There's no "maybe" that that's what the statement means, and that it does not imply intent.:)
As to whether it's true, it's a fair observation that we don't really know. However since the long necks provide no clear advantages other than reach, and provide many downsides that required significant other evolutionary adaptations to support (implying adaptations to support a long neck were selected for, vs not having a long neck), I think it's probably safe to say that while it may not have started as a food-reaching adaptation, that quickly became the dominant selective pressure.
And if you replay "playing the piano" with the general "using tools", then I'd say we absolutely evolved in order to do just that.:)
I would be concerned that constraining the bees in that way might hinder their navigation ability, and just generally throw another variable into the experiment. It's quite possible to construct hard instances of the TSP where distances between nodes are "as the bee flies" in Euclidean space. Not that it couldn't potentially be an interesting experiment.:)
While it appears to be TSP, there are artificial restrictions imposed such that the weights on the edges are dependent on each other. More specifically, they are constrained by the triangle inequality theorem.
That's true in that they are not in the base statement of the problem, but the Euclidean TSP is still a version of the TSP, and is still NP-Hard and NP-Complete in the decision problem version.
In your example, nodes a, b and c fit this theorem, however, the edge between b and d cannot be 1 and still fit. Since a-b is 2 and a-d is 5, b-d can be no less than 3. (a-d <= a-b + b-d) and also no more than 7...
Given the triangle inequality constraint, I'm thinking the greedy algorithm will still suffice to find the shortest path from a specific starting node to every other node in O(n^2) time. O(n^3) to find the shortest path from an arbitrary starting node. I think this constraint is what differentiates this problem from the typical TSP which has no such restriction. What are your thoughts?
My thoughts are that while my half-assed example generated specifically to foil your algorithm while occupying a minimum amount of my time may not follow the triangle inequality, it is still trivial to generate a similar set of nodes that would obey the triangle inequality and still make a greedy algorithm suboptimal.:)
The basic observation here is that the greedy algorithm, by always taking the shortest hop at any given step, does not guarantee that subsequent hops from which it has to choose are not much longer than they would be otherwise. Whereas taking longer hops at a given step may provide for better choices of path in the future. The long chain of nearby cities which ends far away from any other city and thus should be the end of the tour is such an example, and doesn't require non-Euclidean space to be set up. You can't arrive at a globally optimal route simply by making local decisions.
The Euclidean Traveling Salesman (and Manhattan TSP) decision problem is in fact NP-Complete, which I know both because I had to prove it in my algorithms class, and because smarter people than me have proven it. So if the greedy algorithm was an actual solution, and not just a heuristic, then it would be proof that P=NP. But, sadly, it's not.:)
1. Euclidean Traveling Salesman is probably not NP-Complete.
It provably is. Or specifically, it's NP-hard, and the decision version ("is there a solution of less than length x") is NP-complete, just like the general TSP problem.
2. In fact there is a PTAS (polynomial time approximation scheme) for ETSP so the bees could be computing approximate solutions to ETSP.
Seems by far the most likely, yeah.
I propose a new experiment:
1. 1. We can pick a hard cryptographic problem (say factoring a number N). We can take our specific instance N from some large public RSA key.
2. 2. We can easily reduce factoring to TSP to get some specific TSP instance T. This ensures that we pick a hard TSP instance (either that or factoring N and breaking RSA wasn't that hard in the first place). Note that these distances are not necessarily Euclidean!
3. 3. Add a flower for each vertex in T
How are you going to add physical flowers if the distances between them do not follow Euclidean geometry?
I can think of lots of experiments to conduct, but it's hard to know what are worthwhile without knowing exactly what they did. But their experimental setup (having a variety of computer controlled artificial flowers) seems general enough that it would have been pretty easy to set up cases for which well-known heuristics don't provide the optimal solution. That still means the most likely case is just that they're using a different, maybe better heuristic. But it still means we could learn a lot, aside from the basic neurology questions.
So if a "bee-wolf" upgrades the wolf to a bear, would a "bee-bear" (beobeowulf I guess?) turn it into a Tyranosaurus? Or a raptor with an RPG riding a shark?
Whether or not it "typically" is fully connected, the standard form wherein it is proven to be NP-hard is fully connected. It is not any easier to solve in that special case; in general it is harder. Any non-fully connected form can be converted to fully connected form by adding in the necessary connections with arbitrarily large lengths.
Should I even mention all the trouble people had with the buggy and broken Grub2 and Plymouth releases? The way two finger middleclicking in Firefox was broken? Notice that I have yet to bring up the fscking window controls?
Yeah, and I also notice you have yet to bring up anything to do with UI consistency. Bugs are bad, and breaking things like customizing the system tray is a real issue, but only for people who are customizing their UI, which is the opposite of the scenario we're talking about. And grub2. You're seriously bringing up grub2 in a discussion about consistent UI elements.
It never bothered me so much as I actually prefer my window controls to be on left, but I can understand the annoyance for people who have always used their controls on the right. It creates a mental dissonance that costs you time every attempt to close a window when you go to do so and have to recalibrate when you find it isn't there.
And if that actually causes a problem every time rather than the 1st couple times they go to use a window control after the change, then I maintain that they are being done a favor by being forced to generalize their UI knowledge. Their muscle memory will adapt to quickly deal with different windowing systems, and this will only serve them
So yeah... There are all kinds of issues with Canonical's arbitrarily deciding to change things that have nothing to do with the window controls and it's insulting for you to pretend otherwise.
I'm not pretending Ubuntu has no other issues. I'm talking about the topic at hand. Until you can understand that subtle distinction, I give not a flying fuck on a rubber buffalo dong if you're offended, or if you stop using Ubuntu because you're mad at me.
Sometimes its better for angled reflective objects if the light is above the object to reflect into your eye, as opposed to coming from your head and then reflecting away.
But only rarely is this better if it costs you the use of your hand.
If you want light coming from arbitrary angles, get a snake light.:)
No, that's simply not true. The basic form of the TSP consists of a fully connected graph of cities, i.e. you can get from any city directly to any other city.
The non-fully-connected graph is a special case of TSP, and it is these cases that may end up simpler to solve (though often not, other than not being fully connected means there are fewer possible routes to check for the same number of cities).
"The plants evolved so bees could find them" does not imply intent or decision. It simply means it happened, and pollination by bees, and thus the need for the bees to find them, was the selective pressure that drove the evolution. Similarly, "giraffes evolved to graze from trees other animals couldn't" doesn't mean giraffes were looking longingly at the high up branches and wishing they had long necks. It means that's how they adapted to competition via natural selection.
Similarly, statements like "Bee eyes are designed to see UV wavelengths that flowers reflect strongly in" does not imply a "designer", bee, alien, or god, actually designed their eyes. Their eyes have a design, and it has an apparent purpose.
Sometimes words have implications in common English that do not apply to all situations.
. The traveling salesman problem is hard because the graph is not necessarily fully connected so all paths have to be examined individually.
No, sorry, the Traveling Salesman Problem is often presented as a fully-connected graph, and is NP-Hard in that form. It's difficulty does not rely on the possibility that the graph is not fully connected (and is slightly easier if it isn't, because there are fewer possible paths, though this obviously doesn't affect the big-O).
I believe the shortest path on a fully connected graph is found by always choosing the closest non-visited neighbor from the current node. The difference in calculation is O(n!) vs. O(n^2).
Sorry, that doesn't work for some trivial examples. Consider cities A, B, C, D, E, and their pairwise distances as so:
The optimal route is A-B-C-D-E (10). Your algorithm would take A-B-D-E-C (12). Yours would come up with the optimal answer if you started at E (so you have to run the algorithm for each starting point, so it'd be N^3), but I could easily add another instance of the same type of fork so you'd always encounter this hole in your algorithm regardless of where you started.
Basically, by always visiting the nearest city, you can end up putting yourself very far away from the next city when you run out of close ones, when visiting a more distant city could possibly save you time on subsequent trips.
It's what makes this a difficult problem -- that you can't arrive at an optimal global solution by solely making local decisions.
I can understand being pissed about the lack of LAN play in Starcraft, but the only times I ever played Diablo II off of battle.net was when I was playing it with Wine and battle.net didn't work (and LAN play barely worked).
So... I still care about Blizzard and Diablo III because I don't care about what has your panties in a bunch. Sorry.
Actually I imagined hooking up a super powerful LED to some nerve endings on the tips of my finger
Bad choice of location. I mean, it's a great choice for the utility of the flashlight, but poor because you're giving up a region of your most sensitive tactile sense.
Me, my first thought is forehead, like a headlamp, so both hands are free. Maybe with an adjustable lens to point it up and down. Make it so.
Oh but seriously, programmable, animated tattoos would be pretty cool. I might get one of those.:)
Oh no. The window controls moved from the right top of the window to the left top of the window, some icons look slightly different, and the default theme is a slightly different shade of brown. My world, she is rent asunder.
The interface changes from the last few years of Ubuntu updates are in the same order of magnitude as XP to Vista, or Win98 to XP. The "consistency" of Windows is an artifact of them not releasing a new OS for over 5 years. When they do, they of course make different decisions than what they made years before in a previous release. They only maintain "consistency" in the broadest scope, like there's still a Start menu and window frames still have a Close, Minimize, and Maximize buttons. Ubuntu has this too.
I understand that consistency is something people desire in the abstract. I do not believe that lack of consistency is a reason anyone stayed away from Vista. They stayed away because it was crap. Now people are happily using Windows 7, and the fact that some icons look different. Similarly, the Ubuntu releases are not so dissimilar as to actually cause significant confusion. Maybe for 5 seconds -- "where did Minimize go? Oh, there it is."
And frankly, if those 5 seconds of confusion cause a panic, or a desire to avoid that OS from then on, then I believe that you need to be exposed to some inconsistency in the form of new GUI interfaces. Learning to use one and exactly one specific interface is a recipe for obsolescence. Exposure to multiple GUIs results in generalizing your understanding, so then when you sit down in front of a new and seemingly completely different GUI (like your friends Macbook), you aren't lost.
That said, consistency is good, and randomly changing the interface (considered in isolation from why) is undesirable. But that is not why people are avoiding Ubuntu and Linux in general. They're avoiding it for other OSes because those other OSes come pre-installed by OEMs, and support all the software they want to run.
Which kind of leaves me wondering what the point is; the primary market is people who want to play games but not enough to actually buy a graphics card which can do so. Maybe it would be beneficial in the laptop market where many systems can't really play any games at all.
Naaaaah, that's not it.
Just kidding, of course that's it. It's like I'm trying to say: Where is having performance around maybe a low-end discreet card, or at least that feature set, but with the economics and power usage of an integrated solution, actually a good tradeoff?
Where low power usage isn't just nice but a requirement, and where that level of performance and feature set is actually pretty good -- for laptops, that's a good trade-off. There's a reason it's 20-60W and adjusts frequency and power usage dynamically.
In desktops, it's pretty easy to beat the performance of a chip that doesn't have dedicated high-bandwidth DRAM. And it doesn't cost much to have another model with an add-on card. In that space, they'll probably try to play it like they did dual-core CPUS and price them close enough to integrated graphics solutions that the OEM's cost advantage of having one fewer chipset let's AMD have a little margin on the parts, and try to make that level of graphics the default. If a decent and compatible GPU can be assumed in a significant percentage of cases, then people will write more apps for them, which increases demand for the chips. While Intel and others are moving in that direction, AMD is more out on their own here, so it remains to be seen how successful they'll be. And how the real chip's advantages stack up. But the strategy seems sound, and beneficial if it pans out.
The very high-end, sure. For the rest, the cost and battery life advantages will steer it towards Fusion. People these days want a high-end laptop that can play games, but that can also function as a useful portable computer in the absence of a power outlet.
The idea AMD is going for here, and I think there's some merit to it, is a low range desktop type of system. People who want something cheap, but still want to play games. Intel's integrated chips don't work that well for that (though they've improved) so this is to try and fill the market.
Think more mid-to-high-end laptops.
As mentioned in the summary, this is a low-power version of the Phenom II. Not an ultra-low power for consumer electronics or netbooks like Atom or AMD's Bobcat, but still solidly aimed at the mobile market. It provides all the power and cost advantages of a UMA solution plus gets rid of one of the system buses for more savings, while providing good-for-a-laptop graphics without having to significantly re-engineer the motherboard or cooling solution. This is still in theory; demonstrations of engineering samples are nice, but it'll be interesting once the reviewers get their hands on some.
Of course you're also right, since cost and power usage are relevant for desktop. Just not as much, since you're not dealing with battery life, or the form factor that make it difficult to work with discreet graphics. A single line of UMA-based motherboards with optional pci-e graphics card can serve multiple markets with one design and acceptable margins.
Don't waste a wish. This is truly already within your power. You can decide whether or not, and to what degree, you give a damn about what other people think or say. Today, without wishing for a thing.
What people think or say about the flagrantly false things they believe can affect real-world behaviors and decisions with real negative consequences, and it is clearly not within my power to "decide" that this not be so, otherwise we'd be involved in at minimum one fewer war. The Mayan Calendar Doomsday Myth, while not of much real world impact, is a symptom of a broader issue that truly does matter.
So your suggestion is like saying I should save my wish that the hungry children of the world would be fed, because instead I could simply stop caring.
You cannot possibly substantiate "none whatsoever" because we have only a minute sliver of their knowledge to go on.
No, but you can trivially substantiate "none known whatsoever" based on the fact that there's only a minute sliver of knowledge.
Which means there's no evidence of there being such a prophecy, and the doomsday theorists are basing their theory on nothing.
2) There's little to no reason to believe that the calendar wouldn't end on the solstice. Nothing in the article seems to illustrate WHY they would suddenly dispose of any regard for astronomy with the end of this one single calendar.
Yeah, except for the whole thing with the calendar in question not even matching the year much less any other astronomical phenomenon because they were stuck on using multiples of 20 even where it didn't work. So, there's actually quite a bit of reason right there.
Contrariwise, there's no reason to assign any real or hypothetical significance to the end of this calendar other than it being the end of the calendar. My desk calendar ends on Dec. 31. Prophecy of the end of the world? or just time to buy a new calendar? Certainly if you look for astronomical significance, you won't find any. Other than it's very close to one revolution of the earth around the sun since the end of the previous calendar, which you can't even say about the Mayan calendar.
You seem rather worked up over what would be, if you're correct, basically nothing. So what's the deal? Whatever it is, why not save us all some time and simply argue THAT instead?
I wish people believing things despite all evidence pointing in the opposite direction, simply because it appeals to them in some way, was "basically nothing".
Slashdot Mirror
There's no "maybe" that that's what the statement means, and that it does not imply intent. :)
As to whether it's true, it's a fair observation that we don't really know. However since the long necks provide no clear advantages other than reach, and provide many downsides that required significant other evolutionary adaptations to support (implying adaptations to support a long neck were selected for, vs not having a long neck), I think it's probably safe to say that while it may not have started as a food-reaching adaptation, that quickly became the dominant selective pressure.
And if you replay "playing the piano" with the general "using tools", then I'd say we absolutely evolved in order to do just that. :)
I would be concerned that constraining the bees in that way might hinder their navigation ability, and just generally throw another variable into the experiment. It's quite possible to construct hard instances of the TSP where distances between nodes are "as the bee flies" in Euclidean space. Not that it couldn't potentially be an interesting experiment. :)
While it appears to be TSP, there are artificial restrictions imposed such that the weights on the edges are dependent on each other. More specifically, they are constrained by the triangle inequality theorem.
That's true in that they are not in the base statement of the problem, but the Euclidean TSP is still a version of the TSP, and is still NP-Hard and NP-Complete in the decision problem version.
In your example, nodes a, b and c fit this theorem, however, the edge between b and d cannot be 1 and still fit. Since a-b is 2 and a-d is 5, b-d can be no less than 3. (a-d <= a-b + b-d) and also no more than 7...
Given the triangle inequality constraint, I'm thinking the greedy algorithm will still suffice to find the shortest path from a specific starting node to every other node in O(n^2) time. O(n^3) to find the shortest path from an arbitrary starting node. I think this constraint is what differentiates this problem from the typical TSP which has no such restriction. What are your thoughts?
My thoughts are that while my half-assed example generated specifically to foil your algorithm while occupying a minimum amount of my time may not follow the triangle inequality, it is still trivial to generate a similar set of nodes that would obey the triangle inequality and still make a greedy algorithm suboptimal. :)
The basic observation here is that the greedy algorithm, by always taking the shortest hop at any given step, does not guarantee that subsequent hops from which it has to choose are not much longer than they would be otherwise. Whereas taking longer hops at a given step may provide for better choices of path in the future. The long chain of nearby cities which ends far away from any other city and thus should be the end of the tour is such an example, and doesn't require non-Euclidean space to be set up. You can't arrive at a globally optimal route simply by making local decisions.
The Euclidean Traveling Salesman (and Manhattan TSP) decision problem is in fact NP-Complete, which I know both because I had to prove it in my algorithms class, and because smarter people than me have proven it. So if the greedy algorithm was an actual solution, and not just a heuristic, then it would be proof that P=NP. But, sadly, it's not. :)
Coo! Coo! I'm a dove; watch me fly away in dramatic fashion!
1. Euclidean Traveling Salesman is probably not NP-Complete.
It provably is. Or specifically, it's NP-hard, and the decision version ("is there a solution of less than length x") is NP-complete, just like the general TSP problem.
2. In fact there is a PTAS (polynomial time approximation scheme) for ETSP so the bees could be computing approximate solutions to ETSP.
Seems by far the most likely, yeah.
I propose a new experiment:
1. 1. We can pick a hard cryptographic problem (say factoring a number N). We can take our specific instance N from some large public RSA key.
2. 2. We can easily reduce factoring to TSP to get some specific TSP instance T. This ensures that we pick a hard TSP instance (either that or factoring N and breaking RSA wasn't that hard in the first place). Note that these distances are not necessarily Euclidean!
3. 3. Add a flower for each vertex in T
How are you going to add physical flowers if the distances between them do not follow Euclidean geometry?
I can think of lots of experiments to conduct, but it's hard to know what are worthwhile without knowing exactly what they did. But their experimental setup (having a variety of computer controlled artificial flowers) seems general enough that it would have been pretty easy to set up cases for which well-known heuristics don't provide the optimal solution. That still means the most likely case is just that they're using a different, maybe better heuristic. But it still means we could learn a lot, aside from the basic neurology questions.
Huh, I didn't know that.
So if a "bee-wolf" upgrades the wolf to a bear, would a "bee-bear" (beobeowulf I guess?) turn it into a Tyranosaurus? Or a raptor with an RPG riding a shark?
Whether or not it "typically" is fully connected, the standard form wherein it is proven to be NP-hard is fully connected. It is not any easier to solve in that special case; in general it is harder. Any non-fully connected form can be converted to fully connected form by adding in the necessary connections with arbitrarily large lengths.
Should I even mention all the trouble people had with the buggy and broken Grub2 and Plymouth releases? The way two finger middleclicking in Firefox was broken? Notice that I have yet to bring up the fscking window controls?
Yeah, and I also notice you have yet to bring up anything to do with UI consistency. Bugs are bad, and breaking things like customizing the system tray is a real issue, but only for people who are customizing their UI, which is the opposite of the scenario we're talking about. And grub2. You're seriously bringing up grub2 in a discussion about consistent UI elements.
It never bothered me so much as I actually prefer my window controls to be on left, but I can understand the annoyance for people who have always used their controls on the right. It creates a mental dissonance that costs you time every attempt to close a window when you go to do so and have to recalibrate when you find it isn't there.
And if that actually causes a problem every time rather than the 1st couple times they go to use a window control after the change, then I maintain that they are being done a favor by being forced to generalize their UI knowledge. Their muscle memory will adapt to quickly deal with different windowing systems, and this will only serve them
So yeah... There are all kinds of issues with Canonical's arbitrarily deciding to change things that have nothing to do with the window controls and it's insulting for you to pretend otherwise.
I'm not pretending Ubuntu has no other issues. I'm talking about the topic at hand. Until you can understand that subtle distinction, I give not a flying fuck on a rubber buffalo dong if you're offended, or if you stop using Ubuntu because you're mad at me.
Yes but you may want to adjust the angle relative to your face, or to get a good angle for the light without it being a bad angle for your neck.
I say this from personal experience with headlamps, where ones that have an adjustable tilt are much more useful than ones without.
Sometimes its better for angled reflective objects if the light is above the object to reflect into your eye, as opposed to coming from your head and then reflecting away.
But only rarely is this better if it costs you the use of your hand.
If you want light coming from arbitrary angles, get a snake light. :)
No, that's simply not true. The basic form of the TSP consists of a fully connected graph of cities, i.e. you can get from any city directly to any other city.
The non-fully-connected graph is a special case of TSP, and it is these cases that may end up simpler to solve (though often not, other than not being fully connected means there are fewer possible routes to check for the same number of cities).
"The plants evolved so bees could find them" does not imply intent or decision. It simply means it happened, and pollination by bees, and thus the need for the bees to find them, was the selective pressure that drove the evolution. Similarly, "giraffes evolved to graze from trees other animals couldn't" doesn't mean giraffes were looking longingly at the high up branches and wishing they had long necks. It means that's how they adapted to competition via natural selection.
Similarly, statements like "Bee eyes are designed to see UV wavelengths that flowers reflect strongly in" does not imply a "designer", bee, alien, or god, actually designed their eyes. Their eyes have a design, and it has an apparent purpose.
Sometimes words have implications in common English that do not apply to all situations.
. The traveling salesman problem is hard because the graph is not necessarily fully connected so all paths have to be examined individually.
No, sorry, the Traveling Salesman Problem is often presented as a fully-connected graph, and is NP-Hard in that form. It's difficulty does not rely on the possibility that the graph is not fully connected (and is slightly easier if it isn't, because there are fewer possible paths, though this obviously doesn't affect the big-O).
I believe the shortest path on a fully connected graph is found by always choosing the closest non-visited neighbor from the current node. The difference in calculation is O(n!) vs. O(n^2).
Sorry, that doesn't work for some trivial examples. Consider cities A, B, C, D, E, and their pairwise distances as so:
A-B: 2
A-C: 4
A-D: 5
A-E: 7
B-C: 2
B-D: 1
B-E: 3
C-D: 5
C-E: 7
D-E: 2
The optimal route is A-B-C-D-E (10). Your algorithm would take A-B-D-E-C (12). Yours would come up with the optimal answer if you started at E (so you have to run the algorithm for each starting point, so it'd be N^3), but I could easily add another instance of the same type of fork so you'd always encounter this hole in your algorithm regardless of where you started.
Basically, by always visiting the nearest city, you can end up putting yourself very far away from the next city when you run out of close ones, when visiting a more distant city could possibly save you time on subsequent trips.
It's what makes this a difficult problem -- that you can't arrive at an optimal global solution by solely making local decisions.
I can understand being pissed about the lack of LAN play in Starcraft, but the only times I ever played Diablo II off of battle.net was when I was playing it with Wine and battle.net didn't work (and LAN play barely worked).
So... I still care about Blizzard and Diablo III because I don't care about what has your panties in a bunch. Sorry.
If the servers are too stressed they should load some hot tea and yoga programs alongside Diablo 3.
Oh no, we've already had enough trouble with Hot Coffee.
Actually I imagined hooking up a super powerful LED to some nerve endings on the tips of my finger
Bad choice of location. I mean, it's a great choice for the utility of the flashlight, but poor because you're giving up a region of your most sensitive tactile sense.
Me, my first thought is forehead, like a headlamp, so both hands are free. Maybe with an adjustable lens to point it up and down. Make it so.
Oh but seriously, programmable, animated tattoos would be pretty cool. I might get one of those. :)
Oh no. The window controls moved from the right top of the window to the left top of the window, some icons look slightly different, and the default theme is a slightly different shade of brown. My world, she is rent asunder.
The interface changes from the last few years of Ubuntu updates are in the same order of magnitude as XP to Vista, or Win98 to XP. The "consistency" of Windows is an artifact of them not releasing a new OS for over 5 years. When they do, they of course make different decisions than what they made years before in a previous release. They only maintain "consistency" in the broadest scope, like there's still a Start menu and window frames still have a Close, Minimize, and Maximize buttons. Ubuntu has this too.
I understand that consistency is something people desire in the abstract. I do not believe that lack of consistency is a reason anyone stayed away from Vista. They stayed away because it was crap. Now people are happily using Windows 7, and the fact that some icons look different. Similarly, the Ubuntu releases are not so dissimilar as to actually cause significant confusion. Maybe for 5 seconds -- "where did Minimize go? Oh, there it is."
And frankly, if those 5 seconds of confusion cause a panic, or a desire to avoid that OS from then on, then I believe that you need to be exposed to some inconsistency in the form of new GUI interfaces. Learning to use one and exactly one specific interface is a recipe for obsolescence. Exposure to multiple GUIs results in generalizing your understanding, so then when you sit down in front of a new and seemingly completely different GUI (like your friends Macbook), you aren't lost.
That said, consistency is good, and randomly changing the interface (considered in isolation from why) is undesirable. But that is not why people are avoiding Ubuntu and Linux in general. They're avoiding it for other OSes because those other OSes come pre-installed by OEMs, and support all the software they want to run.
Which kind of leaves me wondering what the point is; the primary market is people who want to play games but not enough to actually buy a graphics card which can do so. Maybe it would be beneficial in the laptop market where many systems can't really play any games at all.
Naaaaah, that's not it.
Just kidding, of course that's it. It's like I'm trying to say: Where is having performance around maybe a low-end discreet card, or at least that feature set, but with the economics and power usage of an integrated solution, actually a good tradeoff?
Where low power usage isn't just nice but a requirement, and where that level of performance and feature set is actually pretty good -- for laptops, that's a good trade-off. There's a reason it's 20-60W and adjusts frequency and power usage dynamically.
In desktops, it's pretty easy to beat the performance of a chip that doesn't have dedicated high-bandwidth DRAM. And it doesn't cost much to have another model with an add-on card. In that space, they'll probably try to play it like they did dual-core CPUS and price them close enough to integrated graphics solutions that the OEM's cost advantage of having one fewer chipset let's AMD have a little margin on the parts, and try to make that level of graphics the default. If a decent and compatible GPU can be assumed in a significant percentage of cases, then people will write more apps for them, which increases demand for the chips. While Intel and others are moving in that direction, AMD is more out on their own here, so it remains to be seen how successful they'll be. And how the real chip's advantages stack up. But the strategy seems sound, and beneficial if it pans out.
The very high-end, sure. For the rest, the cost and battery life advantages will steer it towards Fusion. People these days want a high-end laptop that can play games, but that can also function as a useful portable computer in the absence of a power outlet.
No, think value desktops.
A suitable market, but not the primary thrust. Read my post.
With regard to Fusion generally; never, ever bet against integration.
Integration is even more important in the mobile market, for all the reasons I already discussed.
You know what else the Mayans figured out? Jokes.
Yeah well I'd study their teachings a bit more before I quit my day job, if I were you.
The idea AMD is going for here, and I think there's some merit to it, is a low range desktop type of system. People who want something cheap, but still want to play games. Intel's integrated chips don't work that well for that (though they've improved) so this is to try and fill the market.
Think more mid-to-high-end laptops.
As mentioned in the summary, this is a low-power version of the Phenom II. Not an ultra-low power for consumer electronics or netbooks like Atom or AMD's Bobcat, but still solidly aimed at the mobile market. It provides all the power and cost advantages of a UMA solution plus gets rid of one of the system buses for more savings, while providing good-for-a-laptop graphics without having to significantly re-engineer the motherboard or cooling solution. This is still in theory; demonstrations of engineering samples are nice, but it'll be interesting once the reviewers get their hands on some.
Of course you're also right, since cost and power usage are relevant for desktop. Just not as much, since you're not dealing with battery life, or the form factor that make it difficult to work with discreet graphics. A single line of UMA-based motherboards with optional pci-e graphics card can serve multiple markets with one design and acceptable margins.
Don't waste a wish. This is truly already within your power. You can decide whether or not, and to what degree, you give a damn about what other people think or say. Today, without wishing for a thing.
What people think or say about the flagrantly false things they believe can affect real-world behaviors and decisions with real negative consequences, and it is clearly not within my power to "decide" that this not be so, otherwise we'd be involved in at minimum one fewer war. The Mayan Calendar Doomsday Myth, while not of much real world impact, is a symptom of a broader issue that truly does matter.
So your suggestion is like saying I should save my wish that the hungry children of the world would be fed, because instead I could simply stop caring.
How foolish.
You cannot possibly substantiate "none whatsoever" because we have only a minute sliver of their knowledge to go on.
No, but you can trivially substantiate "none known whatsoever" based on the fact that there's only a minute sliver of knowledge.
Which means there's no evidence of there being such a prophecy, and the doomsday theorists are basing their theory on nothing.
2) There's little to no reason to believe that the calendar wouldn't end on the solstice. Nothing in the article seems to illustrate WHY they would suddenly dispose of any regard for astronomy with the end of this one single calendar.
Yeah, except for the whole thing with the calendar in question not even matching the year much less any other astronomical phenomenon because they were stuck on using multiples of 20 even where it didn't work. So, there's actually quite a bit of reason right there.
Contrariwise, there's no reason to assign any real or hypothetical significance to the end of this calendar other than it being the end of the calendar. My desk calendar ends on Dec. 31. Prophecy of the end of the world? or just time to buy a new calendar? Certainly if you look for astronomical significance, you won't find any. Other than it's very close to one revolution of the earth around the sun since the end of the previous calendar, which you can't even say about the Mayan calendar.
You seem rather worked up over what would be, if you're correct, basically nothing. So what's the deal? Whatever it is, why not save us all some time and simply argue THAT instead?
I wish people believing things despite all evidence pointing in the opposite direction, simply because it appeals to them in some way, was "basically nothing".
Eh well I honestly stole the idea from Dr. Venture.