I think Prime Ministers can be different - people won't promote the best as they're scared of being replaced. So you actually get people being promoted above their level of incompetance
The ultrasparc approach sounds sensible - I wonder if the itanium offers anything similar?
The thing that concerns me is that if I have an arbitary object, I need to look up it's vtable.
ie, if the object is in r2, I need to do the equivalent of call (*r2)+4;
Although I can see that the computed address might be stored as a branch target, I can't see how this can be done particularly efficiently on, say, x86.
Also, this would be where the moving around of objects could potentially cause issues. I might be missing something - been a long time since I stared at computer architecture.
I imagine that modern garbage collection, which moves objects around, might actually also make this worse, as the branch target buffer relies on real addresses. That and the fact that a call will usually point to a vtable (thus requiring a pointer lookup) will probably mean that it plays havoc with branch prediction.
Still, somebody's probably come up with some gruesome way to avoid these penalties (I imagine that knowing you could still guess per-object, but because you look up the vtable each time it will cost you in branch target buffer space)
Of course, a smart jit will just avoid virtual method calls. If I do
Foo f = new Foo(); f.blah();
then f will always (aside from people pissing around with a debugger) be an object of type Foo. You could do the same optimisation with private variables and so on.
A really smart JVM might choose to align objects specially on the heap in order to help the branch hardware, but that sounds like a hideously complex optimisation....
Cool, so if I get the Linux kit I can stock up on American PS2 games! Fantastic.
(Note, I'm a Brit, and I wish I wish I wish that we had judges and laws that would give us some of this stuff over here. Maybe I should write to my MP)
(I'm sure this has been said 100s of times... but)
1. You can interpret Java, JIT it, or native compile it (have a play with gcj at some point)..NET might have a very well written JIT, and it might have some advantages that Java bytecode does not have - but this does not make life impossible, it simply makes it more interesting.
2. Java automatically garbage collects. System.gc() is a hint - for example, if you know you have created lots of garbage and you know you have time. The JVM can completely ignore you though.
3. JNI is there for a reason. There is OpenGl amongst other things implemented in this way.
4. The sleep function is implemented like that for correct handling in threading. If you want to cancel a thread, then you can call 'interrupt' on it, which will cause it to throw an InterruptedException. This allows threading to be a little safer than either allowing proper thread cancellation or just letting it sleep.
I've not really had enough time to play with.NET, but I'd like to. I think it's a shame that there isn't really a standard though. Seems a shame that all of these things will have to be rewritten.
Maybe a good project for somebody might be to write some standard APIs that can be used 'as is' from both C# and Java to do the same thing, so that the code can be broadly the same.
I've tried to explain this as well as I can, but I'm tired:|
Imagine I wanted to reduce pollution in an area where people were complaining. Let's say I gave you the right to pollute over your area of land, perhaps for a specific period of time (maybe 10 years or so), as part of owning land. You could sell that right onwards, or perhaps put further restrictions upon it. This would help to reduce the externalities from pollution. Each area would lose out by not allowing people to pollute.
This is what in economics is called a 'property right' - basically turning an abstract concept into a 'good' so that the market can allocate resources more correctly.
Intellectual Property Rights is a similar concept. If there were no restrictions such as copyright, trademarks etc, the market would work less efficiently. So a property right in certain intellectual endevours (marketing, literature (etc), patents) is designed to adjust an area where the market fails. The market fails to give an incentive to creation, thus by giving a property right over the area we can use the market to allocate in it.
I don't really think there is a better term to describe it in terms of economics. I'm not a fan of the term, as I think it lumps a lot of things together when people mean specific ones, but I think it fits neatly when discussing economics.
If I was selling software, I would not make US Senators (whose page is selling them to the public) non-commercial. If they didn't receive massive corporate donations maybe.
I'm sorry, but they are businesses as much as any other.
An alternative way of doing it of course is to take the output of a compiler to some sort of tree-phase. That might be harder to match against though, as you'd need some sort of fuzzy matching. Still it would give us a good clue of the 'identical' code that they're claiming, let alone similar code.
The rest of the responsibilities are fine (and are pretty much enshrined in the US through case law). But morals strikes me as allowing you to push all sorts of things. Mind you, I think that pushing that too hard can contravene article 9, so maybe it's just stopping you from having sex in public:)
You tell them to sue you. They sue you. You point out to the judge that 100Mb is several times the original argument, is significantly larger than required to make their point, and is contrary to the spirit of the law. You win.
All it's saying is that it's stupid for the law to say 'limited to the original' or 'limited to 1Mb' as times will move on. It says flexible, not stupid. Any judge is going to throw out a demand to host a video file in response to a text file.
Well, yes the term Intellectual Property is abused, however it comes from a perfectly justified phrase - 'Intellectual Property Rights' (IPR). Now, whether you agree or disagree with this naming, it's part of economics called (d'uh) property rights, which is the idea that you can improve the functioning of the economy by making certain people owners of non-standard items.
For example, if I owned the air over my house, it might mean that you had to pay for polluting it. If a town owned a local river, you might have to pay us to pollute it, or in fact we could deny you this. That's the concept of property rights, and IPR is in the same vein.
Not to say the term isn't abused - a purely economic term (and an idea designed to help the economy function better) has been hijacked as a perpetual right
I wonder how much of this is due to the ease of developing and how much is due to the nasties of the GNU licensing? I wouldn't mind writing a quick and dirty implementation of the sun.io framework (which various things need to link against) - which, while it might not be perfect, would potentially work, but I don't want to have to fart around with the annoyances of the GNU copyright assignment. Plus I imagine it would be far harder to get by my boss.
This isn't to say I disagree with the GNU copyright assignment, but I think it reduces the number of people who will work on a project like this.
That said, I think GCJ is doing a generally quite good job - it seems to work quite well, and the code seems to run at a fair lick.
Big business is generally in favour of certain types of deregulation.
For example, with the FCC in the States. Deregulation of, say, radio signals, or media which would allow anybody to produce media and compete easily - that would be 'bad' deregulation for large companies. However, allowing them to expand would be 'good' deregulation.
In general, deregulation, in the sense used by 'conservatives' - tends to mean removing the constraints that society imposes upon businesses aiming for fairness and equity. If a company runs a newspaper then they might not be allowed to own all of the newspapers in a city - equally not all of the TV, etc.
Now, there are loads of cases where business's like regulations. For example, the regulations that impose higher costs on new businesses (generally that the larger companies have to do anyway) are perfectly acceptable to them.
Sorry, bit rambling, but it appears that the regulations aimed at pushing companies to use their positions for social good are being thrown away, where the regulations that put them there are completely ignored. Gah
The standard edition runs an 'on-rails' simulation. Like the beginning of Half-life, but completely on the rails.
The expenso edition allows you to pause, and take it off the rails. The argument is that when you pause and go off the rails, it screws up, as if Nvidia have optimised it totally for being 'ON' the rails. They don't have the expenso edition so they probably didn't realise how easy it was to catch them out.
That's the summary anyway. Seems like a bit naughty to try, I think they would have to be pretty mad to try it.
If somebody isn't willing to sell me something that's one of a kind, I think that's acceptable, I see your point.
But if somebody said 'no, sorry, you can't read this cultural work of art, because I don't want you to' then sod them.
This isn't a case of where I am depriving somebody of something that they own. Nobody is selling it to me. When they finally *DO* sell it to me then I'll buy it from them, but until then I'll have to wait.
Equally I've watched most of Futurama off the net because it comes out in the States before it does over here. If they persist in treating non-Americans as second class citizens then they should expect no better.
This isn't meatspace, the entire industry is a mess of rules and regulations designed purely to create a product where none should exist. By your logic I could make a convincing case that people who avoid region coding are evil thieves as well, but I tend to agree with avoiding it myself.
This is more like going to a university and saying 'some of the dormrooms contain my stuff. Give me money or I'll turf you all out'. This isn't 1 single person acting in bad faith, it is potentially a small minority of people acting in bad faith with a large number of innocent people. The innocent people are basically being told that they cannot possibly trust anything their neighbours have done with no information!
SCO should release the information that is relevant, and then sue the culprits - the people who took this material and illegally used it. Otherwise they are merely penalising those who act in nothing but good faith. It also looks to me as if they are attacking Linux - I do not see that they lose *ANYTHING* by revealing the information, aside from maybe the ability to try and charge people for their usage (which would last precisely as long as it would take to remove it all, and also would be grossly against the GPL anyway)
Both of you seem to be very black and white on this.
I download a fair few things - Macgyver, old eighties cartoons that are impossible to find, Buffy and Angel episodes and Futurama. Why? Because I can't watch most of them.
I own Buffy seasons 1-4 on DVD, I want Angel as soon as I can afford it, and Buffy season 5. I'm waiting for some bright soul to realise that releasing the Mysterious Cities of Gold on DVD with English audio would be really smart, I'll still buy stuff. However, if I can't find something or watch it, then I will obtain it through other means.
I see the future as containing significantly more DRM, but also not a complete absolutism. People will cheat - they will always have friends who can obtain something, but it will be so much easier to obtain something legally, rather than trying to find an obscure film on Kazaa - you'll be able to select it in the same way that the iTunes library works, possibly even better. And when these systems fall down, piracy will fill in the cracks. Much like today.
The only really massive downsides is that this might really limit the difference between the theatrical release of a film and when it appears on DVD/Interest stores. But I think that all of the industries will adapt - they've done it before.
The fear with making this Rendevous enabled, or allowing it to be queried is that you don't necessarily want your employer gauging how you feel, and it's only useful if it's honest.
You could always hook it into a Livejournal client for rabid LJ users - since mood tends to be entered you could have that as a global or something. Set your mood and it adapts accordingly.
Should be an easy change - maybe have mood as a sort of dock icon, and allow it to be queried by all of these applications that would then switch profile.
Saving to VCR makes your TiVo useless for the length of the program. Not only that, but unlike Sky+ there is no queueing, so you can't even make good use of the time - I can't queue up 8 hours of shows to be put onto tape - I have to just satisfy myself with one at a time.
There's lots of stuff I want to archive. I might want to tape episodes for longer than I have space for, I might want to keep stuff around for a while - say interesting shows or programs that aren't repeated often enough.
In the end I want to move to MythTV and a relatively thin PC, as I'm imagining the flexibility increases will be immense - but having a PVR is still a lot better than not having one.
Slashdot Mirror
>lest we forget that microsoft did -not- shut down the open source .Net compatible project
Yeah, they're great like that. With MSN Messenger it's all open, and they allow loads of people to connect and stuff.
Oh, wait.
I think Prime Ministers can be different - people won't promote the best as they're scared of being replaced. So you actually get people being promoted above their level of incompetance
The ultrasparc approach sounds sensible - I wonder if the itanium offers anything similar?
The thing that concerns me is that if I have an arbitary object, I need to look up it's vtable.
ie, if the object is in r2, I need to do the equivalent of
call (*r2)+4;
Although I can see that the computed address might be stored as a branch target, I can't see how this can be done particularly efficiently on, say, x86.
Also, this would be where the moving around of objects could potentially cause issues. I might be missing something - been a long time since I stared at computer architecture.
I imagine that modern garbage collection, which moves objects around, might actually also make this worse, as the branch target buffer relies on real addresses. That and the fact that a call will usually point to a vtable (thus requiring a pointer lookup) will probably mean that it plays havoc with branch prediction.
Still, somebody's probably come up with some gruesome way to avoid these penalties (I imagine that knowing you could still guess per-object, but because you look up the vtable each time it will cost you in branch target buffer space)
Of course, a smart jit will just avoid virtual method calls. If I do
Foo f = new Foo();
f.blah();
then f will always (aside from people pissing around with a debugger) be an object of type Foo. You could do the same optimisation with private variables and so on.
A really smart JVM might choose to align objects specially on the heap in order to help the branch hardware, but that sounds like a hideously complex optimisation....
Cool, so if I get the Linux kit I can stock up on American PS2 games! Fantastic.
(Note, I'm a Brit, and I wish I wish I wish that we had judges and laws that would give us some of this stuff over here. Maybe I should write to my MP)
how do you extend it beyond 65536 elements?
By simple do you mean - edit the source code (not that I actually mind doing this, I'm just curious), or change a variable somewhere?
(I'm sure this has been said 100s of times... but)
.NET might have a very well written JIT, and it might have some advantages that Java bytecode does not have - but this does not make life impossible, it simply makes it more interesting.
.NET, but I'd like to. I think it's a shame that there isn't really a standard though. Seems a shame that all of these things will have to be rewritten.
1. You can interpret Java, JIT it, or native compile it (have a play with gcj at some point).
2. Java automatically garbage collects. System.gc() is a hint - for example, if you know you have created lots of garbage and you know you have time. The JVM can completely ignore you though.
3. JNI is there for a reason. There is OpenGl amongst other things implemented in this way.
4. The sleep function is implemented like that for correct handling in threading. If you want to cancel a thread, then you can call 'interrupt' on it, which will cause it to throw an InterruptedException. This allows threading to be a little safer than either allowing proper thread cancellation or just letting it sleep.
I've not really had enough time to play with
Maybe a good project for somebody might be to write some standard APIs that can be used 'as is' from both C# and Java to do the same thing, so that the code can be broadly the same.
I've tried to explain this as well as I can, but I'm tired :|
Imagine I wanted to reduce pollution in an area where people were complaining. Let's say I gave you the right to pollute over your area of land, perhaps for a specific period of time (maybe 10 years or so), as part of owning land. You could sell that right onwards, or perhaps put further restrictions upon it. This would help to reduce the externalities from pollution. Each area would lose out by not allowing people to pollute.
This is what in economics is called a 'property right' - basically turning an abstract concept into a 'good' so that the market can allocate resources more correctly.
Intellectual Property Rights is a similar concept. If there were no restrictions such as copyright, trademarks etc, the market would work less efficiently. So a property right in certain intellectual endevours (marketing, literature (etc), patents) is designed to adjust an area where the market fails. The market fails to give an incentive to creation, thus by giving a property right over the area we can use the market to allocate in it.
I don't really think there is a better term to describe it in terms of economics. I'm not a fan of the term, as I think it lumps a lot of things together when people mean specific ones, but I think it fits neatly when discussing economics.
If I was selling software, I would not make US Senators (whose page is selling them to the public) non-commercial. If they didn't receive massive corporate donations maybe.
I'm sorry, but they are businesses as much as any other.
An alternative way of doing it of course is to take the output of a compiler to some sort of tree-phase. That might be harder to match against though, as you'd need some sort of fuzzy matching. Still it would give us a good clue of the 'identical' code that they're claiming, let alone similar code.
The thing that gives me shivers is 'morals'
:)
The rest of the responsibilities are fine (and are pretty much enshrined in the US through case law). But morals strikes me as allowing you to push all sorts of things. Mind you, I think that pushing that too hard can contravene article 9, so maybe it's just stopping you from having sex in public
You tell them to sue you. They sue you. You point out to the judge that 100Mb is several times the original argument, is significantly larger than required to make their point, and is contrary to the spirit of the law. You win.
All it's saying is that it's stupid for the law to say 'limited to the original' or 'limited to 1Mb' as times will move on. It says flexible, not stupid. Any judge is going to throw out a demand to host a video file in response to a text file.
Well, yes the term Intellectual Property is abused, however it comes from a perfectly justified phrase - 'Intellectual Property Rights' (IPR). Now, whether you agree or disagree with this naming, it's part of economics called (d'uh) property rights, which is the idea that you can improve the functioning of the economy by making certain people owners of non-standard items.
For example, if I owned the air over my house, it might mean that you had to pay for polluting it. If a town owned a local river, you might have to pay us to pollute it, or in fact we could deny you this. That's the concept of property rights, and IPR is in the same vein.
Not to say the term isn't abused - a purely economic term (and an idea designed to help the economy function better) has been hijacked as a perpetual right
I wonder how much of this is due to the ease of developing and how much is due to the nasties of the GNU licensing? I wouldn't mind writing a quick and dirty implementation of the sun.io framework (which various things need to link against) - which, while it might not be perfect, would potentially work, but I don't want to have to fart around with the annoyances of the GNU copyright assignment. Plus I imagine it would be far harder to get by my boss.
This isn't to say I disagree with the GNU copyright assignment, but I think it reduces the number of people who will work on a project like this.
That said, I think GCJ is doing a generally quite good job - it seems to work quite well, and the code seems to run at a fair lick.
Big business is generally in favour of certain types of deregulation.
For example, with the FCC in the States. Deregulation of, say, radio signals, or media which would allow anybody to produce media and compete easily - that would be 'bad' deregulation for large companies. However, allowing them to expand would be 'good' deregulation.
In general, deregulation, in the sense used by 'conservatives' - tends to mean removing the constraints that society imposes upon businesses aiming for fairness and equity. If a company runs a newspaper then they might not be allowed to own all of the newspapers in a city - equally not all of the TV, etc.
Now, there are loads of cases where business's like regulations. For example, the regulations that impose higher costs on new businesses (generally that the larger companies have to do anyway) are perfectly acceptable to them.
Sorry, bit rambling, but it appears that the regulations aimed at pushing companies to use their positions for social good are being thrown away, where the regulations that put them there are completely ignored. Gah
From my reading of the article:
The standard edition runs an 'on-rails' simulation. Like the beginning of Half-life, but completely on the rails.
The expenso edition allows you to pause, and take it off the rails. The argument is that when you pause and go off the rails, it screws up, as if Nvidia have optimised it totally for being 'ON' the rails. They don't have the expenso edition so they probably didn't realise how easy it was to catch them out.
That's the summary anyway. Seems like a bit naughty to try, I think they would have to be pretty mad to try it.
If somebody isn't willing to sell me something that's one of a kind, I think that's acceptable, I see your point.
But if somebody said 'no, sorry, you can't read this cultural work of art, because I don't want you to' then sod them.
This isn't a case of where I am depriving somebody of something that they own. Nobody is selling it to me. When they finally *DO* sell it to me then I'll buy it from them, but until then I'll have to wait.
Equally I've watched most of Futurama off the net because it comes out in the States before it does over here. If they persist in treating non-Americans as second class citizens then they should expect no better.
This isn't meatspace, the entire industry is a mess of rules and regulations designed purely to create a product where none should exist. By your logic I could make a convincing case that people who avoid region coding are evil thieves as well, but I tend to agree with avoiding it myself.
I'm a Brit me :p
Go get yourself a multi-region DVD player.
I think we've got Buffy 1-6, Angel 1-3 atm.
My S1 and S2 of Buff are American though (they charge us about $100 per season!)
I've got that one. I want a DVD of it though :|
The quality of that is watchable, but not perfect. And he doesn't have the documentaries
This is more like going to a university and saying 'some of the dormrooms contain my stuff. Give me money or I'll turf you all out'. This isn't 1 single person acting in bad faith, it is potentially a small minority of people acting in bad faith with a large number of innocent people. The innocent people are basically being told that they cannot possibly trust anything their neighbours have done with no information!
SCO should release the information that is relevant, and then sue the culprits - the people who took this material and illegally used it. Otherwise they are merely penalising those who act in nothing but good faith. It also looks to me as if they are attacking Linux - I do not see that they lose *ANYTHING* by revealing the information, aside from maybe the ability to try and charge people for their usage (which would last precisely as long as it would take to remove it all, and also would be grossly against the GPL anyway)
GCC flags this and tells you to double bracket it in order to avoid the warning.
Both of you seem to be very black and white on this.
I download a fair few things - Macgyver, old eighties cartoons that are impossible to find, Buffy and Angel episodes and Futurama. Why? Because I can't watch most of them.
I own Buffy seasons 1-4 on DVD, I want Angel as soon as I can afford it, and Buffy season 5. I'm waiting for some bright soul to realise that releasing the Mysterious Cities of Gold on DVD with English audio would be really smart, I'll still buy stuff. However, if I can't find something or watch it, then I will obtain it through other means.
I see the future as containing significantly more DRM, but also not a complete absolutism. People will cheat - they will always have friends who can obtain something, but it will be so much easier to obtain something legally, rather than trying to find an obscure film on Kazaa - you'll be able to select it in the same way that the iTunes library works, possibly even better. And when these systems fall down, piracy will fill in the cracks. Much like today.
The only really massive downsides is that this might really limit the difference between the theatrical release of a film and when it appears on DVD/Interest stores. But I think that all of the industries will adapt - they've done it before.
The fear with making this Rendevous enabled, or allowing it to be queried is that you don't necessarily want your employer gauging how you feel, and it's only useful if it's honest.
You could always hook it into a Livejournal client for rabid LJ users - since mood tends to be entered you could have that as a global or something. Set your mood and it adapts accordingly.
Should be an easy change - maybe have mood as a sort of dock icon, and allow it to be queried by all of these applications that would then switch profile.
I also have one of them. They rock.
Saving to VCR makes your TiVo useless for the length of the program. Not only that, but unlike Sky+ there is no queueing, so you can't even make good use of the time - I can't queue up 8 hours of shows to be put onto tape - I have to just satisfy myself with one at a time.
There's lots of stuff I want to archive. I might want to tape episodes for longer than I have space for, I might want to keep stuff around for a while - say interesting shows or programs that aren't repeated often enough.
In the end I want to move to MythTV and a relatively thin PC, as I'm imagining the flexibility increases will be immense - but having a PVR is still a lot better than not having one.