But that freedom is not really intended for YOU, just for the carrier and handset manufacturer. Android TOS is getting tighter and tighter.
But in general I agree with your assessment of Apple's absurdly paranoid lock down of the phone and withholding features that the hardware is fully capable of simply because "Steve says NO".
I have come to believe that the Apple iPhone software development team is an extremely small group of people with the real talent concentrated in an even smaller subset. There is no other rational reason NOT to introduce features in software that the hardware is already capable of.
Face it: your side lost. "fsync everywhere" is an infeasible, untenable, and useless position to take.
And had it been enforced, as soon as all developers went thru and added the fsync calls everywhere it would have become necessary for file system maintainers to no-op fsync calls in order to regain any approximation of prior performance.
Flushing "one file" is not always sufficient. Calling fsync() does not necessarily ensure that the entry in the directory containing the file has also reached disk. For that an explicit fsync() on a file descriptor for the directory is also needed. And perhaps the higher level directory as well.
My point is that when you have a system that is so open to corruption, with so few checks and balances, and so much baggage inherited from institutions that began in the dark ages, it's no surprise that you end up with science that's much less than perfect.
Its also no surprise when you end up with science that is horribly incomplete.
We need to place more emphasis on using the internet as a repository for non-published works. (Like DeepDyve http://www.deepdyve.com/corp/about ).
With this comes the boogie man of the kook "scientist". (Which unfortunately includes any scientist who is not yet published).
We need to start using something like the Web of Trust found in key signing http://en.wikipedia.org/wiki/Web_of_trust to document the credentials of scientists without regard to the content of any specific work.
(Scientists A, B, C, and Institution 1, 2, and 5 sign Professor X's credentials certifying that they know him to possess the training and education to conduct studies in his field, without any indication of approval or disapproval of his current work, but with due regard for any past work of which they may be aware).
With a web of trust you would be able to distinguish the kooks (those with closed webs of trust) from the real scientists (those with open and expanding webs).
There are more than a few incidents of Meta Analysis including the same data set from multiple places simply because it was shopped around for publication under different names.
Meta Analysis combines vaguely related studies, using data sets of suspect quality, which you don't fully understand, which have already undergone filtering and editing you won't find out about, and which were collected under conditions you don't know, for motives you can't be sure of, by people you don't know, and can't possibly trust, which purport to measure issues only approximately related to the sibling studies with which they are homogenized.
"Grubb & Ellisâ(TM) National Data Center Group, who said he can currently identify only six options in the United States for data center users who need more than 7 megawatts of power or more than 50,000 square feet of contiguous space".
This sounds more like mainframe installation space than COLO space. There is simply no need to concentrate COLO farms to this degree.
Sounds fishy to me. Power demands of rack servers have been falling dramatically for years. There would be no problem swapping 4 or 5 servers onto a single U1 using virtual machine technology and using less power over all and requiring less fiber too.
Maybe the problem of said Telco was failure to keep up with technology.
And "Key Markets" makes no sense in the internet world.
Unless of course the "Key Markets" seeing this shortage are having a hard time finding sites willing to COLO their kiddy porn servers for fear of having the entire place seized by US Marshals.
Does it matter if your server is across the country or across the state any more?
Whats the chances that the center you visited was designed for prior generations of hardware, and the current dense racks simply don't need that much space?
Did Virtual Servers even exist when the contract was let?
> also something supply/demand models do not account for.
I'm sorry, but thats just wrongheaded.
Supply and Demand does not "account" for anything. Price does. Lead time, credit availability, competition, quality, speed, permitting issues and, yes, Supply and Demand are ALL accounted for by the final Price. And so is behavior, reluctance to take risk.
It is all encapsulated in the price.
I suggest you have left too much of Econ 101 in the classroom, if you ever attended at all.
That is Not a significant problem here (as far as anyone knows).
Voter registration is a function of State Government.
The ballots are sent and counted by the County Government.
There are reasonable (but not foolproof) checks on the number they of people that can be registered at a given address at the State level.
Also, you might be confusing the problem of fraudulent voter registration with the problem of vote counting.
They are related problems, for sure, but not quite the same thing. If the voter registration system is subject to gaming, the actual voting process is already compromised, and party hacks marching from polling place to polling place with fraudulent IDs has happened often enough to cast doubt on In-Person voting as well.
Temporary shortages of things like this are the ultimate Non-Story.
If we want to really HoZe this industry, let the government step in. Otherwise will the hand wringers please shut up and sit down and let the market handle this!?
Yes, financing may be tight, unless you are Google or Amazon, or Apple, or IBM, or Microsoft. But it it IS a problem you haven't built your business case very well. There are Venture Capitalists out there roaming around looking for ways to put their money to work, and infrastructure is a lot more reliable then next weeks "big thing" software project.
My experience, limited as it may be, suggests that Hawaii is almost, but not quite, thoroughly unlike Texas.
That being said, my County (in Washington state) votes by mail. My ballot finds me. I don't have to find it. Return Postage pre-paid. Sign the outer envelope, secret ballot in the machine open-able inner envelope.
Can this be gamed? Probably, if someone wanted to add the Federal offense of mail theft to the (apparently ignored) crime of Election Fraud, but to do so on any massive scale would be pretty hard.
> no clear way to capitalize on this without being traced.
You presume a level of diligence that does not exist. We can't even get botnets shut down in this country when we know exactly which computers have been compromised, let along be able to trace the problem to the source.
Its really GOOD for the deplorable propagation of inappropriate and insecure voting technology. It should nip it in the bud!
Compared to this scheme, Diebold was an example of bullet proof security. Hacking Diebold for the most part still required physical access to the machines or their memory. Cracking internet voting can be done from the safety for some Russian bot net master's basement.
Slashdot Mirror
But that freedom is not really intended for YOU, just for the carrier and handset manufacturer. Android TOS is getting tighter and tighter.
But in general I agree with your assessment of Apple's absurdly paranoid lock down of the phone and withholding features that the hardware is fully capable of simply because "Steve says NO".
I have come to believe that the Apple iPhone software development team is an extremely small group of people with the real talent concentrated in an even smaller subset. There is no other rational reason NOT to introduce features in software that the hardware is already capable of.
Go get her a Disney cell phone. Parental managed GPS tracking from the web, call limiting, (you can lock it down to only being able to call home).
Google that.
Apple would never leak such utterly crappy shots.
http://www.t3.com/news/android-cupcake-1-5-update-to-enable-video-capture-for-g1?=38641
The hardware is capable, but the software is not yet willing. That will be changed, - see above.
Note that this is the EXACT same situation as the iPhone. You have been able to shoot video on the iPhone once jailbroken.
No. But this seems the wrong place to hid this question.
We avoid anything that has less than 24 months of wide deployment unless there is some absolute pressing need to
Good Idea. Let's all follow this sage advice.
Face it: your side lost. "fsync everywhere" is an infeasible, untenable, and useless position to take.
And had it been enforced, as soon as all developers went thru and added the fsync calls everywhere it would have become necessary for file system maintainers to no-op fsync calls in order to regain any approximation of prior performance.
Flushing "one file" is not always sufficient. Calling fsync() does not necessarily ensure that the entry in the directory containing the file has also reached disk. For that an explicit fsync() on a file descriptor for the directory is also needed. And perhaps the higher level directory as well.
My point is that when you have a system that is so open to corruption, with so few checks and balances, and so much baggage inherited from institutions that began in the dark ages, it's no surprise that you end up with science that's much less than perfect.
Its also no surprise when you end up with science that is horribly incomplete.
We need to place more emphasis on using the internet as a repository for non-published works. (Like DeepDyve http://www.deepdyve.com/corp/about ).
With this comes the boogie man of the kook "scientist". (Which unfortunately includes any scientist who is not yet published).
We need to start using something like the Web of Trust found in key signing http://en.wikipedia.org/wiki/Web_of_trust to document the credentials of scientists without regard to the content of any specific work.
(Scientists A, B, C, and Institution 1, 2, and 5 sign Professor X's credentials certifying that they know him to possess the training and education to conduct studies in his field, without any indication of approval or disapproval of his current work, but with due regard for any past work of which they may be aware).
With a web of trust you would be able to distinguish the kooks (those with closed webs of trust) from the real scientists (those with open and expanding webs).
This would allow us at least some clue as to credentials and knowledge of the scientist under discussion rather than the mere presence of an article in a journal of questionable value. http://blog.bioethics.net/2009/05/merck-makes-phony-peerreview-journal/
Scientific fraud is much harder to find, though no less important, because it has no real world affects.
Sure it does. It just takes longer to surface. It ends up costing other people a lot of time and money. The Scientist tends to get away scott free.
Meta Analysis does that.
There are more than a few incidents of Meta Analysis including the same data set from multiple places simply because it was shopped around for publication under different names.
Meta Analysis combines vaguely related studies, using data sets of suspect quality, which you don't fully understand, which have already undergone filtering and editing you won't find out about, and which were collected under conditions you don't know, for motives you can't be sure of, by people you don't know, and can't possibly trust, which purport to measure issues only approximately related to the sibling studies with which they are homogenized.
It amounts to rigorous analysis of a turd.
I find it odd that this is all based on Meta Analysis, which itself is still highly suspect.
> "would you eat pork with human genes in it"
But I draw the line at mice.
Not everyone does. http://www.fugly.com/videos/6644/guy_eats_live_mouse.html
Story said:
"Grubb & Ellisâ(TM) National Data Center Group, who said he can currently identify only six options in the United States for data center users who need more than 7 megawatts of power or more than 50,000 square feet of contiguous space".
This sounds more like mainframe installation space than COLO space. There is simply no need to concentrate COLO farms to this degree.
Sounds fishy to me.
Power demands of rack servers have been falling dramatically for years. There would be no problem swapping 4 or 5 servers onto a single U1 using virtual machine technology and using less power over all and requiring less fiber too.
Maybe the problem of said Telco was failure to keep up with technology.
And "Key Markets" makes no sense in the internet world.
Unless of course the "Key Markets" seeing this shortage are having a hard time finding sites willing to COLO their kiddy porn servers for fear of having the entire place seized by US Marshals.
Does it matter if your server is across the country or across the state any more?
Whats the chances that the center you visited was designed for prior generations of hardware, and the current dense racks simply don't need that much space?
Did Virtual Servers even exist when the contract was let?
> also something supply/demand models do not account for.
I'm sorry, but thats just wrongheaded.
Supply and Demand does not "account" for anything. Price does.
Lead time, credit availability, competition, quality, speed, permitting issues and, yes, Supply and Demand are ALL accounted for by the final Price. And so is behavior, reluctance to take risk.
It is all encapsulated in the price.
I suggest you have left too much of Econ 101 in the classroom, if you ever attended at all.
The copyright holders and their corrupt organizations may want it.
They use the position to make sure one looks too hard at the invasive digging into people's hard drives and network traffic.
Meanwhile they totally ignore any REAL threats and protection measures. (As can be seen by stories on Slashdot about data thefts left and right).
That is Not a significant problem here (as far as anyone knows).
Voter registration is a function of State Government.
The ballots are sent and counted by the County Government.
There are reasonable (but not foolproof) checks on the number they of people that can be registered at a given address at the State level.
Also, you might be confusing the problem of fraudulent voter registration with the problem of vote counting.
They are related problems, for sure, but not quite the same thing. If the voter registration system is subject to gaming, the actual voting process is already compromised, and party hacks marching from polling place to polling place with fraudulent IDs has happened often enough to cast doubt on In-Person voting as well.
Temporary shortages of things like this are the ultimate Non-Story.
If we want to really HoZe this industry, let the government step in. Otherwise will the hand wringers please shut up and sit down and let the market handle this!?
Yes, financing may be tight, unless you are Google or Amazon, or Apple, or IBM, or Microsoft. But it it IS a problem you haven't built your business case very well. There are Venture Capitalists out there roaming around looking for ways to put their money to work, and infrastructure is a lot more reliable then next weeks "big thing" software project.
My experience, limited as it may be, suggests that Hawaii is almost, but not quite, thoroughly unlike Texas.
That being said, my County (in Washington state) votes by mail. My ballot finds me. I don't have to find it. Return Postage pre-paid. Sign the outer envelope, secret ballot in the machine open-able inner envelope.
Can this be gamed? Probably, if someone wanted to add the Federal offense of mail theft to the (apparently ignored) crime of Election Fraud, but to do so on any massive scale would be pretty hard.
> no clear way to capitalize on this without being traced.
You presume a level of diligence that does not exist. We can't even get botnets shut down in this country when we know exactly which computers have been compromised, let along be able to trace the problem to the source.
You missed the part of TFA that provided Telephone Voting?
I disagree.
Its really GOOD for the deplorable propagation of inappropriate and insecure voting technology. It should nip it in the bud!
Compared to this scheme, Diebold was an example of bullet proof security. Hacking Diebold for the most part still required physical access to the machines or their memory. Cracking internet voting can be done from the safety for some Russian bot net master's basement.
Since you postulate what does not exist, I suppose you are free to postulate the impossible as well.