The thing that most bugs me about this is that there is absolutely no reason for Time Machine to not work with File Vault. A correct implementation would keep an encrypted disk image on the backup device with the same password as the original, have hard links in the image for versioning inside the image and copy files in whenever you ran the backup utility, just as it does on the main filesystem with unencrypted images. Then you would have to either have your password or the file vault master password in order to run the backup, which seems completely understandable. If you think this through you can see the problem. the time machine vault could after a while be hundreds of gigabytes or more in size. every time it updated it would have to decrypt the entire hundreds of gigabytes and create some place to store the decrypted image. Where are you going to mount it, and are you really going to put up with your computer stopping every ten minutes to update the encrypted vault? The it has to re-encrypt all the hundreds of gigabytes. Either that or it would have to decrypt it on the fly for all accesses. This is going to get brutal. Perhaps there's some lighter weight non-monolithic protocol or other ways they could store incremental changes between massive decypts. But given all the problems they had getting file vault working in the first place I can see why maybe they would see that time machine 1.0 does not need file vault too. leave that for 2.0
I thought Time Machine was great when I saw the demos, but it is completely incompatible with File Vault, which means you have to choose between security and safety for your data. You can kind-of use it with File Vault, but it will only run backups when you log out (which is something I only ever do to reboot for software updates), and so is completely useless. huh? So you were a fan of file vault till they came out with time machine and then suddenly File vault is useless? So they have features that cannot be used together, big deal. File vault is mainly for laptops that risk exposure or shared machines in neither case is Time Machine going to be the preferred way of backing up the machine so it's not a big crossections of folks that needs both on one machine I'd say.
The biggest improvement with 10.5 is that Spotlight now actually works. In 10.4 it was so slow that I could generally find files faster without it. With 10.5 it is fast enough to be useful. Yep. in 10.4 it was worthless. piece of crap. I got a new machine with 10.5 and it suddenly worked. I was scratching my head. Did I dod something different on this new machine? I had not heard anyone elese talking about the fact that Spotlight actually worked. So I guess it's not just me.
I keep my dock on the left side, attached to the top-left corner, and the 10.5 dock is about as nice as the 10.4 one, just different. mine hangs to the right.
V.I. Kydor Kropotkin:
Are you trying to tell me every phone in the country is tapped?
Don Masters, CEA Agent:
That's what's in my head.
V.I. Kydor Kropotkin:
Don, this is America, not Russia. WHo could possibly tap every phone in the US?
---------------
dad to son: "Don't you ever bring that gun in the house!" pause then "That's my car gun. My house gun is already in the house, so please return my car gun to the glove compartment."
----------------
Don Masters, CEA Agent:
I'm a CEA agent.
Dr. Sidney Schaefer: [rises from desk, walks over to read ID card]
You ARE a CEA agent. And you really did kill someone!
Don Masters, CEA Agent:
Yeah, and it bothers me sometimes that I don't feel guilty about it. Don't you think that's psychotic behavior?
Dr. Sidney Schaefer:
No I don't! It explains your utter lack of hostility. You can vent your aggressive feelings by actually killing people! It's a sensational solution to the hostility problem.
YAML meets the flat file requirement and can be made to be the persistent DB behind Python, Perl, Ruby.
If it were not for the Flat File requirement then a simple Python shelve or Perl Tie would be the most logical solution since they are both part of the standard library so don't require installing libs on random computers you might use or port to.
of those two Perl Tie is probably the most suited because it's backed by a real DB operating off the disk not fully in memory.
But why not do both: use YAML as the DB backing the Perl Tie.
then you get a nice human readable flat file.
I'm a mac guy too and I am somewhat ticked off by this announcment. For a year or so now they have had a web page saying they are "working hard" to get netflix ondemand service for their mac customers but Apple is preventing them.
How could apple possibly be standing in there way. If they can implement it on a PC or if Amazon NBC and SCIfi channel can get Flash streaming to work why do they have to ask for Apple's permission?
Bah. It was this box they were developing. They had zero interest in making onDemand available to their apple customers.
What is with all this France bashing from the US. You do realize of course that if it were not for France the United States would exist and most likely you would be part of Canada. And if we were part of canada we'd all be speaking english now.
As for the ruling. When do two wrongs make a right? This seems like a very meddlesome court. Sure they are fighting the excesses of bundling when monopoly software is involved. But I'd rather see them attack the root of the problem than set lousy practices like this as precedent.
As they say when elephants fight the grass gets hurt. Asus is the grass.
I compiled the first in situ measurements of the annual temp and pressure cylces on mars (viking lander).
I was always surprised by the mars has water debate when it seemed to me the vapor pressure of the atmosphere was less than the vappor pressure of water.
Thus to my mind if mars had water in any abbundance then it had to be bound up in some mixture that was lowering the vapor pressure.
Apparently there may be another possibility: deep very cold storage.
But either way: no available surface water. No canals. no oceans.
Right. Tell them that you have an obligation to your suppliers and other customers and marketing folks to maintain data security. You can protect this via onion layers on your own web portal and data base set up. But once the DB can potentially be migrated outside, you cannot assure your stakeholdrs you have responsibility for it.
Moreover, if they based financial decisions on their own queries you cannot guarentee them results.
Its just that the simulator for this universe has a cell-size, so anything below a plank length is just being approximated to speed up the calculations. No they tried that and if failed. It turned out that really the simulations of our world are being done in a 6 dimensional world. Since it's six dimensional it's not really a burden on their computers. FOr example, if you lived on a 2-d plane of finite size and tried to simulate another 2-d world, you'd end up like you say having to make the simulation smaller than the world it lives in and hence cell-size effects would pop up and you'd consume a good fraction of all the resources in your 2-d world to represent another 2-d world.
But if you live in a 3-d world then having a bunch of 2-d simmulations is like have a ream of paper. 500 sheets of paper stack up nicely and consume very little of our 3-d world.
in 6-d our 3-d world is a trivial piece of it and computers can easily simmulate it.
No the problem is that there's not an algebraic solution to any polynomial greater than fifth order. Thus they wind up having to numerically approximate the mappings from 6D and this has round off errors from the finite bit floating point representation in Exel 6D.
It's interesting they are only just realizing it. Thermodynamic folks have had to deal with a related issue for a long time.
Almost everything interesting in thermo has to do with a phase transitition popping up somewhere.
THe funny thing is this. There are no phase transitions in the real world. THey only occur on paper continuuum models. However there are a lot of things that look awfully like phase transitions so they are useful to think about.
What am I babbling about. Well phase transitions happen at places where infinite derivatives occur in mappings. And that's all fine on paper where you have an infinite number of states. If you think of states as being something like basis vectors then it' like saying you can write a fourier transform of a square edge with a continuum of frequencies.
But since there's only a finite number of states available to any system, you dont have enough basis vectors to describe a discountinuty.
So phase transitions dont' exist technically speaking. There's always some transition zone around the edge of the transition.
What's the best way to measure one's bandwidth. For example, once you get above 2Mb/sec and live more than 600 miles from a major speed text site, there's not a single speed test meter that works right, in my extensive searches.
Comcast also puts in these 10 second bandwith burst boosts so any test you do has to outlast that if you want to know the sustained rate.
The best way I seem to be able to test things is to find some server and start multiple scp sessions going. But this is plagued by weird artifacts probably having to do with routers at the far end shaping things.
Bit torrent used to be the only way I could actually see anything within a factor of 3 of the bandwidth I pay for. But now I can't even get that speed even when I'm dealing with 100% seeds (I use comcast).
My basic reason for caring is that given I never am able to get within a factor 3 of what I pay for in a sustained way, I'm thinking of downgrading the service level I pay for. But I worry that my service might just get proportionally worse.
Maybe you should report the bug to them and contact them. Anyone can have aproblem. Macs are not problem free. The difference is you know who to blame and who to report the bug to.
I believe their hardware could stand on its own merits and the additional revenue and marketshare couldn't hurt. Microsoft just had an XP SP3 disaster when on some machines with AMD based motherboards the system would endlessly reboot. I have no doubt MS did test this on AMD processors. The problem was that some motherboard makers and vendors improperly used both the intel and amd power management kernel mods simultaneoulsy. This did not show up for 10 years, and SP3.
Apple wants to control the experience. They want to spec high values of hardware. And they don't want to support mutt-hardware and end up like SP3.
What's the number one frustration in calling ANY tech support hotline. Well if you have more than one vendor in the chain then vendor A says it's a graphics card problem, and Vendor B says it's a operating system problem. Meanwhile it's actually a mouse problem because the logitec mouse drivers over wrote some dll the video card was expecting to be an older version.
Not only does no one claim responsibility but they really can't because they don't control it all like apple.
So you pay a tad more for a pleasant experience. Savvy apple folks know which things to buy from apple and which to do themselves. e.g. don't buy apple memory upgrades, but perhaps it may be worth it to buy an apple WiFi (since the system will then handle all the firmware updates for you, and things like optical audio, remote disk mounting over the WiFi will all happen magically and reliably).
As for this latest EFI spoof. Apple, as evidenced by the lack of DRM on their OS and the vulnerable DRM on itunes, tries to use the speedbump model for DRM rather than the Steel Vault model. Any time people start abusing one of their DRMs they tend to issue some new software update that goofs up the current way of gaming the system. Basically a nuiscance which at some point becomes not worth dealing with for the majority of people.
I would predict they have a long road of nuiscance planned for EFI crackers. They only need to plan about 5 years worth of them, because in 5 years there will be new hardware nuiscances that spwan a whole new list of software nuiscances.
Well, if 1970ish is old then fortran libs are alive and kicking. Indeed I was just using a modern Matlab toolkit for computing uniform deviates on the surface of an ellipse. it simply wraps fortran dating to 1973.
This is fresh in my mind because I traced it back to the original published code when I found an error in the math. It computes the deviates wrong! (they are not perfectly uniform).
Ammusingly, I found in the process of doing this search for the origin, that the same code is widely used by global climate modelers to pick random points on the surface of the (ellipsoidal) earth. Maybe that's why 5 day forecasts don't work.
What is it with people. If Linux is better than MS window then surely you won't mind paying more to get it? why are you so fixated on how much people are making and instead no fixated on simply what it's worth to you? so What if linux costs them less to Buy?
Besides which it probably cost them more to implement it than MS since I bet they had to hire an entire Engineering staff, at least one FTE plus support folks.
So If a billion people owned 100 Yotta bytes that's 10^9*10^9*10^9*100*9 = 8E30 bits
there's something like 10^49 atoms on earth, and we'll only be able to access the crust of which only 5% is iron, and 80% of the earth is covered with water. so if we assume as a wild as guess that perhaps a part in a trillion of the earth can be made into disk drives then we have
1E37 atoms available for disk drives.
if each yottbyte drive weighs say 1/5 of a kilo and we assume it's built out mainly carbon and has say a mean weight of 20 amu per atom then this is like 6E21 atoms
therefore one could build no more than 1E15 drives all total.
Thinking about this number it also makes me wonder about how McDonalds got all those hamburgers.
Maybe I boofed the math or assumptions. Good thing this is slashdot and I know people will kindly correct me
positive cash flow does not mean that investors are satisfied with the return. They could put their money in another investment and get a higher ROI. Yahoo's funding will dry up and it's stock price will crash. Microsoft can then snap it up. Additionally, yahoo is still in a growth phase despite the positive cash flow. They won't last long if they cannot raise new capital.
Slashdot Mirror
Is that the one Michael Jackson sings about?
three clicks to to hell:
slashdot
slashdot effect
Larry Niven
Hell
The President's Analyst
Other great lines from the president's analyst:
V.I. Kydor Kropotkin:
Are you trying to tell me every phone in the country is tapped?
Don Masters, CEA Agent:
That's what's in my head.
V.I. Kydor Kropotkin:
Don, this is America, not Russia.
WHo could possibly tap every phone in the US?
---------------
dad to son: "Don't you ever bring that gun in the house!"
pause then "That's my car gun. My house gun is already in the house, so please return my car gun to the glove compartment."
----------------
Don Masters, CEA Agent:
I'm a CEA agent.
Dr. Sidney Schaefer: [rises from desk, walks over to read ID card]
You ARE a CEA agent. And you really did kill someone!
Don Masters, CEA Agent:
Yeah, and it bothers me sometimes that I don't feel guilty about it. Don't you think that's psychotic behavior?
Dr. Sidney Schaefer:
No I don't! It explains your utter lack of hostility. You can vent your aggressive feelings by actually killing people! It's a sensational solution to the hostility problem.
Best line in that movie. "Who could possibly bug every phone in the united states?" TPC
YAML meets the flat file requirement and can be made to be the persistent DB behind Python, Perl, Ruby.
If it were not for the Flat File requirement then a simple Python shelve or Perl Tie would be the most logical solution since they are both part of the standard library so don't require installing libs on random computers you might use or port to.
of those two Perl Tie is probably the most suited because it's backed by a real DB operating off the disk not fully in memory.
But why not do both: use YAML as the DB backing the Perl Tie.
then you get a nice human readable flat file.
You need YAML
I think is exactly what you are requesting.
I'm a mac guy too and I am somewhat ticked off by this announcment. For a year or so now they have had a web page saying they are "working hard" to get netflix ondemand service for their mac customers but Apple is preventing them.
How could apple possibly be standing in there way. If they can implement it on a PC or if Amazon NBC and SCIfi channel can get Flash streaming to work why do they have to ask for Apple's permission?
Bah. It was this box they were developing. They had zero interest in making onDemand available to their apple customers.
I hate liars.
And if not for the US, paris would be called pufter-grad and we'd all be speaking russian
As for the ruling. When do two wrongs make a right? This seems like a very meddlesome court. Sure they are fighting the excesses of bundling when monopoly software is involved. But I'd rather see them attack the root of the problem than set lousy practices like this as precedent.
As they say when elephants fight the grass gets hurt. Asus is the grass.
I compiled the first in situ measurements of the annual temp and pressure cylces on mars (viking lander).
I was always surprised by the mars has water debate when it seemed to me the vapor pressure of the atmosphere was less than the vappor pressure of water.
Thus to my mind if mars had water in any abbundance then it had to be bound up in some mixture that was lowering the vapor pressure.
Apparently there may be another possibility: deep very cold storage.
But either way: no available surface water. No canals. no oceans.
Right. Tell them that you have an obligation to your suppliers and other customers and marketing folks to maintain data security. You can protect this via onion layers on your own web portal and data base set up. But once the DB can potentially be migrated outside, you cannot assure your stakeholdrs you have responsibility for it.
Moreover, if they based financial decisions on their own queries you cannot guarentee them results.
But if you live in a 3-d world then having a bunch of 2-d simmulations is like have a ream of paper. 500 sheets of paper stack up nicely and consume very little of our 3-d world.
in 6-d our 3-d world is a trivial piece of it and computers can easily simmulate it.
No the problem is that there's not an algebraic solution to any polynomial greater than fifth order. Thus they wind up having to numerically approximate the mappings from 6D and this has round off errors from the finite bit floating point representation in Exel 6D.
It's interesting they are only just realizing it. Thermodynamic folks have had to deal with a related issue for a long time.
Almost everything interesting in thermo has to do with a phase transitition popping up somewhere.
THe funny thing is this. There are no phase transitions in the real world. THey only occur on paper continuuum models. However there are a lot of things that look awfully like phase transitions so they are useful to think about.
What am I babbling about. Well phase transitions happen at places where infinite derivatives occur in mappings. And that's all fine on paper where you have an infinite number of states. If you think of states as being something like basis vectors then it' like saying you can write a fourier transform of a square edge with a continuum of frequencies.
But since there's only a finite number of states available to any system, you dont have enough basis vectors to describe a discountinuty.
So phase transitions dont' exist technically speaking. There's always some transition zone around the edge of the transition.
I think this is what they are talking about here.
What's the best way to measure one's bandwidth. For example, once you get above 2Mb/sec and live more than 600 miles from a major speed text site, there's not a single speed test meter that works right, in my extensive searches.
Comcast also puts in these 10 second bandwith burst boosts so any test you do has to outlast that if you want to know the sustained rate.
The best way I seem to be able to test things is to find some server and start multiple scp sessions going. But this is plagued by weird artifacts probably having to do with routers at the far end shaping things.
Bit torrent used to be the only way I could actually see anything within a factor of 3 of the bandwidth I pay for. But now I can't even get that speed even when I'm dealing with 100% seeds (I use comcast).
My basic reason for caring is that given I never am able to get within a factor 3 of what I pay for in a sustained way, I'm thinking of downgrading the service level I pay for. But I worry that my service might just get proportionally worse.
Maybe you should report the bug to them and contact them. Anyone can have aproblem. Macs are not problem free. The difference is you know who to blame and who to report the bug to.
Apple wants to control the experience. They want to spec high values of hardware. And they don't want to support mutt-hardware and end up like SP3.
What's the number one frustration in calling ANY tech support hotline. Well if you have more than one vendor in the chain then vendor A says it's a graphics card problem, and Vendor B says it's a operating system problem. Meanwhile it's actually a mouse problem because the logitec mouse drivers over wrote some dll the video card was expecting to be an older version.
Not only does no one claim responsibility but they really can't because they don't control it all like apple.
So you pay a tad more for a pleasant experience. Savvy apple folks know which things to buy from apple and which to do themselves. e.g. don't buy apple memory upgrades, but perhaps it may be worth it to buy an apple WiFi (since the system will then handle all the firmware updates for you, and things like optical audio, remote disk mounting over the WiFi will all happen magically and reliably).
As for this latest EFI spoof. Apple, as evidenced by the lack of DRM on their OS and the vulnerable DRM on itunes, tries to use the speedbump model for DRM rather than the Steel Vault model. Any time people start abusing one of their DRMs they tend to issue some new software update that goofs up the current way of gaming the system. Basically a nuiscance which at some point becomes not worth dealing with for the majority of people.
I would predict they have a long road of nuiscance planned for EFI crackers. They only need to plan about 5 years worth of them, because in 5 years there will be new hardware nuiscances that spwan a whole new list of software nuiscances.
Well, if 1970ish is old then fortran libs are alive and kicking. Indeed I was just using a modern Matlab toolkit for computing uniform deviates on the surface of an ellipse. it simply wraps fortran dating to 1973.
This is fresh in my mind because I traced it back to the original published code when I found an error in the math. It computes the deviates wrong! (they are not perfectly uniform).
Ammusingly, I found in the process of doing this search for the origin, that the same code is widely used by global climate modelers to pick random points on the surface of the (ellipsoidal) earth. Maybe that's why 5 day forecasts don't work.
It take HP 30 bits to show color? Ha! my old Apple II could do it in just 8 bits. HP has a lot of catching up to do.
Part of the project actually is to determine what can make the game more fun. For me the funnest part is competing against others.
What is it with people. If Linux is better than MS window then surely you won't mind paying more to get it? why are you so fixated on how much people are making and instead no fixated on simply what it's worth to you? so What if linux costs them less to Buy?
Besides which it probably cost them more to implement it than MS since I bet they had to hire an entire Engineering staff, at least one FTE plus support folks.
Fixing my own math. I see TFS is wrong and a yottabyte is not 10^18 but is in fact 10^24. So is a billion people have a drive that's
1E34 drives.
so that's 100 atoms per drive.
Some how I don't think so.
So If a billion people owned 100 Yotta bytes that's 10^9*10^9*10^9*100*9 = 8E30 bits
there's something like 10^49 atoms on earth, and we'll only be able to access the crust of which only 5% is iron, and 80% of the earth is covered with water. so if we assume as a wild as guess that perhaps a part in a trillion of the earth can be made into disk drives then we have
1E37 atoms available for disk drives.
if each yottbyte drive weighs say 1/5 of a kilo and we assume it's built out mainly carbon and has say a mean weight of 20 amu per atom then this is like
6E21 atoms
therefore one could build no more than
1E15 drives all total.
Thinking about this number it also makes me wonder about how McDonalds got all those hamburgers.
Maybe I boofed the math or assumptions. Good thing this is slashdot and I know people will kindly correct me
positive cash flow does not mean that investors are satisfied with the return. They could put their money in another investment and get a higher ROI. Yahoo's funding will dry up and it's stock price will crash. Microsoft can then snap it up. Additionally, yahoo is still in a growth phase despite the positive cash flow. They won't last long if they cannot raise new capital.