Now that I am forced to articulate these disadvantages, I realize that they aren't as much problems with MI, as they are with c++'s implementation of MI. In c++ it is all too easy to use MI incorrectly and screw things up with improper casting. When used properly, it is certainly a very powerful language feature, and if you have used it without problems then you work with a more educated programming staff than I.
The other main disadvantage I've heard of is that it is a particularly difficult feature for compiler writers to implement. Although I've never written a compiler and don't intend to, I certainly think this is an important factor in language design. Simpler compilers means less buggy compilers with greater cross-vendor compatibility.
If you have working code that uses MI, then I think that's great, don't throw it away by any means. I just think it's a good idea to think twice about including it in new languages when there are other language constructs that can do just as good without some of the problems.
Which is why they do include support for interfaces. I don't think you understand the disadvantages of multiple inheritance. I've never found a situation where I actually required the additional functionality that multiple inheritance allows and coudn't be done better with just interfaces. Plus, if you were gonna copy multiple inheritance from c++ you'd need to copy all those nasty casting operators.
I didn't find how it deals with endianness specifically, but the idea is that D replaces a text-only preprocessor with actual sybolic features of the language. They definitely have support for conditional compilation and inline assembly code so I'm sure you could do whatever you need to handle endianness without a text-based preprocessor. The preprocessor itself isn't necessarily archaic, it's just unneccesary if your language isn't archaic.
Since we have no proof of anything beyond the Universe, this is just a chasing of a simple definition. Without the Universe in a 3D viewable environment and being just IT, then we can't define the shape meaningfully.
I'm not a mathematician, but my roommate was and he explained this to me once. These descriptions do not require anything beyond the observable universe to exist, they are merely technical statements comparing the characteristics of our universe to a flat euclidean universe, which are conveniently (confusingly) worded to sound like visual descriptions. The statement "the universe is shaped like a funnel" is still meaningful in that sense, even though no one can ever view the universe from the 4 dimensional perspective that would be required to actually see a funnel shape.
Is it just me, or are there no actual source files released? The only downloads I can find are binary only. This seems like an attempt to get some P.R. out of making something freely downloadable, which they do all the time anyway.
there *must* be something in for Microsoft for them to release the source of something
you say this like it's a bad thing... so what if there's "something in it" for microsoft? if there's even a glimmer of them opening their minds about open source, and realizing there's something worthwhile there, then I think it's a good thing.
Not true, they already had a buffer overflow protection that compilers could use a long time ago: the BOUND instruction. I don't think any mainstream compilers used it (at least not by default). This new fix is to allow an updated OS to protect against old apps that buffer overflow and minimize the damage. Besides, I think compilers already flag code and data segments within most executable file formats, but the OSs just ignore it for now because there's no hardware mechanism to enforce this distinction.
I will tell you that in the State of Georgia, we/do/ have the right to ask the name and information of any person in any public place
I have no doubt that your interpretation of Georgia law is correct, but the officer in this case acted legally under Nevada law as well. The point of this case is the argument that such laws are unconstitutional. The aclu brief argues that while officers have the right to ask for identification, people cannot be compelled to reply to their inquiries without probable cause.
I bet you're right though, the laws in most states are probably similar and will all be affected by the result of this case.
Frankly, I'd prefer to see neural-interface match-ups because then the games become less of a matter of how well you can properly wield a mouse, but it relies more on strategy.
I don't see why this would be the case. Just because the interface wouldn't be based on a mouse and keyboard doesn't mean that different people wouldn't have varying levels of skill operating the interface. It's easily conceivable that people's succeptiblity to biofeedback signals would vary just as widely as hand eye coordination.
That may be true, but the fact that the FSF now owns the copyrights means that the FSF can take over license enforcement. That's why they really did it.
you are correct, but you miss the point of the article. the article makes a much stronger claim, that even if you do "accidentally create a derived work" and redistribute it under a proprietary license (which is possible if you don't have perfect knowledge of what all your employees are doing) then the worst that can happen is that you get fined and forced to stop distributing your GPL-violating derived work. this argument destroys the popular claim that the punishment for GPL violations would be to force your proprietary additions to be released under the GPL.
actually, shouldn't we try to keep this quiet? doesn't this mean that we can manipulate elections now without the general public finding out? say goodbye to DMCA, UCITA, etc...
automated dependency resolution (and a lot more) for rpm based systems: Red Carpet
and if you need something that ximian doesn't have in their database you just pull it's dependency libraries from redcarpet and build it from an srpm. things definitely aren't as tough as they used to be
I have no experience with BSD, and I'm sure ports is great, but it isn't the only game in town anymore
There are no technical limitations, it's just a cost/benefit analysis. Right now a cluster that could render a cinematic scene in realtime isn't worth the money it would cost. Probably at some point (probably in the next 10 years) it will be cheap to render cinematic quality images of certain scenes in real time, but by then people will want to render much more complex environments (like fun-houses, scenes with an awful lot of reflections and refractions) just because they can, even if it means they have to sacrifice realtime rendering. If you really want to render a scene photorealistically, there is an essentially infinite amount of detail you could try to include and as a result, I think there will always be a market for non-realtime rendering products.
Quite true, but I think you're missing his point, which was that although the technology may be getting cheaper, it is becoming increasingly more costly to produce good content (he was just plain bullshitting about distribution) in the sense that finding good artists these days is harder, and movies and games have skyrocketing budgets. All of these factors add up to make it more expensive for content producers.
The saddest thing of all is not the failures of the current space program, as disturbing as they might be. The saddest thing is that we have lost the spirit and the system and methodology which yielded our greatest triumphs.
no, the saddest thing of all is the fact that that post makes me wish I knew how to use a slide rule. I'm such an inferior geek.
ps - alcohol is the best catalyst for karma combustion ever, here's to getting drunk on a tuesday night!
Slashdot Mirror
Now that I am forced to articulate these disadvantages, I realize that they aren't as much problems with MI, as they are with c++'s implementation of MI. In c++ it is all too easy to use MI incorrectly and screw things up with improper casting. When used properly, it is certainly a very powerful language feature, and if you have used it without problems then you work with a more educated programming staff than I.
The other main disadvantage I've heard of is that it is a particularly difficult feature for compiler writers to implement. Although I've never written a compiler and don't intend to, I certainly think this is an important factor in language design. Simpler compilers means less buggy compilers with greater cross-vendor compatibility.
If you have working code that uses MI, then I think that's great, don't throw it away by any means. I just think it's a good idea to think twice about including it in new languages when there are other language constructs that can do just as good without some of the problems.
I don't get it, what is wrong with D's templates. Admittedly, I haven't used D, but it sounds like it has templates to me...
Which is why they do include support for interfaces. I don't think you understand the disadvantages of multiple inheritance. I've never found a situation where I actually required the additional functionality that multiple inheritance allows and coudn't be done better with just interfaces. Plus, if you were gonna copy multiple inheritance from c++ you'd need to copy all those nasty casting operators.
I didn't find how it deals with endianness specifically, but the idea is that D replaces a text-only preprocessor with actual sybolic features of the language. They definitely have support for conditional compilation and inline assembly code so I'm sure you could do whatever you need to handle endianness without a text-based preprocessor. The preprocessor itself isn't necessarily archaic, it's just unneccesary if your language isn't archaic.
Since we have no proof of anything beyond the Universe, this is just a chasing of a simple definition. Without the Universe in a 3D viewable environment and being just IT, then we can't define the shape meaningfully.
I'm not a mathematician, but my roommate was and he explained this to me once. These descriptions do not require anything beyond the observable universe to exist, they are merely technical statements comparing the characteristics of our universe to a flat euclidean universe, which are conveniently (confusingly) worded to sound like visual descriptions. The statement "the universe is shaped like a funnel" is still meaningful in that sense, even though no one can ever view the universe from the 4 dimensional perspective that would be required to actually see a funnel shape.
Is it just me, or are there no actual source files released? The only downloads I can find are binary only. This seems like an attempt to get some P.R. out of making something freely downloadable, which they do all the time anyway.
there *must* be something in for Microsoft for them to release the source of something
you say this like it's a bad thing... so what if there's "something in it" for microsoft? if there's even a glimmer of them opening their minds about open source, and realizing there's something worthwhile there, then I think it's a good thing.
If they do that, then they risk getting their patent overturned in court. I think it's unlikely that they'll be that stupid.
Not true, they already had a buffer overflow protection that compilers could use a long time ago: the BOUND instruction. I don't think any mainstream compilers used it (at least not by default). This new fix is to allow an updated OS to protect against old apps that buffer overflow and minimize the damage. Besides, I think compilers already flag code and data segments within most executable file formats, but the OSs just ignore it for now because there's no hardware mechanism to enforce this distinction.
I will tell you that in the State of Georgia, we /do/ have the right to ask the name and information of any person in any public place
I have no doubt that your interpretation of Georgia law is correct, but the officer in this case acted legally under Nevada law as well. The point of this case is the argument that such laws are unconstitutional. The aclu brief argues that while officers have the right to ask for identification, people cannot be compelled to reply to their inquiries without probable cause.
I bet you're right though, the laws in most states are probably similar and will all be affected by the result of this case.
Frankly, I'd prefer to see neural-interface match-ups because then the games become less of a matter of how well you can properly wield a mouse, but it relies more on strategy.
I don't see why this would be the case. Just because the interface wouldn't be based on a mouse and keyboard doesn't mean that different people wouldn't have varying levels of skill operating the interface. It's easily conceivable that people's succeptiblity to biofeedback signals would vary just as widely as hand eye coordination.
That may be true, but the fact that the FSF now owns the copyrights means that the FSF can take over license enforcement. That's why they really did it.
Wow.... according to this gcc C sucks : (
actually gcc on windows sucks, big suprise there
As others pointed out, I was the one who was confused here.
funny how none of the moderators noticed this, well done fellow karma whore!
here, this should explain it to you.
you are correct, but you miss the point of the article. the article makes a much stronger claim, that even if you do "accidentally create a derived work" and redistribute it under a proprietary license (which is possible if you don't have perfect knowledge of what all your employees are doing) then the worst that can happen is that you get fined and forced to stop distributing your GPL-violating derived work. this argument destroys the popular claim that the punishment for GPL violations would be to force your proprietary additions to be released under the GPL.
actually, shouldn't we try to keep this quiet? doesn't this mean that we can manipulate elections now without the general public finding out? say goodbye to DMCA, UCITA, etc...
automated dependency resolution (and a lot more) for rpm based systems: Red Carpet
and if you need something that ximian doesn't have in their database you just pull it's dependency libraries from redcarpet and build it from an srpm. things definitely aren't as tough as they used to be
I have no experience with BSD, and I'm sure ports is great, but it isn't the only game in town anymore
don't blame me, I voted for Kodos...
seriously, and he didn't even include a link to the real smallest printer...
talk about dickhead....
There are no technical limitations, it's just a cost/benefit analysis. Right now a cluster that could render a cinematic scene in realtime isn't worth the money it would cost. Probably at some point (probably in the next 10 years) it will be cheap to render cinematic quality images of certain scenes in real time, but by then people will want to render much more complex environments (like fun-houses, scenes with an awful lot of reflections and refractions) just because they can, even if it means they have to sacrifice realtime rendering. If you really want to render a scene photorealistically, there is an essentially infinite amount of detail you could try to include and as a result, I think there will always be a market for non-realtime rendering products.
what, you don't like september? but halloween's right around the corner!
Quite true, but I think you're missing his point, which was that although the technology may be getting cheaper, it is becoming increasingly more costly to produce good content (he was just plain bullshitting about distribution) in the sense that finding good artists these days is harder, and movies and games have skyrocketing budgets. All of these factors add up to make it more expensive for content producers.
The new 'nv' driver for my Nvidia MX 2 Card causes...
Dude, you have an MX? No wonder you posted as AC, I'd be embarrassed too.
The saddest thing of all is not the failures of the current space program, as disturbing as they might be. The saddest thing is that we have lost the spirit and the system and methodology which yielded our greatest triumphs.
no, the saddest thing of all is the fact that that post makes me wish I knew how to use a slide rule. I'm such an inferior geek.
ps - alcohol is the best catalyst for karma combustion ever, here's to getting drunk on a tuesday night!