then, I'd imagine they'd set those subdomain cookies to expire quickly so you'd have to act on them fast.
Ironically, the Wikimedia foundation can't sue them for violating the GFDL. This is due to the copyright being owned by only those individuals who specifically wrote what was plagiarized. It's for this reason that the FSF requires that all GNU projects have copyright ownership transferred over to them; so they can sue on behalf of the content creators.
Apple has briefly surpass Dell by a tiny amount. How does this show that at all? For a major portion of Dell's history it was beating out Apple. If there's any evidence here, it's that Jobs doesn't have the vision to run a company like Dell.
The debunking steve gibson website has actually been widely supported even by anti-microsoft groups because of Steve Gibson's stupid rants about how raw socket support in Windows were the end of the world and should be disabled. This after he got hacked by a 13 year old script kiddie on his windows system. That, and he has a product (SpinRite) which is full of snake oil claims and even does things which are damaging to your data. And IMHO you didn't actually read the website, you just attacked it out of sheer ignorance and wanting to dismiss claims made by anyone attacking someone who criticizes microsoft. You let your anti-Microsoft zealotry blind you. Pretty much everyone in the security community knows who Gibson is, and they almost unaninmously consider him a laughing stock.
Oh and since he's being cited as an expert, it's perfectly valid to attack his credibility as one. He doesn't include his source or testing methodology, so I'm not going to take his "expert" word that he got it right. Remember, this is the same guy who was trying to make a "raw sockets" tester for Windows, thought he had it working, but didn't actually because he didn't know that he needed to bind() the socket first. Raw sockets 101. Someone on his own forums had to teach him the basics.
I want to see the source. Gibson is notorious for making idiotic and ignorant errors in his own code. I remember a forum exchange where he had to be given special instructions on how to correct his code that tested a supposed vulnerability he found with raw sockets. He didn't know he had to call bind() basically. LOL.
Why did they have to name it @Home? All the other projects refer to distributed computing at home. This refers to distributed human work, so it should get a different suffix. Something related to clickworkers, mechanical turk or whatever would do fine.
I should probably point out that the project doesn't actually start until March 1st. You can preregister now though. What's interesting is that this also has tests you need to pass to be able to participate. This is different from NASA's clickworkers project where anyone could just run the java applet without registration and fill things out at will. The overall accuracy for clickworkers was higher than that for trained professionals, but perhaps there's something different with Stardust@Home that requires people to pass a test.
THanks for clarifying that. I was confused as to how it could be new to do any general type of molecular simulation on a computer. Molecules obey completely known laws, so in the worst case scenario you can always do a very computationally intensive atom-for-atom simulation. That could simulate any chemical reaction that didn't involve breaking apart atoms.
Why would air traffic control systems require such frequent reboots? Why can't they rely on existing OSS operating systems which do have very good uptimes? What is the air traffic software doing that's crashing the system?
Pretty sure this is wrong. Ring 0 and ring 3 are the only rings used.
I checked and you're right, I'm not sure how I got that idea. In any case, the protection modes they run in doesn't really make them part of the kernel, just the "kernel space." If they were part of the kernel you'd need the source to write a driver, but that's not the case. Yes, it's bad to have them all running the lowest protection level, but really ring 1 isn't *that* big of an improvement as you can still clobber and crash the system pretty easily (by accident).
Not to expend great effort on features that might never be implemented, but to make potential future implementation at least feasible, and perhaps easy (depending on cost). What period of time elapsed between the hardware availability of "no execute", and the windows implementation? 2 years was it?
Yes, but if you're already developing the "next generation" product, it doesn't make much sense to put the effort into current one from a business perspective (since that would require a non-trivial change). It would be *nice* to have, say, Windows 2xxx or XP to have it added, but from a business perspective it makes more sense to them.
But if MS keeps re-designing the architecture, they must have a reason (other than giving hardware makers new API challenges). Vista's redesign of drivers seems focused on reliability and error reporting.
You can always improve things, I don't think any OS is an exception. They *should* have done this before, but you can say that about any improvements. The fact is, they're adding significant improvements, regardless of whether or not you consider them late.
Windows drivers run in ring 1. I'm not sure about Linux, but I think they do too.
The fact that it took until just recently to implement No Execute protection (in hardware and software) shows that security was not a core concern of Cutler or the others that designed it.
That's because the x86 chips lacked the capability until recently. You can't fault the OS for what the hardware doesn't allow.
The fact that this is the 3rd redesign of the driver architecture in 15 years makes me wonder if they got it "right" this time.
What's wrong with the driver system? NT is and always has been a microkernel design. Drivers were heavily layered around it to provide all the functionality.
But if they have a specific local person in mind, then they could just view Amazon wishlists and other things manually. Old investigative techniques would work better anyway. The only benefit of data mining is to find out about criminals you wouldn't otherwise know about, but if that means getting tens of thousands of people buying books on marijuana in a given city, the data is USELESS. Don't you realize how expensive it would be to just start investigating totally random people because they kinda-sorta-maybe might be criminals?
It's like the way Copernicus swept away the huge complexity of the Ptolemaic astronomer's theory of planetary orbits, all those cycles and epicycles, with the simple and powerful idea of the elliptical orbit. Or how Mendeleev replaced the 18th century's bewildering lists of correlations between chemical properties of substances with the simple and powerful organizational principles of the Periodic Table.
Ah, but it's important to note that neither of those have the prestige of Einstein, not even close. Most lay persons have probably never heard of Mendeleev. You also forgot Newton who himself made several important "Nobel prize worthy" discoveries. I would even go so far as to suggest that Newton was really the father of modern physics (yes--I realize that modern physics typically really refers to 20th century and beyond) because his work was such a change from what was mostly just pseduoscientific nonsense. To the best of my knowledge, astronomy was the only major scientific physics endeavor at the time. This suggests that the prestige has a large social aspect that extends beyond just the genius of one's work and is, in fact, more important than the achievement itself.
Even in my own experience as a theoretician I find the truly brilliant ideas are not complex. They're insights that drastically simplify and clarify. They're the kind of things that, when you understand them, make you slap your head in awe and envy.
But given some of the mind boggling and counter-intuitive phenomenon of quantum mechanics, how do you think it can be all explained with something so simple? Is there any evidence that theories are getting simpler?
Additionally, were pre-GR theories more complex than GR itself? From what I know of the history, the popular failed theories weren't overcompliced monoliths, they were simple ones that were just wrong.
Keep in mind that even GR was wrong and only encompassed limited phenomenon. When it comes to cases of relatively "flat space," SR simplifies to Newton's gravitation equation. The reason stuff like this happens is because Newton's work was for describing a small amount of phenomenon (or phenomenon within limited constraints/assumptions). SR expanded that further, loosening the constraints. SR, of course, is more complex than Newtonian gravitation as a result of expansion. Now we have M-theory which expands the described phenomenon by an order of magnitude and likewise it's an order of magnitude more complicated.
So historically, we see things getting more complex not because people aren't as brilliant, but because past research was only working with constrained/limited phenomenon.
So what do you think of M-theory? Do you really think there is a simpler mathematical way of unifying all of those phenomena?
Of course, this might not be true -- it might instead be the case that the basic structure of the universe is simply too complex for ordinary humans to understand even its principles. But I find this hard to believe (for no logical reason, I admit).
It's not that it's too complex for humans to understand, it's just that it takes more humans to understand than before and it's not something we can really conceptualize/visualize (as a whole) intuitively anymore. However, we can definitely fit more mathematics to fit the data. We are getting to the point where it's just too complex to imagine it all in our heads, so we either restrict ourselves to conceptual models for specific aspects of a greater phenomenon OR we just like the mathematics do the work, not bothering to conceptualize what's going on.
He only received a nobel prize for one of those, so your statement is subjective rubbish. AS others have stated, physics has gotten much more complicated over time, so we don't see discoveries nearly that revolutionary come from a single person. Anything radically new will be orders of magnitude more complicated than any of his work was. Your statement suggests that you've never actually studied GR, you probably just read some summaries in lay person books.
I have taken GR as an undergraduate and I can tell you that it's pretty much standard to learn and physics graduates tend to understand it pretty well. On othe other hand, if you look at new theories like M-theory, you'll notice that, as other posters stated, you can get your PhD just trying to analyze one aspect of it. The complexity of scientific knowledge increases exponentially over time. This means that you get PhDs in sub-sub-specialities now.
This is, of course, ignoring all past contributions of geniuses. Someone like you who obviously only has a lay understanding of the subject is completely unaware of past geniuses of the 19th century like Maxwell, Gauss and Farady who all made important contributions. They did "nobel prize worthy" work, but were never recognized as such because it was before the Nobel Prize existed. This is really a problem with lay understandings, they just hear about how great Einstein was then ignore all the other great geniuses in history who students in physics and engineering subjects learn about.
Just look at the long list of Nobel Prize winners and nominees now. It's so big that I doubt most scientists even know who half of them are. They encompass so many disciplines.
Right, because we all know how effective JCP is at addressing consumer comlaints. Let's ignore how many eons it took to get generics implemented in Java. Java still doesn't have operator overloading to this day, despite furious insistence.
And Microsoft may have signed a contract, but that doesn't change the fact that Sun will sue anyone regardless of whether or not they signed it. Their licensing agreement (which no one actually signs) forbids anyone from doing what Microsoft did.
Are you against trademarks and their protection? And unlike your uninformed guess, Microsoft do not let you call things.Net
You misread that page, it's saying you can't name your product something.NET. It's not saying you can't say your product is.NET compliant, which is something Sun forbids with Java. In fact, that very page gives recommendations of how to say that your product uses.NET.
Note that Mono calls itself a.NET implementation, even adding its own APIs, and Microsoft has no problem with that because they don't name themselves Mono.NET. Not that Microsoft enforces that anyway, there are a million products that use.NET in their name and to my knowlege, not one has been sued.
Sun has already had to give in to other members of the JCP over technology. Again you show yourself to be uninformed.
When? Over what?
Unless you count RFCs like 1014 (XDR), 1057 (RPC), 1094/1813 (NFS), UltraSparc...
Way to take my statement out of context. I was talking about _Java related_ standards. Microsoft has C# and the CLR as open standards. No parts of Java are part of an open standard.
Witness Microsoft's non-compliance to the C++ specification when it comes to writing Windows apps in Visual Studio for instance.
Uh, have you even used Visual Studio? Are you talking about compiler specific additions? Because all compilers do that, including the ever popular gcc. In fact, there's a de facto standard directive (#pragma) for compiler specific flags even.
If you're referring to slightly imperfect compliance, well there are no compilers that are that good. Comeau and EDG had the earliest compilers that implemented most of the C++ standard. Those were pretty much the only two that could have been considered 'mostly compliant' around 1998 when the final standard was released.
gcc itself didn't support most of it until it was into the 3.x series. 2.95.x had serious compliance problems. Visual Studio had good support since 7.0, as compared to 6.0 which was released before the C++ standard was even finalized.
What is less open about the JCP than ECMA? ECMA is just a rubber-stamp organization, there is no spec development going on there - that all happens at Microsoft and it's associates like HP.
Perhaps that the JCP is run by Sun and EMCA isn't affiliated with any organization (including Microsoft)? It's nice to see that you acknowledge that it's a collaborative effort with other organizations and that it's not just closed off like Sun. oh wait, you didn't know that open standards are usually developed by representatives of large corporations? Maybe you should take a look through the authors of RFCs.
So now suddenly vendor-provided API documentation is enough to make something standardized? What happened to the requirements you leveled at Sun and the JCP?
Of course it is. I don't think you know what "standard" means. Having something well documented and specified makes it standardized. I never said Sun didn't have standards, I just said they were CLOSED standards. Try working on that reading comprehension.
Of course they sued! Microsoft broke a CONTRACT! They had SIGNED a CONTRACT saying they should NOT make incompatible changes to their implementation and they DID! What does that have to do with Java vs. C#?
What contract did they sign? You don't understand; it's a legal mandate by Sun. You don't have to sign anything. If I wrote my own Microsoft-like implementation and it became popular, Sun would sue my ass too, despite me not signing any contract.
And what does this have to do with Java vs. C#? Well maybe you should ask the guy who I'm replying to which started this all. He was asserting that C# was bad because Microsoft would probably sue people. Since Microsoft has sued ZERO and Sun has sued SEVERAL, I'd say that argument holds no water.
No, they do not. They control who can call their Java VM's "Java compliant."
Right, so you can't actually call it Java, you have to come up with some other name for it, which is stupid. Even if it's backwards compatible with Java, you can't call it "Java compliant," which is stupid. Strangely, Microsoft doesn't place this restriction on.NET. Looks like Microsoft is more leanient.
No, they are not "closely controlled" - they collaborate extensively through the JCP, and indeed much of Java 5 was defined that way. OS developers like the Apache Group are behind many of the new parts of the spec...
JCP is a joke. Sun has the ultimate and final say and they have so many purists working for them that they refuse to listen to anything that's not in line with their personal views. It took many years of *numerous* complaints before they finally implemented template (generics) support. So the lesson is: you don't get large changes in it until so many people start complaining that they risk losing market share if they don't..NET, on the other hand, is constantly being improved. With 2.0 out and 3.0 coming out in a few years, they're very open to improvement.
Microsoft actually listens to customer demands because they want to people to use their technology. They don't have a joke like JCP that feigns interest in what their customers say.
Which their license explicitly grants you the use of, when implementing JVMs...
Same goes for Microsoft's license, what's your point? You were *just* whining about microsoft granting royalty free licenses, this is just plain hypocritical.
Unlike MS, who has the option of potentially GPL-incompatible RAND licensing for C# and the CLR, and doesn't standardize the rest of.NET APIs at all... what are you smoking, really?
They *are* standardized, you are just flat out lying now. A proprietary standard is still a standard. MSDN details the full standard. And there's no more reason to believe RAND licensing is incompatible with the GPL than Java's is. Thank Mr. Internets legal expert. As I said: Microsoft has sued ZERO people over.NET, Sun has sued several over JAva. Who do we trust not to bring lawsuits?
Now, let's ignore the fact that, AS YOU IGNORED, Sun has released ZERO open standards. They are all closed. Swing and miss.
You are basically saying you can take your.NET app to some other server? Or ever will be able to, at some point in the future? Lies this obvious just make you look foolish.
Yes, of course. Given the numerous people who have done this without problem, your argument holds no water.
Except all those class not found errors.... Oops, your API didn't come with you... I guess you don't mind rewriting everything?
WTF are you smoking? If you wrote for.NET 1.0, it will work on any other.NET 1.0 system. And newer versions are backwards compatible. This isn't any different with Java. If I write something that uses Java 1.5 features and APIs, it will break on a 1.4 or earlier system.
What are you smoking? Seriously? No APIs have been dropped yet. They have added new, never removed the old...
I said dropped *support*. They won't bother updating them or providing any help with them anymore. Microsoft hasn't dropped any APIs in.NET either.
I notice you ignored the point about the spec and the standard and the VM. Not surprising.
What about them? The specifications for Fortran and COBOL have been around for eons. That didn't prevent them from being phased out though.
Oh and following your own logic, Java isn't standardized, because their standards aren't open. Non-open standards aren't standards in your mind.
So you are aware this is a blatant lie, right?
Oh it's OK. Just point me to the other vendor that implements
There is no future in C#, because it's Microsoft's toy, and it will always be Microsoft's toy.
Wild, rampant and unsupported speculation.
If they want they can take it and go home.
Sun can do this to an even greater extent, since they control who is allowed to implement VMs. Their standards are 100% proprietary and closely controlled. They also protect themselves with patents. Strangely, Microsoft actually has open standards, giving it a distinct advantage. Sun is in a much better position to lock Java down and prevent people from using it.
With Java you can take your code anywhere.
Same with C#.
Basically this adds up critical mass. The language is never going away.
Sort of like Sun's constantly changing GUI APIs which become very popular then have support for them dropped, right?
Basically this adds up critical mass. The language is never going away.
Like COBOL and Fortran, right? Because we know how popular and well supported those languages are today despite being extremely popular in their heyday.
Java is well specified and unencumbered. Even the source of Sun's VM is available (though not under the GPL, at least you can read it, see what's going on in the VM, and fix bugs),
C#/.NET is also well specified and even more unencumbered. Java uses completely closed standards. You keep whining that only 2 of 3 parts of C#/.NET are part of open standards, but Java has 0 open standards. C#/.NET wins hands down.
And Sun's VM is not open source. Have fun with that, especially when reading means that you can't take the risk of coding your own VM, because of potential copyright infringement.
there are Gnu implementations that are farther along already than Mono - and I doubt Mono will catch up.
No, it's quite the opposite, the GNU classpath implementations have only recently come to a half-decent level of completeness. We also know that the open source attempts at VMs have been abysmal. Hell, even commercial alternatives to Java's tend to be a pain to set up and suck donkey balls.
What's that thing called? Oh yeah, evidence. When you make bold claims, you better back them up.
Based purely on raw numbers of job offers, if you're looking to make money off this skill you would be flipping crazy to learn C#... although OTOH once you know one, the other won't be too difficult.
Let's see these numbers then. Evidence.
In many cases the supposed advantages of C# are a wash or even bad ideas - such as their pointless and absurd practice of mixing VM and non-VM code at every opportunity, and allowing unsafe code to be mixed in...
Unsafe code mixing is RARE and requires you specifically telling the compiler to allow it. But hey, don't let things like facts get in your way.
C# people claim their runtime is language agnostic. It is not. It's C* agnostic. Any language significantly different from a C/C++/Java-like language can't be supported efficiently.
What's this hypothesis based on? Evidence helps, it's that thing you keep refusing to provide. So please tell me why Lisp couldn't be implemented effeciently in it. Do you even know how the runtimes worked? They were designed specifically to be LANGUAGE agnostic, not C* agnostic. That's why they are stack based instead of register based. If they just wanted to work as effeciently as possible for C* languages, it would have been made register based, since that would be more effecient and easier to implement.
Perhaps their best path will be to stop trying to be compatible and diverge into a kind of "dirty.NET"... All fun and games until MS sues them.
Sun has already sued people, including Microsoft, for making altered versions of their APIs. So far, Microsoft has sued no one for that. I guess MS wins.
And if you dismiss this as a conspiracy theory... and go to embrace the patented, "standardized" platform
This interface adds more information than Wikipedia does, so you can't link it up. The best you could do is a simplistic modification of Google Maps (or something similar) and just tag individual points as linked to articles. In fact, someone has already done that, but it's not that useful. This adds all kinds of capabilities on top of simple geographic point links.
And this is all the icing on the cake to an already good quality controlled encyclopedia. I have no idea why it would be something "they put on their CDs," that makes no sense at all. If their articles are written by experts, it doesn't matter what medium it's presented in in regards to article quality.
I get sick of Slashdotted not RTFAing. Digital Universe doesn't open up until next year. They have had experts working on it for nearly a year now, but it's not been disclosed to the public.
No, it's not basically like Wikipedia. I really get sick of Slashdotters not reading the article or doing minimal research. This new encyclopedia, aside from having much better funding (and being self-sustaining financially) has a novel interface.
They have this "ManyOne" globe browser deal that basically lets you move all over the globe (like in Google Earth) and select arbitrary portions at arbitrary times in history and have it show you geographical/visual data, as well as an article that's relevent to it.
You missed the point. Those arguments result in compromising on what is NPOV. This means that the zealots' views get disproportionately represented, especially if they're a very vocal fringe minority. Not just that, but POV phrasing ends up getting inserted because they are forced to compromise on how to phrase and organize the article.
After GEM, which very obviously copied Apple, there were no new implementations of the concept until Konfabulator came along.
The fact that it was so long afterward should tell you that people had forgotten about Desktop Accessories. Most people, even MAc users, had no idea what they were until someone brought it up again in this context.
The technology, until a year or two ago when Konfabulator was ported to Windows, never existed for Windows. I've never seen it implemented in X11.
But there are similar things for X11. Even earlier window managers included very basic "mini apps" like a little clock and that moving eyes thing, and a calculator. These came along with Twm, which was created in 1987. They may even predate it, but I'm not sure. Of course, those are too liberal to fit any reasonable definition of something "widget like."
What Konfabulator brought to the table was HTML and Javascript. Beyond that, the apps it supports are barely different in functionality from those Mac OS 1 supported.
Oh, but it's the underlying architecture that makes it what it is. "Difference in functionality" is m eaningless without some unified framework to tie it all together. How those widgets interoperate with eachother is exactly what makes them widgets. A simple bunch of independent mini apps don't really qualify, as those could always be written in any language for any OS without the need for a special "widget engine."
As per your own description, you're relying on a very liberal interpretation of what it means to be a "widget engine," which suggests that it wasn't a rip off at all. They were nothing more than a hack to allow shitty multitasking in a non-multitasking OS. So in essence, it's just a multitasking driver. Other than the API necessary to multi-task, there is no unified API. You could easily make the programs look like whatever, there were no technical restricitons or designs in place to enforce it (beyond social mandate). So all they really are is independent, multitasking programs intended to look vaguely similar.
Slashdot Mirror
They're not 0-day, LJ is just really slow in keeping up with these things. I noticed one recently fixed XSS bug was fixed in Horde/IMP over a year ago.
The new cookie system hasn't actually been implemented yet. They've only implemented the first part, which involves changing the subdomain system. The new cookie system will involve cookies for each subdomain and various redirects to ensure that you can only compromise individual journal access at best. Even then, I'd imagine they'd set those subdomain cookies to expire quickly so you'd have to act on them fast.
Javascript and related features are disabled by default. The issue is that they were able to evade the Javascript filters.
Ironically, the Wikimedia foundation can't sue them for violating the GFDL. This is due to the copyright being owned by only those individuals who specifically wrote what was plagiarized. It's for this reason that the FSF requires that all GNU projects have copyright ownership transferred over to them; so they can sue on behalf of the content creators.
Apple has briefly surpass Dell by a tiny amount. How does this show that at all? For a major portion of Dell's history it was beating out Apple. If there's any evidence here, it's that Jobs doesn't have the vision to run a company like Dell.
The debunking steve gibson website has actually been widely supported even by anti-microsoft groups because of Steve Gibson's stupid rants about how raw socket support in Windows were the end of the world and should be disabled. This after he got hacked by a 13 year old script kiddie on his windows system. That, and he has a product (SpinRite) which is full of snake oil claims and even does things which are damaging to your data. And IMHO you didn't actually read the website, you just attacked it out of sheer ignorance and wanting to dismiss claims made by anyone attacking someone who criticizes microsoft. You let your anti-Microsoft zealotry blind you. Pretty much everyone in the security community knows who Gibson is, and they almost unaninmously consider him a laughing stock.
Oh and since he's being cited as an expert, it's perfectly valid to attack his credibility as one. He doesn't include his source or testing methodology, so I'm not going to take his "expert" word that he got it right. Remember, this is the same guy who was trying to make a "raw sockets" tester for Windows, thought he had it working, but didn't actually because he didn't know that he needed to bind() the socket first. Raw sockets 101. Someone on his own forums had to teach him the basics.
I want to see the source. Gibson is notorious for making idiotic and ignorant errors in his own code. I remember a forum exchange where he had to be given special instructions on how to correct his code that tested a supposed vulnerability he found with raw sockets. He didn't know he had to call bind() basically. LOL.
Why did they have to name it @Home? All the other projects refer to distributed computing at home. This refers to distributed human work, so it should get a different suffix. Something related to clickworkers, mechanical turk or whatever would do fine.
I should probably point out that the project doesn't actually start until March 1st. You can preregister now though. What's interesting is that this also has tests you need to pass to be able to participate. This is different from NASA's clickworkers project where anyone could just run the java applet without registration and fill things out at will. The overall accuracy for clickworkers was higher than that for trained professionals, but perhaps there's something different with Stardust@Home that requires people to pass a test.
THanks for clarifying that. I was confused as to how it could be new to do any general type of molecular simulation on a computer. Molecules obey completely known laws, so in the worst case scenario you can always do a very computationally intensive atom-for-atom simulation. That could simulate any chemical reaction that didn't involve breaking apart atoms.
Why would air traffic control systems require such frequent reboots? Why can't they rely on existing OSS operating systems which do have very good uptimes? What is the air traffic software doing that's crashing the system?
Pretty sure this is wrong. Ring 0 and ring 3 are the only rings used.
I checked and you're right, I'm not sure how I got that idea. In any case, the protection modes they run in doesn't really make them part of the kernel, just the "kernel space." If they were part of the kernel you'd need the source to write a driver, but that's not the case. Yes, it's bad to have them all running the lowest protection level, but really ring 1 isn't *that* big of an improvement as you can still clobber and crash the system pretty easily (by accident).
Not to expend great effort on features that might never be implemented, but to make potential future implementation at least feasible, and perhaps easy (depending on cost). What period of time elapsed between the hardware availability of "no execute", and the windows implementation? 2 years was it?
Yes, but if you're already developing the "next generation" product, it doesn't make much sense to put the effort into current one from a business perspective (since that would require a non-trivial change). It would be *nice* to have, say, Windows 2xxx or XP to have it added, but from a business perspective it makes more sense to them.
But if MS keeps re-designing the architecture, they must have a reason (other than giving hardware makers new API challenges). Vista's redesign of drivers seems focused on reliability and error reporting.
You can always improve things, I don't think any OS is an exception. They *should* have done this before, but you can say that about any improvements. The fact is, they're adding significant improvements, regardless of whether or not you consider them late.
Windows and Linux only use rings 0 and 3.
Windows drivers run in ring 1. I'm not sure about Linux, but I think they do too.
The fact that it took until just recently to implement No Execute protection (in hardware and software) shows that security was not a core concern of Cutler or the others that designed it.
That's because the x86 chips lacked the capability until recently. You can't fault the OS for what the hardware doesn't allow.
The fact that this is the 3rd redesign of the driver architecture in 15 years makes me wonder if they got it "right" this time.
What's wrong with the driver system? NT is and always has been a microkernel design. Drivers were heavily layered around it to provide all the functionality.
But if they have a specific local person in mind, then they could just view Amazon wishlists and other things manually. Old investigative techniques would work better anyway. The only benefit of data mining is to find out about criminals you wouldn't otherwise know about, but if that means getting tens of thousands of people buying books on marijuana in a given city, the data is USELESS. Don't you realize how expensive it would be to just start investigating totally random people because they kinda-sorta-maybe might be criminals?
It's like the way Copernicus swept away the huge complexity of the Ptolemaic astronomer's theory of planetary orbits, all those cycles and epicycles, with the simple and powerful idea of the elliptical orbit. Or how Mendeleev replaced the 18th century's bewildering lists of correlations between chemical properties of substances with the simple and powerful organizational principles of the Periodic Table.
Ah, but it's important to note that neither of those have the prestige of Einstein, not even close. Most lay persons have probably never heard of Mendeleev. You also forgot Newton who himself made several important "Nobel prize worthy" discoveries. I would even go so far as to suggest that Newton was really the father of modern physics (yes--I realize that modern physics typically really refers to 20th century and beyond) because his work was such a change from what was mostly just pseduoscientific nonsense. To the best of my knowledge, astronomy was the only major scientific physics endeavor at the time. This suggests that the prestige has a large social aspect that extends beyond just the genius of one's work and is, in fact, more important than the achievement itself.
Even in my own experience as a theoretician I find the truly brilliant ideas are not complex. They're insights that drastically simplify and clarify. They're the kind of things that, when you understand them, make you slap your head in awe and envy.
But given some of the mind boggling and counter-intuitive phenomenon of quantum mechanics, how do you think it can be all explained with something so simple? Is there any evidence that theories are getting simpler?
Additionally, were pre-GR theories more complex than GR itself? From what I know of the history, the popular failed theories weren't overcompliced monoliths, they were simple ones that were just wrong.
Keep in mind that even GR was wrong and only encompassed limited phenomenon. When it comes to cases of relatively "flat space," SR simplifies to Newton's gravitation equation. The reason stuff like this happens is because Newton's work was for describing a small amount of phenomenon (or phenomenon within limited constraints/assumptions). SR expanded that further, loosening the constraints. SR, of course, is more complex than Newtonian gravitation as a result of expansion. Now we have M-theory which expands the described phenomenon by an order of magnitude and likewise it's an order of magnitude more complicated.
So historically, we see things getting more complex not because people aren't as brilliant, but because past research was only working with constrained/limited phenomenon.
So what do you think of M-theory? Do you really think there is a simpler mathematical way of unifying all of those phenomena?
Of course, this might not be true -- it might instead be the case that the basic structure of the universe is simply too complex for ordinary humans to understand even its principles. But I find this hard to believe (for no logical reason, I admit).
It's not that it's too complex for humans to understand, it's just that it takes more humans to understand than before and it's not something we can really conceptualize/visualize (as a whole) intuitively anymore. However, we can definitely fit more mathematics to fit the data. We are getting to the point where it's just too complex to imagine it all in our heads, so we either restrict ourselves to conceptual models for specific aspects of a greater phenomenon OR we just like the mathematics do the work, not bothering to conceptualize what's going on.
He only received a nobel prize for one of those, so your statement is subjective rubbish. AS others have stated, physics has gotten much more complicated over time, so we don't see discoveries nearly that revolutionary come from a single person. Anything radically new will be orders of magnitude more complicated than any of his work was. Your statement suggests that you've never actually studied GR, you probably just read some summaries in lay person books.
I have taken GR as an undergraduate and I can tell you that it's pretty much standard to learn and physics graduates tend to understand it pretty well. On othe other hand, if you look at new theories like M-theory, you'll notice that, as other posters stated, you can get your PhD just trying to analyze one aspect of it. The complexity of scientific knowledge increases exponentially over time. This means that you get PhDs in sub-sub-specialities now.
This is, of course, ignoring all past contributions of geniuses. Someone like you who obviously only has a lay understanding of the subject is completely unaware of past geniuses of the 19th century like Maxwell, Gauss and Farady who all made important contributions. They did "nobel prize worthy" work, but were never recognized as such because it was before the Nobel Prize existed. This is really a problem with lay understandings, they just hear about how great Einstein was then ignore all the other great geniuses in history who students in physics and engineering subjects learn about.
Just look at the long list of Nobel Prize winners and nominees now. It's so big that I doubt most scientists even know who half of them are. They encompass so many disciplines.
Right, because we all know how effective JCP is at addressing consumer comlaints. Let's ignore how many eons it took to get generics implemented in Java. Java still doesn't have operator overloading to this day, despite furious insistence.
And Microsoft may have signed a contract, but that doesn't change the fact that Sun will sue anyone regardless of whether or not they signed it. Their licensing agreement (which no one actually signs) forbids anyone from doing what Microsoft did.
Are you against trademarks and their protection? And unlike your uninformed guess, Microsoft do not let you call things .Net
.NET compliant, which is something Sun forbids with Java. In fact, that very page gives recommendations of how to say that your product uses .NET.
.NET implementation, even adding its own APIs, and Microsoft has no problem with that because they don't name themselves Mono.NET. Not that Microsoft enforces that anyway, there are a million products that use .NET in their name and to my knowlege, not one has been sued.
You misread that page, it's saying you can't name your product something.NET. It's not saying you can't say your product is
Note that Mono calls itself a
Sun has already had to give in to other members of the JCP over technology. Again you show yourself to be uninformed.
When? Over what?
Unless you count RFCs like 1014 (XDR), 1057 (RPC), 1094/1813 (NFS), UltraSparc...
Way to take my statement out of context. I was talking about _Java related_ standards. Microsoft has C# and the CLR as open standards. No parts of Java are part of an open standard.
Witness Microsoft's non-compliance to the C++ specification when it comes to writing Windows apps in Visual Studio for instance.
Uh, have you even used Visual Studio? Are you talking about compiler specific additions? Because all compilers do that, including the ever popular gcc. In fact, there's a de facto standard directive (#pragma) for compiler specific flags even.
If you're referring to slightly imperfect compliance, well there are no compilers that are that good. Comeau and EDG had the earliest compilers that implemented most of the C++ standard. Those were pretty much the only two that could have been considered 'mostly compliant' around 1998 when the final standard was released.
gcc itself didn't support most of it until it was into the 3.x series. 2.95.x had serious compliance problems. Visual Studio had good support since 7.0, as compared to 6.0 which was released before the C++ standard was even finalized.
What is less open about the JCP than ECMA? ECMA is just a rubber-stamp organization, there is no spec development going on there - that all happens at Microsoft and it's associates like HP.
Perhaps that the JCP is run by Sun and EMCA isn't affiliated with any organization (including Microsoft)? It's nice to see that you acknowledge that it's a collaborative effort with other organizations and that it's not just closed off like Sun. oh wait, you didn't know that open standards are usually developed by representatives of large corporations? Maybe you should take a look through the authors of RFCs.
So now suddenly vendor-provided API documentation is enough to make something standardized? What happened to the requirements you leveled at Sun and the JCP?
Of course it is. I don't think you know what "standard" means. Having something well documented and specified makes it standardized. I never said Sun didn't have standards, I just said they were CLOSED standards. Try working on that reading comprehension.
Of course they sued! Microsoft broke a CONTRACT! They had SIGNED a CONTRACT saying they should NOT make incompatible changes to their implementation and they DID! What does that have to do with Java vs. C#?
What contract did they sign? You don't understand; it's a legal mandate by Sun. You don't have to sign anything. If I wrote my own Microsoft-like implementation and it became popular, Sun would sue my ass too, despite me not signing any contract.
And what does this have to do with Java vs. C#? Well maybe you should ask the guy who I'm replying to which started this all. He was asserting that C# was bad because Microsoft would probably sue people. Since Microsoft has sued ZERO and Sun has sued SEVERAL, I'd say that argument holds no water.
No, they do not. They control who can call their Java VM's "Java compliant."
.NET. Looks like Microsoft is more leanient.
.NET, on the other hand, is constantly being improved. With 2.0 out and 3.0 coming out in a few years, they're very open to improvement.
.NET APIs at all... what are you smoking, really?
.NET, Sun has sued several over JAva. Who do we trust not to bring lawsuits?
.NET app to some other server? Or ever will be able to, at some point in the future? Lies this obvious just make you look foolish.
.NET 1.0, it will work on any other .NET 1.0 system. And newer versions are backwards compatible. This isn't any different with Java. If I write something that uses Java 1.5 features and APIs, it will break on a 1.4 or earlier system.
.NET either.
Right, so you can't actually call it Java, you have to come up with some other name for it, which is stupid. Even if it's backwards compatible with Java, you can't call it "Java compliant," which is stupid. Strangely, Microsoft doesn't place this restriction on
No, they are not "closely controlled" - they collaborate extensively through the JCP, and indeed much of Java 5 was defined that way. OS developers like the Apache Group are behind many of the new parts of the spec...
JCP is a joke. Sun has the ultimate and final say and they have so many purists working for them that they refuse to listen to anything that's not in line with their personal views. It took many years of *numerous* complaints before they finally implemented template (generics) support. So the lesson is: you don't get large changes in it until so many people start complaining that they risk losing market share if they don't.
Microsoft actually listens to customer demands because they want to people to use their technology. They don't have a joke like JCP that feigns interest in what their customers say.
Which their license explicitly grants you the use of, when implementing JVMs...
Same goes for Microsoft's license, what's your point? You were *just* whining about microsoft granting royalty free licenses, this is just plain hypocritical.
Unlike MS, who has the option of potentially GPL-incompatible RAND licensing for C# and the CLR, and doesn't standardize the rest of
They *are* standardized, you are just flat out lying now. A proprietary standard is still a standard. MSDN details the full standard. And there's no more reason to believe RAND licensing is incompatible with the GPL than Java's is. Thank Mr. Internets legal expert. As I said: Microsoft has sued ZERO people over
Now, let's ignore the fact that, AS YOU IGNORED, Sun has released ZERO open standards. They are all closed. Swing and miss.
You are basically saying you can take your
Yes, of course. Given the numerous people who have done this without problem, your argument holds no water.
Except all those class not found errors.... Oops, your API didn't come with you... I guess you don't mind rewriting everything?
WTF are you smoking? If you wrote for
What are you smoking? Seriously? No APIs have been dropped yet. They have added new, never removed the old...
I said dropped *support*. They won't bother updating them or providing any help with them anymore. Microsoft hasn't dropped any APIs in
I notice you ignored the point about the spec and the standard and the VM. Not surprising.
What about them? The specifications for Fortran and COBOL have been around for eons. That didn't prevent them from being phased out though.
Oh and following your own logic, Java isn't standardized, because their standards aren't open. Non-open standards aren't standards in your mind.
So you are aware this is a blatant lie, right?
Oh it's OK. Just point me to the other vendor that implements
There is no future in C#, because it's Microsoft's toy, and it will always be Microsoft's toy.
.NET"... All fun and games until MS sues them.
Wild, rampant and unsupported speculation.
If they want they can take it and go home.
Sun can do this to an even greater extent, since they control who is allowed to implement VMs. Their standards are 100% proprietary and closely controlled. They also protect themselves with patents. Strangely, Microsoft actually has open standards, giving it a distinct advantage. Sun is in a much better position to lock Java down and prevent people from using it.
With Java you can take your code anywhere.
Same with C#.
Basically this adds up critical mass. The language is never going away.
Sort of like Sun's constantly changing GUI APIs which become very popular then have support for them dropped, right?
Basically this adds up critical mass. The language is never going away.
Like COBOL and Fortran, right? Because we know how popular and well supported those languages are today despite being extremely popular in their heyday.
Java is well specified and unencumbered. Even the source of Sun's VM is available (though not under the GPL, at least you can read it, see what's going on in the VM, and fix bugs),
C#/.NET is also well specified and even more unencumbered. Java uses completely closed standards. You keep whining that only 2 of 3 parts of C#/.NET are part of open standards, but Java has 0 open standards. C#/.NET wins hands down.
And Sun's VM is not open source. Have fun with that, especially when reading means that you can't take the risk of coding your own VM, because of potential copyright infringement.
there are Gnu implementations that are farther along already than Mono - and I doubt Mono will catch up.
No, it's quite the opposite, the GNU classpath implementations have only recently come to a half-decent level of completeness. We also know that the open source attempts at VMs have been abysmal. Hell, even commercial alternatives to Java's tend to be a pain to set up and suck donkey balls.
What's that thing called? Oh yeah, evidence. When you make bold claims, you better back them up.
Based purely on raw numbers of job offers, if you're looking to make money off this skill you would be flipping crazy to learn C#... although OTOH once you know one, the other won't be too difficult.
Let's see these numbers then. Evidence.
In many cases the supposed advantages of C# are a wash or even bad ideas - such as their pointless and absurd practice of mixing VM and non-VM code at every opportunity, and allowing unsafe code to be mixed in...
Unsafe code mixing is RARE and requires you specifically telling the compiler to allow it. But hey, don't let things like facts get in your way.
C# people claim their runtime is language agnostic. It is not. It's C* agnostic. Any language significantly different from a C/C++/Java-like language can't be supported efficiently.
What's this hypothesis based on? Evidence helps, it's that thing you keep refusing to provide. So please tell me why Lisp couldn't be implemented effeciently in it. Do you even know how the runtimes worked? They were designed specifically to be LANGUAGE agnostic, not C* agnostic. That's why they are stack based instead of register based. If they just wanted to work as effeciently as possible for C* languages, it would have been made register based, since that would be more effecient and easier to implement.
Perhaps their best path will be to stop trying to be compatible and diverge into a kind of "dirty
Sun has already sued people, including Microsoft, for making altered versions of their APIs. So far, Microsoft has sued no one for that. I guess MS wins.
And if you dismiss this as a conspiracy theory... and go to embrace the patented, "standardized" platform
This interface adds more information than Wikipedia does, so you can't link it up. The best you could do is a simplistic modification of Google Maps (or something similar) and just tag individual points as linked to articles. In fact, someone has already done that, but it's not that useful. This adds all kinds of capabilities on top of simple geographic point links.
And this is all the icing on the cake to an already good quality controlled encyclopedia. I have no idea why it would be something "they put on their CDs," that makes no sense at all. If their articles are written by experts, it doesn't matter what medium it's presented in in regards to article quality.
I get sick of Slashdotted not RTFAing. Digital Universe doesn't open up until next year. They have had experts working on it for nearly a year now, but it's not been disclosed to the public.
No, it's not basically like Wikipedia. I really get sick of Slashdotters not reading the article or doing minimal research. This new encyclopedia, aside from having much better funding (and being self-sustaining financially) has a novel interface.
o v&b=high
They have this "ManyOne" globe browser deal that basically lets you move all over the globe (like in Google Earth) and select arbitrary portions at arbitrary times in history and have it show you geographical/visual data, as well as an article that's relevent to it.
You can check out a movie of the browser here: http://www.manyone.net/player_qt.html?m=preview.m
It's really cool. I had actually thought of something like it before, but didn't have the time to implement it.
You missed the point. Those arguments result in compromising on what is NPOV. This means that the zealots' views get disproportionately represented, especially if they're a very vocal fringe minority. Not just that, but POV phrasing ends up getting inserted because they are forced to compromise on how to phrase and organize the article.
After GEM, which very obviously copied Apple, there were no new implementations of the concept until Konfabulator came along.
h emacdesktop/desktopaccessories
The fact that it was so long afterward should tell you that people had forgotten about Desktop Accessories. Most people, even MAc users, had no idea what they were until someone brought it up again in this context.
The technology, until a year or two ago when Konfabulator was ported to Windows, never existed for Windows. I've never seen it implemented in X11.
But there are similar things for X11. Even earlier window managers included very basic "mini apps" like a little clock and that moving eyes thing, and a calculator. These came along with Twm, which was created in 1987. They may even predate it, but I'm not sure. Of course, those are too liberal to fit any reasonable definition of something "widget like."
What Konfabulator brought to the table was HTML and Javascript. Beyond that, the apps it supports are barely different in functionality from those Mac OS 1 supported.
Oh, but it's the underlying architecture that makes it what it is. "Difference in functionality" is m eaningless without some unified framework to tie it all together. How those widgets interoperate with eachother is exactly what makes them widgets. A simple bunch of independent mini apps don't really qualify, as those could always be written in any language for any OS without the need for a special "widget engine."
As per your own description, you're relying on a very liberal interpretation of what it means to be a "widget engine," which suggests that it wasn't a rip off at all. They were nothing more than a hack to allow shitty multitasking in a non-multitasking OS. So in essence, it's just a multitasking driver. Other than the API necessary to multi-task, there is no unified API. You could easily make the programs look like whatever, there were no technical restricitons or designs in place to enforce it (beyond social mandate). So all they really are is independent, multitasking programs intended to look vaguely similar.
I suggest you take a look at this page for an idea of what they were like: http://www.guidebookgallery.org/articles/atouroft
http://en.wikipedia.org/wiki/Desk_Accessory
That claim about Desktop Accessories has been debunked, they're not like Konfabulator: http://arstechnica.com/reviews/os/macosx-10.4.ars/ 17