What's remarkable is that unlike Avatar or Pixar's latest and greatest is that this is a CG movie that doesn't involve giant expensive renderfarms and really brings CG techniques to the masses of amateur film editors.
You've never needed renderfarms, the limiting factor is time.
Sure, there's a lot of crap independent art(film, music, still visuals, etc), but, this empowers the possibility of a CG equivalent of Clerks or The Blair Witch Project
I thought the whole reason those were "good" is because they accomplished so much with so little... Would "Blair Witch" have been worth your time if it had a CGI witch, or "Clerks" if it actually had a scene of two plumbers arguing on the Death Star? Those films got attention because they were so cheap, and were basically a novelty act, "Blair" in particular.
The primary constraining factor on good motion pictures is the relative scarcity of good writing. The fact that CGI is commoditized is nothing new, thus we have stacks of terrible CGI movies with bad plots, or standard dramas and comedies with dozens of seamless CGI shots (to change eye color, remove wrinkles, add mattes etc.)
This is effectively what the error codes are for. On a raw level there's no such thing as a "missing bit," you either have a zero or a one, but error codes can tell you if it's the correct value or not. If enough of the data and redundant Reed-Solomon codes are on the media, the incorrect value can be corrected, and if there isn't enough, for small errors the player can interpolate. Because a Red Book CD carries a very specific kind of high redundancy data, PCM audio, the reader can then use various strategies to recover something that sounds remotely like what was their originally, if not exactly.
A big drawback of the "interpolation" scheme is that CDs can sound excellent and then suddenly start to fail catastrophically; with tapes and records they would slowly wear down over time, but CDs are much more all-or-nothing. A colleague was telling me recently about engineering sessions in the 80s and having to work with DASH machines, which were big 24 track digital audio tape machines, that used Reed-Solomon codes to allow you to edit the digital tape with razor blades. You had to remove the front panel if you wanted to see the LEDs for the error correction system, and you'd watch those very carefully over the course of the session to make sure these weren't working too hard, and if they were, you'd take the tape and run off a clone before you started having dropouts.
This is very different than a data CD-R -- data CDRs still have error correction and redundant data coding, but if these fail the original file will be corrupt, period, which is why, for added safety, you might create parfiles or something similar.
I guess the upshot of this is that CDs, as they currently operate, particularly CD-Rs but even glass masters, aren't such a hot medium for archival. The failure modes are too severe and can leave you with a moldy loaf instead of half a loaf.
Since CDs have Reed-Solomon and parity for error correction, and even if samples fail the player will interpolate, you can have a pretty ruined disk before it won't play anymore. It is all or nothing once it starts to fail though-- at the point the interpolation can no longer repair a dirty section, the CD will simply drop out.
I also recently (yesterday actually!) opened an old DVD+R (with an HFS volume) from 2002 and rearchived it to a new DVD. It still read perfectly, but it's been stored in a cool dark place, and has been mounted maybe 10 times.
We're talking about using the MAC address outside of the scope of networking here; for all intents and purposes it is a serial number for your computer, readable by any application on the system. No one ever changes (or probably even knows how to change) their MAC address.
No random hey, here's the latest scoop on 0x38df803's location, the local temperature, and the last nine people she called.
As the developer of a hacky (and permanently broken since 4.0) iPhone call log application, I can tell you that the iOS API does not allow an app to access the recently called list. And the local temperature is no trick, since it's an f() of the location.
These work but they use simple gain compression to crunch loud program down -- the tradeoff is that it makes legitimate loud content sound terrible. The channel is capable of a certain dynamic range and the people who make the television programs, the actual shows, want to be able to make things as loud or as quiet as the medium will tolerate. The problem with audio automatic gain correction is that it doesn't know when the audio is from a commercial (loud is a nuisance) or from the programming (loud is desirable), without elaborate and brittle hacks.
How about instead of legislating the details of the presentation layer, the law simply mandated that commercials (actually, all broadcast television) be sent with appropriate metadata that identifies the content.
They do, it's called DIALNORM, and the advertisers simply game the setting in order to pump their stuff.
In a first-past-the-pole winner take all contest, your best strategy is to vote for the viable candidate most closely aligned to your interests. Voting for a third party candidate, who cannot win because he simply lacks support, when someone who shares most of your positions loses due to the lack of your vote, is just a waste.
This is just how this sort of system works -- if you want every vote to "count" you should be agitating for a more proportional allocation of votes to issues.
Pointers. Some people just can't this concept. Dereferencing, etc, are most likely alien concepts. Also, header files and the general mechanics of Obj-C/C/C++ are very different from Java/.NET type languages.
There definitely exists a class of programmers who don't grok things like "functional programming" or "actor model with multiple dispatch" or "tail calls," but for the ones who don't understand how a "pointer reference" works, there is little hope. Even PHP has references.
Also things like "+" and "-" to differentiate class vs instance methods seems silly.Why not use static like everything else does? And whats the deal with the @property and @synthesize stuff? Not really a fan.
They had to add features in such a way that any C files was also a valid Objective-C file. That was the design decision. Also, class methods in ObjC aren't static, they bind to the caller's class, not the class they're defined in. If you don't like "@property", you will get really pissed off by Java annotations, which are another example of a non-compulsory feature that "makes the code ugly" and many people "aren't really a fan of" but accomplishes a specific purpose in a practical way.
Oppenheimer was an American born in New York City. Einstein took the oath to achieve American citizenship in 1940.
Oppenheimer was an astrophysicist who was hired for his administrative abilities, Einstein had nothing to do with the atom bomb program, aside from signing a letter (which he did not write). Niels Bohr, Enrico Fermi, Teller, Ulam, Von Newmann, Bethe all left Europe and became Americans, it is true, but it's important to recognize they came to America for essentially negative reasons -- their home wouldn't tolerate them anymore. If Germany had been merely totalitarian and persecuted Poles instead of Jews, do we dare guess how many of "our" atomic scientists would have simply stayed in Germany? Most of these people also made their critical insights while still in Europe under the auspices of European governments, like Lise Meitner.
By that logic, the Cypriots must have the most powerful military in the world.
This doesn't follow.
Not really. The problems within the USSR were largely caused by pressures due to their participation within the Cold War. In a sense, the U.S. won the Cold War by out-producing the Soviets.
This is still debated, and even granting that it's true, it's basically impossible to apply this lesson to conflicts with, say Iraq or Iran. Or Al Qaeda. I was reading a quote from George F. Kennan recently:
The suggestion that any American administration had the power to influence decisively the course of a tremendous domestic-political upheaval in another great country on another side of the globe is intrinsically silly and childish. No great country has that sort of influence on the internal developments of any other one.
I'll take an actual cold warrior's opinion over some glib, handwaving slashdotter.
The First Gulf War was nothing but a display of muscle to show Saddam Hussein that he didn't know who he was messing with.
And an opportunity for us to promise to come to the aid of anti-governemnt Kurds and Shiites in the North and South, which we promptly refused to support and allowed to be slaughtered, belatedly imposing no-fly zones. And an opportunity for the US and UK to impose ineffective and internally radicalizing sanctions which hollowed out Iraqi society. And occasionally drop bombs under the auspices of "Desert Fox" et al. And draw Hussein into closer alliances with muslim militants.
It depends on how you define success. If by "success" you mean did the U.S. achieve regime change? No failure there. If by "success" do you mean did the U.S. achieve peace in Iraq? If so, I'm fairly sure that was never a goal of the U.S. military.
As long as we define success in terms that would be unrecognizable to someone who was present at the decision to go to war, we have succeeded. And it only cost $800 billion and a few hundred thousand lives, and we are left with a nation state that teeters on the edge of sectarian civil war, and will likely settle as a client of Iran.
Again, no. The goals in Afghanistan were: 1) overthrow the Taliban (check) 2) bring various members of Al Qaeda to justice (check) 3) capture Osama Bin Laden. The status of the 3rd item is, at best, inconclusive, but the other 2 goals have been largely achieved.
And it only cost $300 million, maybe 40k lives, and has occupied our military for 9 years. The magic thing about war, of course, is that it evades all cost-benefit analysis. No matter how many hajis you kill, it never seems to make the cockpit doors any stronger.
But let's not beat around the bush. The project of redefining success is to protect the stainless reputation of our military, despite the fact that the US's strategic position in the world has been in
Military social structure uses bizarre and arbitrary rights of passage, shibboleths, coded jargon (like "war fighter," because too many people think they know what a "soldier" is, means, and does), and social signifiers in order to maintain hierarchy and moral legitimacy. Shocking.
Recourse to audio-visual aids, like videos, "presentations" and PowerPoint are a serious indicator that the speaker does not completely understand the concept involved, or is attempting to force an unsupportable conclusion.
Not in this case, though. I'm currently seeding the video...
For someone only partially familiar with video and audio encoding, this was a particularly clear and informative video. It also serves as an excellent example of the direction more Open Source efforts need to take. Mini lectures that bring some human explanation.
That's fair, and of course they're no doubt trying to emulate the success projects like Ruby on Rails have had with demonstration webcasts, but really what many people are looking for is something they can sit down with a cuppa coffee and a tuna sandwich and read. I doesn't help that there's really no reading material whatsoever on open source audio and video coding efforts, aside from man pages and forums -- completely unstructured and un-comprehensive.
Given that the closed source, proprietary society model is rapidly taking over everything else, those who want a modicum of freedom expression and fair markets...
If there is anything more annoying than the vegan-like moral posturing of open formats advocates, it's their unrelenting persecution complex at the hands of the Big Boys, and if there's anything more annoying than their persecution complex, it's their nervous libertarian tick of always justifying what they want in terms of "markets."
The problem with a transcript is the information density is very low, because it's still built around the typical talking-head conventions. I saw the wiki page, and it's exactly what I didn't want. I wanted what they wanted to say, not what everyone on the internet helpfully wants to add.
I wish they'd just written an ebook, I think sometimes people lose sight of trying to impart useful information, and get wrapped up in making the information "fun and accessible." You could probably get twice the amount of information reading in an hour rather than watching someone mug in a video -- making it "entertaining" will only make the information more accessible to people who are likely never to use it.
PS. Available as OGG and WebM. Is Xiph working for Google now?
“Though the aircraft is not a practical method of transport, it is also meant to act as an inspiration to others to use the strength of their body and the creativity of their mind to follow their dreams.”
There you go, it ain't much, but then again creativity is a pretty expensive and scarce commodity.
And 'cheap underpowered laptops' are what a lot of us bought them for. Cheap so I don't have to justify buying one and don't have to be too upset if I lose it, small so I can carry it whenever I travel, and low-powered so the battery lasts for much longer than a typical full-size laptop.
Most people won't buy a "cheap underpowrered laptop," but they will buy an expensive tablet with teh snappy. It's really about expectations -- something that's "cheap" generally won't sell to a large segment of a market for any price, because of the implication of shoddy quality. There's a difference between a "shoddy" computer and merely "slow" one of course, but it's not really relevant to most people, and something Apple tends to bery good at is calibrating expectations through clever positioning.
(Nota bene, I think Apple might actually be the only technology company that even tries to position its products to actual use cases, price points and user expectations, and doesn't just sell its products with a generic "It does what the competitors does, only better!")
Instead of regulating prudence, the government basically forced banks to support untenable inner city mortgages which should have never been approved in the first place.
The assertion that the Community Reinvestment Act caused the housing bubble and ensuing recession is a disproven canard.
Thus:
The Community Reinvestment Act (CRA) encourages banks to expand mortgage lending
in the communities in which they have branch offices, subject to maintaining overall levels of
financial safety and soundness. Some have argued that this regulation forced banks to lower their
credit standards and engage in riskier mortgage products in order to extend credit to lower-income
individuals, who perhaps should not have received such loans. However, data provided by the
Home Mortgage Disclosure Act (HMDA) reveal that loans covered by the CRA accounted for
only a fraction of mortgage lending to lower-income borrowers and neighborhoods. This is
especially true of higher-priced, or subprime, mortgages.1 CRA assessment-area lending
accounted for only nine percent of higher-priced loans to lower-income borrowers and
neighborhoods, while independent mortgage companies accounted for the majority. Further, the
subprime share of assessment-area loans made to lower-income borrowers and lower-income
neighborhoods was lower than the subprime share for all loans made between 2004 and 2006.
The vast majority of subprime home mortgage lending was made by organizations not bound by the CRA. Not to mention lending for commercial real estate development.
History never repeats, but it does rhyme. It seems like the last 20 years in information technology have been nothing but people reinventing a college campus Unix infrastructure, except over HTML instead of VT100. Most of the real "innovation" has been in business models and the way the help desk works.
Slashdot Mirror
You've never needed renderfarms, the limiting factor is time.
I thought the whole reason those were "good" is because they accomplished so much with so little... Would "Blair Witch" have been worth your time if it had a CGI witch, or "Clerks" if it actually had a scene of two plumbers arguing on the Death Star? Those films got attention because they were so cheap, and were basically a novelty act, "Blair" in particular.
The primary constraining factor on good motion pictures is the relative scarcity of good writing. The fact that CGI is commoditized is nothing new, thus we have stacks of terrible CGI movies with bad plots, or standard dramas and comedies with dozens of seamless CGI shots (to change eye color, remove wrinkles, add mattes etc.)
This is effectively what the error codes are for. On a raw level there's no such thing as a "missing bit," you either have a zero or a one, but error codes can tell you if it's the correct value or not. If enough of the data and redundant Reed-Solomon codes are on the media, the incorrect value can be corrected, and if there isn't enough, for small errors the player can interpolate. Because a Red Book CD carries a very specific kind of high redundancy data, PCM audio, the reader can then use various strategies to recover something that sounds remotely like what was their originally, if not exactly.
A big drawback of the "interpolation" scheme is that CDs can sound excellent and then suddenly start to fail catastrophically; with tapes and records they would slowly wear down over time, but CDs are much more all-or-nothing. A colleague was telling me recently about engineering sessions in the 80s and having to work with DASH machines, which were big 24 track digital audio tape machines, that used Reed-Solomon codes to allow you to edit the digital tape with razor blades. You had to remove the front panel if you wanted to see the LEDs for the error correction system, and you'd watch those very carefully over the course of the session to make sure these weren't working too hard, and if they were, you'd take the tape and run off a clone before you started having dropouts.
This is very different than a data CD-R -- data CDRs still have error correction and redundant data coding, but if these fail the original file will be corrupt, period, which is why, for added safety, you might create parfiles or something similar.
I guess the upshot of this is that CDs, as they currently operate, particularly CD-Rs but even glass masters, aren't such a hot medium for archival. The failure modes are too severe and can leave you with a moldy loaf instead of half a loaf.
I also recently (yesterday actually!) opened an old DVD+R (with an HFS volume) from 2002 and rearchived it to a new DVD. It still read perfectly, but it's been stored in a cool dark place, and has been mounted maybe 10 times.
We're talking about using the MAC address outside of the scope of networking here; for all intents and purposes it is a serial number for your computer, readable by any application on the system. No one ever changes (or probably even knows how to change) their MAC address.
As the developer of a hacky (and permanently broken since 4.0) iPhone call log application, I can tell you that the iOS API does not allow an app to access the recently called list. And the local temperature is no trick, since it's an f() of the location.
These work but they use simple gain compression to crunch loud program down -- the tradeoff is that it makes legitimate loud content sound terrible. The channel is capable of a certain dynamic range and the people who make the television programs, the actual shows, want to be able to make things as loud or as quiet as the medium will tolerate. The problem with audio automatic gain correction is that it doesn't know when the audio is from a commercial (loud is a nuisance) or from the programming (loud is desirable), without elaborate and brittle hacks.
They do, it's called DIALNORM, and the advertisers simply game the setting in order to pump their stuff.
In a first-past-the-pole winner take all contest, your best strategy is to vote for the viable candidate most closely aligned to your interests. Voting for a third party candidate, who cannot win because he simply lacks support, when someone who shares most of your positions loses due to the lack of your vote, is just a waste.
This is just how this sort of system works -- if you want every vote to "count" you should be agitating for a more proportional allocation of votes to issues.
They must be running their webserver on an Arduino.
There definitely exists a class of programmers who don't grok things like "functional programming" or "actor model with multiple dispatch" or "tail calls," but for the ones who don't understand how a "pointer reference" works, there is little hope. Even PHP has references.
They had to add features in such a way that any C files was also a valid Objective-C file. That was the design decision. Also, class methods in ObjC aren't static, they bind to the caller's class, not the class they're defined in. If you don't like "@property", you will get really pissed off by Java annotations, which are another example of a non-compulsory feature that "makes the code ugly" and many people "aren't really a fan of" but accomplishes a specific purpose in a practical way.
Oppenheimer was an astrophysicist who was hired for his administrative abilities, Einstein had nothing to do with the atom bomb program, aside from signing a letter (which he did not write). Niels Bohr, Enrico Fermi, Teller, Ulam, Von Newmann, Bethe all left Europe and became Americans, it is true, but it's important to recognize they came to America for essentially negative reasons -- their home wouldn't tolerate them anymore. If Germany had been merely totalitarian and persecuted Poles instead of Jews, do we dare guess how many of "our" atomic scientists would have simply stayed in Germany? Most of these people also made their critical insights while still in Europe under the auspices of European governments, like Lise Meitner.
By that logic, the Cypriots must have the most powerful military in the world.
This doesn't follow.
This is still debated, and even granting that it's true, it's basically impossible to apply this lesson to conflicts with, say Iraq or Iran. Or Al Qaeda. I was reading a quote from George F. Kennan recently:
I'll take an actual cold warrior's opinion over some glib, handwaving slashdotter.
And an opportunity for us to promise to come to the aid of anti-governemnt Kurds and Shiites in the North and South, which we promptly refused to support and allowed to be slaughtered, belatedly imposing no-fly zones. And an opportunity for the US and UK to impose ineffective and internally radicalizing sanctions which hollowed out Iraqi society. And occasionally drop bombs under the auspices of "Desert Fox" et al. And draw Hussein into closer alliances with muslim militants.
As long as we define success in terms that would be unrecognizable to someone who was present at the decision to go to war, we have succeeded. And it only cost $800 billion and a few hundred thousand lives, and we are left with a nation state that teeters on the edge of sectarian civil war, and will likely settle as a client of Iran.
And it only cost $300 million, maybe 40k lives, and has occupied our military for 9 years. The magic thing about war, of course, is that it evades all cost-benefit analysis. No matter how many hajis you kill, it never seems to make the cockpit doors any stronger.
But let's not beat around the bush. The project of redefining success is to protect the stainless reputation of our military, despite the fact that the US's strategic position in the world has been in
Military social structure uses bizarre and arbitrary rights of passage, shibboleths, coded jargon (like "war fighter," because too many people think they know what a "soldier" is, means, and does), and social signifiers in order to maintain hierarchy and moral legitimacy. Shocking.
This is a user manual, it has very little for the developer.
Recourse to audio-visual aids, like videos, "presentations" and PowerPoint are a serious indicator that the speaker does not completely understand the concept involved, or is attempting to force an unsupportable conclusion.
Not in this case, though. I'm currently seeding the video...
That's fair, and of course they're no doubt trying to emulate the success projects like Ruby on Rails have had with demonstration webcasts, but really what many people are looking for is something they can sit down with a cuppa coffee and a tuna sandwich and read. I doesn't help that there's really no reading material whatsoever on open source audio and video coding efforts, aside from man pages and forums -- completely unstructured and un-comprehensive.
If there is anything more annoying than the vegan-like moral posturing of open formats advocates, it's their unrelenting persecution complex at the hands of the Big Boys, and if there's anything more annoying than their persecution complex, it's their nervous libertarian tick of always justifying what they want in terms of "markets."
Sorry, I'm in a fuddy duddy mood tonight, I commend your patience.
The problem with a transcript is the information density is very low, because it's still built around the typical talking-head conventions. I saw the wiki page, and it's exactly what I didn't want. I wanted what they wanted to say, not what everyone on the internet helpfully wants to add.
I wish they'd just written an ebook, I think sometimes people lose sight of trying to impart useful information, and get wrapped up in making the information "fun and accessible." You could probably get twice the amount of information reading in an hour rather than watching someone mug in a video -- making it "entertaining" will only make the information more accessible to people who are likely never to use it.
PS. Available as OGG and WebM. Is Xiph working for Google now?
I'd like to think the carry-alls had wings, too, but you're right, I believe the book is silent on the issue.
There you go, it ain't much, but then again creativity is a pretty expensive and scarce commodity.
Just in time for Yueh to leave us a pair of stillsuits in the back. The article doesn't mention if it's big enough to lift a spice harvester however.
Oh a million deaths are not enough for Yueh!
Most people won't buy a "cheap underpowrered laptop," but they will buy an expensive tablet with teh snappy. It's really about expectations -- something that's "cheap" generally won't sell to a large segment of a market for any price, because of the implication of shoddy quality. There's a difference between a "shoddy" computer and merely "slow" one of course, but it's not really relevant to most people, and something Apple tends to bery good at is calibrating expectations through clever positioning.
(Nota bene, I think Apple might actually be the only technology company that even tries to position its products to actual use cases, price points and user expectations, and doesn't just sell its products with a generic "It does what the competitors does, only better!")
The assertion that the Community Reinvestment Act caused the housing bubble and ensuing recession is a disproven canard.
Thus:
The vast majority of subprime home mortgage lending was made by organizations not bound by the CRA. Not to mention lending for commercial real estate development.
History never repeats, but it does rhyme. It seems like the last 20 years in information technology have been nothing but people reinventing a college campus Unix infrastructure, except over HTML instead of VT100. Most of the real "innovation" has been in business models and the way the help desk works.
If you want to read a big-overview-of-all-physics book like this, except from a non-crackpot, consider Roger Penrose's The Road to Reality .