I was looking for somewhere to attach this comment, and you're it.
"Cloud" is the modern term for a mainframe, time-sharing-like model.
One advantage is that your data lives on a server somewhere, meaning someone else is responsible for backup, and you can access it from any "terminal" (typically a web browser, but could also be things like the Steam client).
Another advantage is a potential pricing model for developers -- Amazon EC2 charges per hour of server time used, at a very flat rate. If you only use an hour, you only pay ten cents.
The big advantage of an infrastructure like EC2 is shown in pathological cases, like websites which tend to receive more traffic at certain times of the day. So every night, you can shut down whatever capacity you don't need, and stop paying for it -- and Amazon can then allocate it to someone else who needs it at that time, possibly overnight.
At a different level, you see the same pattern with web applications -- you don't need a computer more powerful than it needs to be to run Firefox. The server can do whatever computing you need that isn't already happening locally -- but most GUI apps spend a lot of time waiting for the user. So when you do a search in Gmail, that takes some server CPU -- but while you're examining the results, that server is off running someone else's search.
Here's the problem: None of these advantages apply to these guys.
The "my data is elsewhere" advantage is irrelevant. Steam already provides this. So long as I remember a username and password, I can download all my Steam games, along with all their savegames and settings.
The idea that a piece of hardware might not fully be utilized by a single user, and could thus be re-allocated, is similarly irrelevant here. Unless they have some sort of weird economies of scale where one video card can serve a thousand users, and cost less than a thousand times the cost of one normal video card... they're pretty much stuck with one machine per gamer.
And since these machines will have to be geographically close to the gamer to be at all viable, there's going to be very little gain from half the gamers going to sleep just as the other half wakes up. You're still going to get the bulk of your traffic from large groups of gamers coming home and logging in at about the same time.
The only advantage is the not-having-to-think-about-maintenance bit, which is pretty weak against consoles. A console is something even John Q. Gamer can unpack and plug in himself. Having to do it every four years is really not that big a deal.
So that's a very long way of saying: I agree with you, "cloud" is being abused. This is clearly someone trying to cash in on the buzzword, without really understanding what it's good for -- it would be like creating an XML representation of a waveform from a sound file.
This doesn't mean XML is worthless, or that it lacks meaning. It just means that someone drank a little too much kool-aid.
Similarly, "cloud computing", as vague as it can be, is really about a couple of related concepts that are concrete enough to write down. This is just something that it's really not suited for.
However, even a fiber optics line I'd have my doubts.
Doing some quick calculations:
The highest number I've gotten for Blu-Ray maximum bandwidth is 54 megabits per second. I've seen torrents much smaller that still looked good.
Assuming uncapped, that's actually doable. Fiber is typically 100 mbits per second, and I'm sure some places offer gigabit.
However, encoding time is on the order of hours or days, and is certainly not live. So the real problem is latency -- take 50 ms from your LCD monitor, plus whatever a wireless controller ads, plus the latency between you and their servers, plus the lag for them to render, capture, and encode, then decode back at the client... that's easily getting up to 200 ms, which I'd consider unplayable.
Also, unless the $50/year includes games, it makes little economic sense, either. These systems are designed to last some four years or so. A Wii can be had for $160, according to a quick Google; this would be $200. A Wii can work when your Internet is down, or when your internet is not fiber. And a Wii actually has games already -- not as many as its competitors, but some.
Where I could see this working is in a LAN environment
Not really. LANs are typically 100 mbits, or if you're willing to spend money on a good switch, gigabit. Same situation as fiber.
The only advantage of a LAN is, with a good switch, you aren't using everyone else's bandwidth, but if you're proposing this:
make some kind of "xbox360server" to host all the games as basically virtual machines across a lan,
That's still likely to be a single port, which means now everyone on the LAN is limited to a combined 100 mbits for their video. It means the concept of a LAN party just got very, very impractical.
And WTF would be the point, if it's a console anyway? In what way is that "xbox360server" better than a real Xbox 360?
As for their "no piracy" claim, as a consumer, that doesn't make me want to sign up for the service. That makes me want to go far away, into the open arms of indie developers, who typically ship with reduced or no DRM.
For example, in part 5, while trying to refute Hovind's ice shield theory, he uses the diagram to estimate how big Hovind's ice shield around the earth is and continues with this size for the rest of the video. But previously in the video Hovind had said he didn't know how big it was but guessed at "10 or 20 or 30 inches thick."
The problem is that Hovind was not consistent. 10 or 20 or 30 inches would've cracked apart due to barometric pressure changes. The only other numbers we have to work with are the diagram, which shows much more water than there's really room for on the surface of the planet.
Follow the calculation, and the diagram is actually showing less ice than you would need for the shield to last as long as Hovind claims it did (2000 years). So, you cannot put enough ice there to survive that long, without also putting enough ice there to have the planet flooded under about a thousand kilometers of water.
Ok, that gives you the flood, but after the first kilometer of ice, the planet would be completely dark before that flood. It also doesn't explain where all that water went -- how did the flood end?
That doesn't cover all the points he made, but the points which were made are actually consistent, unless you can point to a thickness of ice that would actually make sense.
He quotes somebody (Hovind I'm guessing) as saying "Scientists have been trying desperately to find water on other planets. However this search is futile..." and then the video maker cuts the quote off.
Actually, that sounds more like VenomFangX, who is quoted (with video) later on as saying "This planet... just so happens to have 100% of the water in the solar system."
That seems to be pretty explicitly confirming the earlier statement. In fact, he continues: "Here's an interesting thing about water: Where did it all come from? We can't find a speck of H20 in outer space..."
In other words: Watch the entire video. Or, if it makes you happy, look up VenomFangX. This is not quote mining -- he really is quite frequently that moronic.
Maybe it makes me a hypocrite, but I actually couldn't make it through one video. About halfway through, he tries to make the point that we are claiming that the brain is conscious, but that can't be true, because a single atom isn't conscious, and you can't make something conscious out of something that's not.
Um... what? Replace the word "conscious" with anything else. How about: A single atom of paint isn't green, and you can't make something green out of something not green. Sure you can -- just mix blue and yellow paint. I'm sure you can find your own examples.
If you can find an actual example of someone being misquoted, or their position being misrepresented, go ahead. I haven't, so far.
Nobody knows for sure why the universe is slightly unbalanced between matter and antimatter. Some Big Bang models predict an equal amount of matter and antimatter which should have long since annihilated each other.
The rest of that discussion... We are both speaking in generalities. Specific examples might help, but I don't really know enough about this to be sure.
Assuming constancy or linearity in nature is an exercise in foolishness, because few things in nature behave in a linearly predictable manner.
Newton's first law: An object in motion remains in motion, unless acted on by an outside force. That predicts a pretty constant, linear line until you have an outside force to affect it.
What's more, carbon dating is itself based on exponential decay.
Erlang is not pure, true -- it requires programmers to explicitly break tasks up into processes.
It then makes those processes so efficient that there's hardly a performance hit to doing so.
The other advantage is that it's trivial to scale this beyond a single machine, let alone a single core. How easy does Haskell make RPC?
Now, it's true that it helps if your compiler can figure it out for you. The question is whether it's easier to grasp purely functional programming, or the concept of actors. I would argue actors are easier, especially with things like Reia.
People observe matter and energy behave in certain, often puzzling ways and then try to interpret this behavior through the lens of what we think we know. You can clearly observe not what appears to be a black hole or a quasar, but certain highly energetic phenomena taking place in the distant reaches of the cosmos. The black hole or quasars are human interpretations
Certainly. And so is this text.
What you are actually observing are the electrical impulses hitting your brain from your optic nerve, which you interpret to be light, which you interpret to be coming from something called a "computer monitor". On that monitor, you see some shapes which you interpret to be text and windows, which you further interpret to be a representation of the website Slashdot.org.
Many of these could theoretically be wrong -- for instance, you could be plugged into the matrix, and the outside world could be nothing like you imagine. Many have actually been shown to be wrong occasionally -- someone could be performing a man-in-the-middle attack of some sort, and thus the page you are viewing, which you assume originated at slashdot.org, actually originated somewhere else.
But you don't live your life assuming this. Even Hume didn't. You interpret it through the lens of what you know -- or rather, what you've chosen to assume, based on prior observations -- and you can only assume that your memories of those prior observations are legitimate.
The biggest difference between scientific discoveries and personal ones we make as children (Balloons pop! They're loud when they pop!) is that scientific discoveries tend to take a lot more thought to interpret and understand, as they are the product of generations of effort.
Then again, some scientific discoveries, as wild as they might have seemed at the time (Lightning is electricity! Electricity is related to magnetism!) are not only confirmed experimentally, they're relied on in everyday life (chances are, a hard drive was involved somewhere between me typing this, and you reading it.) You'd be ridiculed for claiming that electromagnetism isn't proven -- yet when you start arguing against evolution (which you rely on for much of modern medicine), you get congressmen agreeing with you, or at least saying "maybe".
In essence, the Big Bang theory says: "First there was nothing and then it exploded".
Look it up sometime. There are actually some interesting things that we know about exactly what the Universe looked like several seconds after the big bang -- in fact, I believe we know what it looks like a few fractions of a second after it.
This pretty much requires a multithreaded program, and those are notoriously difficult to write.
Or a multi-process program. Most of what makes threads difficult to write is shared memory, and you really don't need much to be shared to call fsync on a filehandle and then notify the parent process.
All you have to do is make sure that transactions are written to permanent storage in the same order they called commit - that is, if transaction A called commit before transaction B, transaction B will not be logged if the transaction A won't,
That is what I meant by "ordering all transactions" -- order the transactions with respect to each other, no need to order actual writes within them.
However, you'd also want to order them with respect to visibility. You wouldn't want pieces of transaction B to be visible to the process which committed transaction A, and then have transaction B fail.
But for sheer compatibility, can you think of a downside to my suggestion:
What I'm beginning to think might be the best solution is to automatically wrap all FS activity in some ideally-sized transaction, and make sure those transactions are ordered with respect to each other. Can you think of a reason this would perform worse than explicit transactions?
The only downside I can see is that fsync would be equivalent to sync, and thus likely slower.
perfect chance meeting of various bits of 'life goo'
Without actually researching it, this seems plausible. Consider that the oceans -- that is, most of the planet's surface -- were literally teeming with the basic building blocks of life. All it takes is one single-celled organism, no matter how crude, and suddenly, you have tons of life, seemingly out of nowhere.
What created the earth? Ok, what created the universe? Ok, what created the big bang
It is actually quite possible that the big bang had no cause, at least not in our own kind of time. Hawking had a cool model of the Universe as a perfect space-time sphere, meaning time had a beginning and end, at opposite sides of the sphere. He actually disproved it later, but it gives you an idea.
As to how that sphere came into existence? "I don't know" is the acceptable scientific answer; "God did it" is an alright religious theory, just don't be too disappointed if it's disproven. In fact, Hawking did later disprove that whole sphere idea.
Ultimately, I don't really mind the thought of God creating the big bang. At least that is actually compatible with science, even if it's not itself science. If that's what you're teaching your children, at least they'll pay attention in science class, instead of asking stupid questions about Intelligent Design.
Eventually, it boils down to "We're not sure".
Fundamentally, yes. But there's a lot of certainty before that, and these are things we can eventually understand.
For instance, quantum theory gives us a better understanding of subatomic dynamics -- but before that, we pretty much knew an atom is made of protons, neutrons, and electrons, and how many were in each. I'm not even sure we knew about quarks before we used that basic knowledge of the atom to build bombs and nuclear reactors.
That in no way leads to your claim of:
In other words, there's no proof of how it happened.
However, up to a few very tiny fractions of a second before the Big Bang, there's quite a lot of evidence for how it happened.
In light of the history of how science tends to refine its understanding of the universe, it seems somewhat unlikely that we will find the pieces we don't know filled in with "God did it", and it seems ludicrous to think that we would be so entirely wrong as to find out that the world is actually six thousand years old.
Don't be too quick to jump on the "it's proven" bandwagen.
I wasn't. I actually spent most of my life believing that science was so often wrong, that there might be this one thing I was more knowledgeable about than them. It would vary, of course, what I assumed that one thing to be...
So I did my homework, and came to the conclusion that they tend to know what they're talking about. In general, when you have a group of people who have each dedicated their life to thinking critical about a particular problem, and they overwhelmingly arrive at the same conclusion, they're probably right, or at least close to right.
Take gravity. You might say that newton was wrong -- in fact, he was merely not as accurate as he could've been.
And when creationists attack evolution, the statements that don't immediately get them laughed out of the discussion are vagaries like "There's debate about evolution! People disagree on some of the finer details!" Well, yes, they do -- in the hopes that they can refine the theory. There seems little chance anyone will be able to wholly disprove it -- just as Mercury doesn't disprove Newton gravity, it just requires Relativity to refine it a bit.
No, it doesn't kill the argument logically, but it absolutely neuters it. If I may play devil's advocate for a moment:
These computers (PowerPC Macs) are x years old, and ONLY due to using non-open, rights restricted formats, they can't play this, this and this
So what? My computer is new. Shouldn't new computers be able to do more stuff?
if it was open it'd play.
Still not getting me to care.
The computer you get NOW may not play closed video formats in the near future just because some vendor doesn't want to support it any more.
Well, in the future, I'll just buy a new computer and it'll support those formats, duh! I don't expect this computer to last forever.
And in the future, I'll want to watch the game that's new then. If I need this game, it'll be on YouTube.
Support open formats, and you can use the current computer indefinitely.
That sounds nice, but if by "support open formats" you mean "don't use closed ones", I'd rather watch my sports than not.
I'd also point out the numerous rights restrictions schemes recently that have shut down, leaving people completely unable to use the (mainly music) they paid for
That's why you go for the subscription services. If one closes down, you just switch to another, and re-download all your stuff.
Finally, I comment "Do you want to be able to move music on and off whatever video or music player you'd like? Closed formats try to prevent this, open formats do not."
Answer: I don't really care, that's complicated and hard to do anyway.
Eugh. Now, speaking as myself...
I do avoid all DRM in media, with two exceptions: Other people's systems (if I'm at someone's house, I won't insist it's a Mythbox before I use their DVR), and several video game DRM schemes (Steam actually makes a really good trade -- you have to be on the network all the time, but the network actually does benefit you, and it's not like you were going to play an open source game instead.)
Oh, and I suppose you could count Windows itself, though I only boot it for games.
But it's really hard to make that sale to people who haven't already been burned by DRM. Even once you do, it's even harder to convince people that this is something they should be enraged about, rather than something they should be annoyed about.
Again: They see it as a feature that wasn't added, rather than a feature that was killed. Remember, for us, the ability to transfer large files between computers is as "easy" as scp, ftp, etc. For them -- once they understand that email can't handle it -- a brand-new service like yousendit, or copying everything to an external hard drive, is a brilliant innovation.
The most important point to communicate, and the most difficult, is just how cool our lives get when you take away DRM. I try to demonstrate that as effectively as I can, by occasionally saying things like "Oh, that movie isn't on my laptop, but it's on the server upstairs. Ten minutes and it'll be transfered down over the wireless."
And even then, it's still on the basis of individual features, not "freedom", that Joe understands this. Meaning that as soon as Microsoft adds enough DRM that you can do that (but only between Windows Media compliant machines), it deflates the whole argument.
The creationists and evolutionists BELIEVE a different set of witnesses.
The funny thing is, the creationsts' "witnesses" were a bunch of humans who wrote a book a few hundred years ago.
The evolutionists' "witnesses" are extrapolation from known scientific principles.
Put another way: Is it possible to know experimentally or observationally that a black hole exists? You can't see it, you can only see the way it affects the space around it. And you clearly can't experiment on it. Yet most of us agree that they exist.
I suppose it depends what you mean by "observation", then, right? I can clearly observe what appears to be a black hole, or a quasar, or a supernova. I can also observe what appears to be an accurate carbon-dating. In both cases, I'm looking at some particles being detected well after the fact -- and I'm not even looking at those directly, I'm looking at what my instruments tell me they are.
It's also worth mentioning: Evolution actually does conform to basic laws of physics, at least as far back as the Big Bang, and we're starting to understand that, too.
Creation really doesn't, unless we assume that the universe suddenly popped into being with everything set up just so, just to tempt the faith of scientists in the future by making it appear that there was evolution, and that the universe is billions of years old. But we actually have no evidence except some really questionable testimony that there is a being capable of doing that -- whereas we have all the evidence in the world (so to speak) that evolution did happen.
I would suggest that a lot of this has to do with people in the "bible belt" taking your attitude, and keeping quiet.
Here, a quick Google finds this study -- even if you include "religious unaffiliated" as a religion, that still leaves 10.3% of the population either explicitly atheist/agnostic, or without religion.
For comparison: Even in the Bible Belt, people are starting to accept homosexuality. They might not like it, and they might not like gay marriage, but they're willing to at least accept that it doesn't make a person inherently wrong.
And according to another quick Google, and another website that is clearly biased towards "traditional values", only 2.3% are homosexual.
You can run the numbers yourself, but either way, 10.3% is kind of a huge number. That's some 31 million people in this country.
I'm not saying that we should be mocking religion, but I do think it might be a good idea for atheists to "come out of the closet", so to speak. We're more hated than homosexuals, and we're a larger group. Really, the only reason the "bible belt" can remain so willfully ignorant is because they haven't been exposed to it.
This is exactly as informed and insightful an understanding of evolution as Ben Stein's description of "lightning striking a puddle of mud."
In other words, it's not actually about evolution, and it's also an incredibly poor understanding of abiogenesis.
Either way, there's no solid proof.
However, there is quite a bit more evidence to support abiogenesis, and a truly massive amount of evidence for evolution. Modern biology relies on evolution, in fact.
On the other hand, there is absolutely zero proof of the Bible's Genesis.
Oh, and for that matter:
if you agree with them and don't argue, you have a degree.
Even in high school, teachers rewarded me for asking questions, even if it led to a debate, so long as I was thinking.
Ultimately, no one really cares what irrational beliefs you hold -- the vast majority of scientists are religious. The important point is to understand the difference between an unfounded belief and actual science.
Obviously not all avid gamers are maladjusted troglodytes. But I've known plenty who are.
Given how much of the, I don't know, 13-30 male population plays video games, I'd suggest that any attempt to stereotype them is going to fail.
And that's just the male population.
As we all know, correlation is not causation. But you haven't really even shown correlation, just wild speculation based on a stereotype and (maybe) a couple of anecdotes.
I've known plenty of people who have disappeared into their favorite games--maybe even becoming part of a different community on-line--and allowing all of their IRL social contacts suffer.
Here's what I've found about myself: When I am avoiding doing something, or avoiding my life in general, I am entirely capable of falling into games, TV, books, movies, even Project Euler. If the games weren't there, I'd find something else.
So the correct approach to this isn't to blame the games, it's to confront the real problem of whatever it is I'm avoiding.
But that's just me. I can't really speak for others because, as I said, it's such a huge demographic.
I do call them Americans, except when I need to make a point that relies on race. If you replace "African Americans" with "Americans" in my post, does it make as much sense?
The problem is that there is no way which will work everywhere, except fsync - and guess what fsync does to performance?
That is true.
POSIX API needs to be expanded to include support for transactions. Until it's done, expect people to rely on hacks, because they have nothing else to rely on.
Except they do -- fsync, as above, since they have no way of detecting whether the filesystem behaves properly.
But the application doesn't need that guarantee. It simply needs the guarantee that either the file is updated or left alone.
That is true. But there are also going to be dependencies, and this is where I'm not really sure what to do.
That is: The application might update file a, and then update file b only if file a succeeded. Or a different application might update file b based on information read from file a. This means the OS has to either order all transactions, or somehow detect which transactions depend on others.
What I'm beginning to think might be the best solution is to automatically wrap all FS activity in some ideally-sized transaction, and make sure those transactions are ordered with respect to each other. Can you think of a reason this would perform worse than explicit transactions?
Well, a company named Real Networks ships a fully supported Linux player and you see the feedback they get.
Mostly because RealPlayer absolutely sucks, on any platform. It's cool that they ship a Linux player, but it's still not something I want.
Kind of like how MySpace probably works well on Firefox. Great, I can use it on Linux, but I still really don't like it.
H264 and MPEG4 SP and even simple http streaming took over, it is only MS not admitting it. Even Real moved to AAC (plus) and some variant of h264 in rv10.
Firstly: It's actually between these and WMV.
Second: While h.264 and aac support is generally better than the latest wmv (or vc-1) and the latest wma codecs, keep in mind that all of these are exactly as proprietary, and exactly as patent-laden.
So, if you really want to be open, put theora and vorbis in either ogm or matroska.
Finally: All of this is really quite irrelevant. It's not the codec, it's the encapsulation. In this case, they are likely wrapping vanilla vc-1 and wma9 (or whatever) in a wmv container, and encrypting that. While I agree that it's the 90's again, they aren't even trying to hide that they have the content, and you need to install the software to get the content. They're not even trying to sell it to the consumer as being cool or innovative -- as far as I can tell, they try to sell Silverlight to developers, and then try to sell the finished Silverlight product to users.
Which still kills the argument for Average Joe -- if he's switching to Mac, he's probably switching to a new Intel Mac, not a hand-me-down PPC. Silverlight will work fine, and the anti-DRM speech will get a "what's your point?"
Because ultimately, if the argument is that you can't watch it on the PS3 (running Linux) on your TV, Average Joe will see that as a feature that "they" didn't add, whereas we see it as a basic right they took away.
Anyone who tries to tell you that all nations and races are equal are doing so for their own advantage.
Really? I'd think that if we hadn't given equal rights to African Americans, we might have an "advantage" in the sense you're describing. After all, slaves could be useful.
No, the main advantage is a moral one -- that we aren't complete barbarians.
A world without hate is an imprisoned world that I sure wouldn't want to live in.
The right time being the hundredths of a second between the commit of the file data and the commit of the directory data, not 60 seconds.
And if you fsync'd, the right time would be zero, on either ext3 or ext4. Or XFS, for that matter.
If I fsync after every write, I can get reliability in ext2.
No you can't. Reliability in ext2 would force you to sync not just your file, but whole directory structures -- and even then, you'd only be safe until something else starts writing.
I put up with the performance hit from ext3 and ext4 because I want the reliability in the filesystem instead of having to build it into every part of every application.
Too late.
All the journaling guarantees is that if you lose power, you won't have to fsck -- you'll get a filesystem which is internally consistent. Oh, and it also guarantees that you won't see circular directory entries, or an entire directory falling off the face of the planet, and other nastiness.
Whether it's consistent with respect to your application is completely outside the scope of the FS journaling, and is the responsibility of your application. Put it in a library, use a database, whatever -- but it's not the filesystem's fault that you failed to read the spec, nor is it very smart of you to code to ext3 instead of POSIX.
Application developers reasonably expect that writes to the disk which happen far apart in time will happen in order.
First: Where is this in the spec? At all?
Second: It's not "far apart in time". It's within a few fractions of a second.
It seems dead obvious, at least to me, that the update of the directory entry should be deferred until after ext4 flushes that part of the file written prior to the change in the directory entry.
Most uses of the filesystem, this really doesn't matter.
Let me put it this way -- why are you assuming it's file write, then a directory entry change? That makes sense in this case, but suppose I do this:
cd / tar -xjpf foo.tar.bz2 #creates foo/ touch/var/lock/whatever # (and aquire the lock, somehow) ln foo/etc/foo/etc ln foo/usr/bin/*/usr/bin... # or better, crawl it deliberately rm -rf foo rm/var/lock/whatever
Now, since most of those are directory operations, not actual file writes, suppose those hit disk out of order. That means I might have foo/usr/bin/* installed, but not foo/etc/foo. But the 'rm' might've claimed etc/foo, so I can't exactly roll the transaction forward by crawling foo, I have to start over.
Never mind the fact that while I'm doing this, the rest of the filesystem is in an inconsistent state, with respect to my contrived package manager.
So what this means is, with the way the POSIX spec is now, if you want to rely on stuff like that, you have to assume every directory entry hits disk in order -- or you have to flush to disk yourself at a few critical points.
Now, if you can call 'sync' immediately after installing everything, but before blowing away the installation folder ('foo'), I believe the above scheme actually works. In fact, it's more efficient that way, because the filesystem gets to reorder the directory structure before flushing. It's not as efficient as it could be, but it's better.
The issue that FS authors, well any authors of any system programs/tools/etc need to understand is that your tool being usable is the #1 important thing.
Part of usability is performance. This is a significant performance improvement.
So, if you do something that really screws that over, well then you probably did it wrong. Doesn't matter if you fully documented it, doesn't matter if it technically "follows the spec" what matters is that it isn't usable.
The real problem here is that application developers were relying on a hack that happened to work on ext3, but not everywhere else.
Let me ask you this -- should XFS change the way it behaves because of this? EXT4 is doing exactly what XFS has done for decades.
I mean I could write a spec for a file system that says "No write is guaranteed to be written to disk until the OS is shut down, everything can be cached in RAM for an indefinite amount of time." However that'd be real flaky and lead to data loss.
No, that's actually precisely what the spec says, with one exception: You can guarantee it to be written to disk by calling fsync.
I'd give these guys more credit if I was aware of any other major OS/FS combo that did shit like this, but I'm not.
Only because you haven't looked.
In fact, there's a mount option to turn this behavior on in ext3.
Slashdot Mirror
I was looking for somewhere to attach this comment, and you're it.
"Cloud" is the modern term for a mainframe, time-sharing-like model.
One advantage is that your data lives on a server somewhere, meaning someone else is responsible for backup, and you can access it from any "terminal" (typically a web browser, but could also be things like the Steam client).
Another advantage is a potential pricing model for developers -- Amazon EC2 charges per hour of server time used, at a very flat rate. If you only use an hour, you only pay ten cents.
The big advantage of an infrastructure like EC2 is shown in pathological cases, like websites which tend to receive more traffic at certain times of the day. So every night, you can shut down whatever capacity you don't need, and stop paying for it -- and Amazon can then allocate it to someone else who needs it at that time, possibly overnight.
At a different level, you see the same pattern with web applications -- you don't need a computer more powerful than it needs to be to run Firefox. The server can do whatever computing you need that isn't already happening locally -- but most GUI apps spend a lot of time waiting for the user. So when you do a search in Gmail, that takes some server CPU -- but while you're examining the results, that server is off running someone else's search.
Here's the problem: None of these advantages apply to these guys.
The "my data is elsewhere" advantage is irrelevant. Steam already provides this. So long as I remember a username and password, I can download all my Steam games, along with all their savegames and settings.
The idea that a piece of hardware might not fully be utilized by a single user, and could thus be re-allocated, is similarly irrelevant here. Unless they have some sort of weird economies of scale where one video card can serve a thousand users, and cost less than a thousand times the cost of one normal video card... they're pretty much stuck with one machine per gamer.
And since these machines will have to be geographically close to the gamer to be at all viable, there's going to be very little gain from half the gamers going to sleep just as the other half wakes up. You're still going to get the bulk of your traffic from large groups of gamers coming home and logging in at about the same time.
The only advantage is the not-having-to-think-about-maintenance bit, which is pretty weak against consoles. A console is something even John Q. Gamer can unpack and plug in himself. Having to do it every four years is really not that big a deal.
So that's a very long way of saying: I agree with you, "cloud" is being abused. This is clearly someone trying to cash in on the buzzword, without really understanding what it's good for -- it would be like creating an XML representation of a waveform from a sound file.
This doesn't mean XML is worthless, or that it lacks meaning. It just means that someone drank a little too much kool-aid.
Similarly, "cloud computing", as vague as it can be, is really about a couple of related concepts that are concrete enough to write down. This is just something that it's really not suited for.
However, even a fiber optics line I'd have my doubts.
Doing some quick calculations:
The highest number I've gotten for Blu-Ray maximum bandwidth is 54 megabits per second. I've seen torrents much smaller that still looked good.
Assuming uncapped, that's actually doable. Fiber is typically 100 mbits per second, and I'm sure some places offer gigabit.
However, encoding time is on the order of hours or days, and is certainly not live. So the real problem is latency -- take 50 ms from your LCD monitor, plus whatever a wireless controller ads, plus the latency between you and their servers, plus the lag for them to render, capture, and encode, then decode back at the client... that's easily getting up to 200 ms, which I'd consider unplayable.
Also, unless the $50/year includes games, it makes little economic sense, either. These systems are designed to last some four years or so. A Wii can be had for $160, according to a quick Google; this would be $200. A Wii can work when your Internet is down, or when your internet is not fiber. And a Wii actually has games already -- not as many as its competitors, but some.
Where I could see this working is in a LAN environment
Not really. LANs are typically 100 mbits, or if you're willing to spend money on a good switch, gigabit. Same situation as fiber.
The only advantage of a LAN is, with a good switch, you aren't using everyone else's bandwidth, but if you're proposing this:
make some kind of "xbox360server" to host all the games as basically virtual machines across a lan,
That's still likely to be a single port, which means now everyone on the LAN is limited to a combined 100 mbits for their video. It means the concept of a LAN party just got very, very impractical.
And WTF would be the point, if it's a console anyway? In what way is that "xbox360server" better than a real Xbox 360?
As for their "no piracy" claim, as a consumer, that doesn't make me want to sign up for the service. That makes me want to go far away, into the open arms of indie developers, who typically ship with reduced or no DRM.
For example, in part 5, while trying to refute Hovind's ice shield theory, he uses the diagram to estimate how big Hovind's ice shield around the earth is and continues with this size for the rest of the video. But previously in the video Hovind had said he didn't know how big it was but guessed at "10 or 20 or 30 inches thick."
The problem is that Hovind was not consistent. 10 or 20 or 30 inches would've cracked apart due to barometric pressure changes. The only other numbers we have to work with are the diagram, which shows much more water than there's really room for on the surface of the planet.
Follow the calculation, and the diagram is actually showing less ice than you would need for the shield to last as long as Hovind claims it did (2000 years). So, you cannot put enough ice there to survive that long, without also putting enough ice there to have the planet flooded under about a thousand kilometers of water.
Ok, that gives you the flood, but after the first kilometer of ice, the planet would be completely dark before that flood. It also doesn't explain where all that water went -- how did the flood end?
That doesn't cover all the points he made, but the points which were made are actually consistent, unless you can point to a thickness of ice that would actually make sense.
He quotes somebody (Hovind I'm guessing) as saying "Scientists have been trying desperately to find water on other planets. However this search is futile..." and then the video maker cuts the quote off.
Actually, that sounds more like VenomFangX, who is quoted (with video) later on as saying "This planet... just so happens to have 100% of the water in the solar system."
That seems to be pretty explicitly confirming the earlier statement. In fact, he continues: "Here's an interesting thing about water: Where did it all come from? We can't find a speck of H20 in outer space..."
In other words: Watch the entire video. Or, if it makes you happy, look up VenomFangX. This is not quote mining -- he really is quite frequently that moronic.
Maybe it makes me a hypocrite, but I actually couldn't make it through one video. About halfway through, he tries to make the point that we are claiming that the brain is conscious, but that can't be true, because a single atom isn't conscious, and you can't make something conscious out of something that's not.
Um... what? Replace the word "conscious" with anything else. How about: A single atom of paint isn't green, and you can't make something green out of something not green. Sure you can -- just mix blue and yellow paint. I'm sure you can find your own examples.
If you can find an actual example of someone being misquoted, or their position being misrepresented, go ahead. I haven't, so far.
You can use actors with purely functional programming.
Certainly.
But my understanding was that Haskell parallizes by having the compiler analyze your program. I assume actors are not part of the language itself.
Nobody knows for sure why the universe is slightly unbalanced between matter and antimatter. Some Big Bang models predict an equal amount of matter and antimatter which should have long since annihilated each other.
The rest of that discussion... We are both speaking in generalities. Specific examples might help, but I don't really know enough about this to be sure.
Assuming constancy or linearity in nature is an exercise in foolishness, because few things in nature behave in a linearly predictable manner.
Newton's first law: An object in motion remains in motion, unless acted on by an outside force. That predicts a pretty constant, linear line until you have an outside force to affect it.
What's more, carbon dating is itself based on exponential decay.
Erlang is not pure, true -- it requires programmers to explicitly break tasks up into processes.
It then makes those processes so efficient that there's hardly a performance hit to doing so.
The other advantage is that it's trivial to scale this beyond a single machine, let alone a single core. How easy does Haskell make RPC?
Now, it's true that it helps if your compiler can figure it out for you. The question is whether it's easier to grasp purely functional programming, or the concept of actors. I would argue actors are easier, especially with things like Reia.
People observe matter and energy behave in certain, often puzzling ways and then try to interpret this behavior through the lens of what we think we know. You can clearly observe not what appears to be a black hole or a quasar, but certain highly energetic phenomena taking place in the distant reaches of the cosmos. The black hole or quasars are human interpretations
Certainly. And so is this text.
What you are actually observing are the electrical impulses hitting your brain from your optic nerve, which you interpret to be light, which you interpret to be coming from something called a "computer monitor". On that monitor, you see some shapes which you interpret to be text and windows, which you further interpret to be a representation of the website Slashdot.org.
Many of these could theoretically be wrong -- for instance, you could be plugged into the matrix, and the outside world could be nothing like you imagine. Many have actually been shown to be wrong occasionally -- someone could be performing a man-in-the-middle attack of some sort, and thus the page you are viewing, which you assume originated at slashdot.org, actually originated somewhere else.
But you don't live your life assuming this. Even Hume didn't. You interpret it through the lens of what you know -- or rather, what you've chosen to assume, based on prior observations -- and you can only assume that your memories of those prior observations are legitimate.
The biggest difference between scientific discoveries and personal ones we make as children (Balloons pop! They're loud when they pop!) is that scientific discoveries tend to take a lot more thought to interpret and understand, as they are the product of generations of effort.
Then again, some scientific discoveries, as wild as they might have seemed at the time (Lightning is electricity! Electricity is related to magnetism!) are not only confirmed experimentally, they're relied on in everyday life (chances are, a hard drive was involved somewhere between me typing this, and you reading it.) You'd be ridiculed for claiming that electromagnetism isn't proven -- yet when you start arguing against evolution (which you rely on for much of modern medicine), you get congressmen agreeing with you, or at least saying "maybe".
In essence, the Big Bang theory says: "First there was nothing and then it exploded".
Look it up sometime. There are actually some interesting things that we know about exactly what the Universe looked like several seconds after the big bang -- in fact, I believe we know what it looks like a few fractions of a second after it.
This pretty much requires a multithreaded program, and those are notoriously difficult to write.
Or a multi-process program. Most of what makes threads difficult to write is shared memory, and you really don't need much to be shared to call fsync on a filehandle and then notify the parent process.
All you have to do is make sure that transactions are written to permanent storage in the same order they called commit - that is, if transaction A called commit before transaction B, transaction B will not be logged if the transaction A won't,
That is what I meant by "ordering all transactions" -- order the transactions with respect to each other, no need to order actual writes within them.
However, you'd also want to order them with respect to visibility. You wouldn't want pieces of transaction B to be visible to the process which committed transaction A, and then have transaction B fail.
But for sheer compatibility, can you think of a downside to my suggestion:
What I'm beginning to think might be the best solution is to automatically wrap all FS activity in some ideally-sized transaction, and make sure those transactions are ordered with respect to each other. Can you think of a reason this would perform worse than explicit transactions?
The only downside I can see is that fsync would be equivalent to sync, and thus likely slower.
perfect chance meeting of various bits of 'life goo'
Without actually researching it, this seems plausible. Consider that the oceans -- that is, most of the planet's surface -- were literally teeming with the basic building blocks of life. All it takes is one single-celled organism, no matter how crude, and suddenly, you have tons of life, seemingly out of nowhere.
Describing this as a "puddle of slime" is kind of like claiming that a single drop of water can cover the world.
What created the earth? Ok, what created the universe? Ok, what created the big bang
It is actually quite possible that the big bang had no cause, at least not in our own kind of time. Hawking had a cool model of the Universe as a perfect space-time sphere, meaning time had a beginning and end, at opposite sides of the sphere. He actually disproved it later, but it gives you an idea.
As to how that sphere came into existence? "I don't know" is the acceptable scientific answer; "God did it" is an alright religious theory, just don't be too disappointed if it's disproven. In fact, Hawking did later disprove that whole sphere idea.
Ultimately, I don't really mind the thought of God creating the big bang. At least that is actually compatible with science, even if it's not itself science. If that's what you're teaching your children, at least they'll pay attention in science class, instead of asking stupid questions about Intelligent Design.
Eventually, it boils down to "We're not sure".
Fundamentally, yes. But there's a lot of certainty before that, and these are things we can eventually understand.
For instance, quantum theory gives us a better understanding of subatomic dynamics -- but before that, we pretty much knew an atom is made of protons, neutrons, and electrons, and how many were in each. I'm not even sure we knew about quarks before we used that basic knowledge of the atom to build bombs and nuclear reactors.
That in no way leads to your claim of:
In other words, there's no proof of how it happened.
However, up to a few very tiny fractions of a second before the Big Bang, there's quite a lot of evidence for how it happened.
In light of the history of how science tends to refine its understanding of the universe, it seems somewhat unlikely that we will find the pieces we don't know filled in with "God did it", and it seems ludicrous to think that we would be so entirely wrong as to find out that the world is actually six thousand years old.
Don't be too quick to jump on the "it's proven" bandwagen.
I wasn't. I actually spent most of my life believing that science was so often wrong, that there might be this one thing I was more knowledgeable about than them. It would vary, of course, what I assumed that one thing to be...
So I did my homework, and came to the conclusion that they tend to know what they're talking about. In general, when you have a group of people who have each dedicated their life to thinking critical about a particular problem, and they overwhelmingly arrive at the same conclusion, they're probably right, or at least close to right.
Take gravity. You might say that newton was wrong -- in fact, he was merely not as accurate as he could've been.
And when creationists attack evolution, the statements that don't immediately get them laughed out of the discussion are vagaries like "There's debate about evolution! People disagree on some of the finer details!" Well, yes, they do -- in the hopes that they can refine the theory. There seems little chance anyone will be able to wholly disprove it -- just as Mercury doesn't disprove Newton gravity, it just requires Relativity to refine it a bit.
It's also worth mentioning that we're talking abo
No, it doesn't kill the argument logically, but it absolutely neuters it. If I may play devil's advocate for a moment:
These computers (PowerPC Macs) are x years old, and ONLY due to using non-open, rights restricted formats, they can't play this, this and this
So what? My computer is new. Shouldn't new computers be able to do more stuff?
if it was open it'd play.
Still not getting me to care.
The computer you get NOW may not play closed video formats in the near future just because some vendor doesn't want to support it any more.
Well, in the future, I'll just buy a new computer and it'll support those formats, duh! I don't expect this computer to last forever.
And in the future, I'll want to watch the game that's new then. If I need this game, it'll be on YouTube.
Support open formats, and you can use the current computer indefinitely.
That sounds nice, but if by "support open formats" you mean "don't use closed ones", I'd rather watch my sports than not.
I'd also point out the numerous rights restrictions schemes recently that have shut down, leaving people completely unable to use the (mainly music) they paid for
That's why you go for the subscription services. If one closes down, you just switch to another, and re-download all your stuff.
Finally, I comment "Do you want to be able to move music on and off whatever video or music player you'd like? Closed formats try to prevent this, open formats do not."
Answer: I don't really care, that's complicated and hard to do anyway.
Eugh. Now, speaking as myself...
I do avoid all DRM in media, with two exceptions: Other people's systems (if I'm at someone's house, I won't insist it's a Mythbox before I use their DVR), and several video game DRM schemes (Steam actually makes a really good trade -- you have to be on the network all the time, but the network actually does benefit you, and it's not like you were going to play an open source game instead.)
Oh, and I suppose you could count Windows itself, though I only boot it for games.
But it's really hard to make that sale to people who haven't already been burned by DRM. Even once you do, it's even harder to convince people that this is something they should be enraged about, rather than something they should be annoyed about.
Again: They see it as a feature that wasn't added, rather than a feature that was killed. Remember, for us, the ability to transfer large files between computers is as "easy" as scp, ftp, etc. For them -- once they understand that email can't handle it -- a brand-new service like yousendit, or copying everything to an external hard drive, is a brilliant innovation.
The most important point to communicate, and the most difficult, is just how cool our lives get when you take away DRM. I try to demonstrate that as effectively as I can, by occasionally saying things like "Oh, that movie isn't on my laptop, but it's on the server upstairs. Ten minutes and it'll be transfered down over the wireless."
And even then, it's still on the basis of individual features, not "freedom", that Joe understands this. Meaning that as soon as Microsoft adds enough DRM that you can do that (but only between Windows Media compliant machines), it deflates the whole argument.
I'd suggest that you'd make a substantial contribution to the so-called "scientific" justifications that creationists occasionally make.
Of course, you'd have to be prepared for your "discovery" to be torn to shreds by people who actually understand science.
The creationists and evolutionists BELIEVE a different set of witnesses.
The funny thing is, the creationsts' "witnesses" were a bunch of humans who wrote a book a few hundred years ago.
The evolutionists' "witnesses" are extrapolation from known scientific principles.
Put another way: Is it possible to know experimentally or observationally that a black hole exists? You can't see it, you can only see the way it affects the space around it. And you clearly can't experiment on it. Yet most of us agree that they exist.
I suppose it depends what you mean by "observation", then, right? I can clearly observe what appears to be a black hole, or a quasar, or a supernova. I can also observe what appears to be an accurate carbon-dating. In both cases, I'm looking at some particles being detected well after the fact -- and I'm not even looking at those directly, I'm looking at what my instruments tell me they are.
It's also worth mentioning: Evolution actually does conform to basic laws of physics, at least as far back as the Big Bang, and we're starting to understand that, too.
Creation really doesn't, unless we assume that the universe suddenly popped into being with everything set up just so, just to tempt the faith of scientists in the future by making it appear that there was evolution, and that the universe is billions of years old. But we actually have no evidence except some really questionable testimony that there is a being capable of doing that -- whereas we have all the evidence in the world (so to speak) that evolution did happen.
I would suggest that a lot of this has to do with people in the "bible belt" taking your attitude, and keeping quiet.
Here, a quick Google finds this study -- even if you include "religious unaffiliated" as a religion, that still leaves 10.3% of the population either explicitly atheist/agnostic, or without religion.
For comparison: Even in the Bible Belt, people are starting to accept homosexuality. They might not like it, and they might not like gay marriage, but they're willing to at least accept that it doesn't make a person inherently wrong.
And according to another quick Google, and another website that is clearly biased towards "traditional values", only 2.3% are homosexual.
You can run the numbers yourself, but either way, 10.3% is kind of a huge number. That's some 31 million people in this country.
I'm not saying that we should be mocking religion, but I do think it might be a good idea for atheists to "come out of the closet", so to speak. We're more hated than homosexuals, and we're a larger group. Really, the only reason the "bible belt" can remain so willfully ignorant is because they haven't been exposed to it.
or that the world evolved from a puddle of slime.
This is exactly as informed and insightful an understanding of evolution as Ben Stein's description of "lightning striking a puddle of mud."
In other words, it's not actually about evolution, and it's also an incredibly poor understanding of abiogenesis.
Either way, there's no solid proof.
However, there is quite a bit more evidence to support abiogenesis, and a truly massive amount of evidence for evolution. Modern biology relies on evolution, in fact.
On the other hand, there is absolutely zero proof of the Bible's Genesis.
Oh, and for that matter:
if you agree with them and don't argue, you have a degree.
Even in high school, teachers rewarded me for asking questions, even if it led to a debate, so long as I was thinking.
Ultimately, no one really cares what irrational beliefs you hold -- the vast majority of scientists are religious. The important point is to understand the difference between an unfounded belief and actual science.
Obviously not all avid gamers are maladjusted troglodytes. But I've known plenty who are.
Given how much of the, I don't know, 13-30 male population plays video games, I'd suggest that any attempt to stereotype them is going to fail.
And that's just the male population.
As we all know, correlation is not causation. But you haven't really even shown correlation, just wild speculation based on a stereotype and (maybe) a couple of anecdotes.
I've known plenty of people who have disappeared into their favorite games--maybe even becoming part of a different community on-line--and allowing all of their IRL social contacts suffer.
Here's what I've found about myself: When I am avoiding doing something, or avoiding my life in general, I am entirely capable of falling into games, TV, books, movies, even Project Euler. If the games weren't there, I'd find something else.
So the correct approach to this isn't to blame the games, it's to confront the real problem of whatever it is I'm avoiding.
But that's just me. I can't really speak for others because, as I said, it's such a huge demographic.
if a site such as isoHunt allows people to find a pirated copy of movies such as Watchmen or The Dark Knight, is it breaching Canadian copyright law?
I don't get it.
Are they trying to subtly make a point that only certain movies should be protected?
Or do they really feel that the general public doesn't know what a "movie" is, and could use some examples?
Maybe it's a nitpick, but something about that language just seems gratuitous, yet most news media seems to do just that.
I do call them Americans, except when I need to make a point that relies on race. If you replace "African Americans" with "Americans" in my post, does it make as much sense?
The problem is that there is no way which will work everywhere, except fsync - and guess what fsync does to performance?
That is true.
POSIX API needs to be expanded to include support for transactions. Until it's done, expect people to rely on hacks, because they have nothing else to rely on.
Except they do -- fsync, as above, since they have no way of detecting whether the filesystem behaves properly.
But the application doesn't need that guarantee. It simply needs the guarantee that either the file is updated or left alone.
That is true. But there are also going to be dependencies, and this is where I'm not really sure what to do.
That is: The application might update file a, and then update file b only if file a succeeded. Or a different application might update file b based on information read from file a. This means the OS has to either order all transactions, or somehow detect which transactions depend on others.
What I'm beginning to think might be the best solution is to automatically wrap all FS activity in some ideally-sized transaction, and make sure those transactions are ordered with respect to each other. Can you think of a reason this would perform worse than explicit transactions?
Well, a company named Real Networks ships a fully supported Linux player and you see the feedback they get.
Mostly because RealPlayer absolutely sucks, on any platform. It's cool that they ship a Linux player, but it's still not something I want.
Kind of like how MySpace probably works well on Firefox. Great, I can use it on Linux, but I still really don't like it.
H264 and MPEG4 SP and even simple http streaming took over, it is only MS not admitting it. Even Real moved to AAC (plus) and some variant of h264 in rv10.
Firstly: It's actually between these and WMV.
Second: While h.264 and aac support is generally better than the latest wmv (or vc-1) and the latest wma codecs, keep in mind that all of these are exactly as proprietary, and exactly as patent-laden.
So, if you really want to be open, put theora and vorbis in either ogm or matroska.
Finally: All of this is really quite irrelevant. It's not the codec, it's the encapsulation. In this case, they are likely wrapping vanilla vc-1 and wma9 (or whatever) in a wmv container, and encrypting that. While I agree that it's the 90's again, they aren't even trying to hide that they have the content, and you need to install the software to get the content. They're not even trying to sell it to the consumer as being cool or innovative -- as far as I can tell, they try to sell Silverlight to developers, and then try to sell the finished Silverlight product to users.
Which still kills the argument for Average Joe -- if he's switching to Mac, he's probably switching to a new Intel Mac, not a hand-me-down PPC. Silverlight will work fine, and the anti-DRM speech will get a "what's your point?"
Because ultimately, if the argument is that you can't watch it on the PS3 (running Linux) on your TV, Average Joe will see that as a feature that "they" didn't add, whereas we see it as a basic right they took away.
Anyone who tries to tell you that all nations and races are equal are doing so for their own advantage.
Really? I'd think that if we hadn't given equal rights to African Americans, we might have an "advantage" in the sense you're describing. After all, slaves could be useful.
No, the main advantage is a moral one -- that we aren't complete barbarians.
A world without hate is an imprisoned world that I sure wouldn't want to live in.
How so?
The right time being the hundredths of a second between the commit of the file data and the commit of the directory data, not 60 seconds.
And if you fsync'd, the right time would be zero, on either ext3 or ext4. Or XFS, for that matter.
If I fsync after every write, I can get reliability in ext2.
No you can't. Reliability in ext2 would force you to sync not just your file, but whole directory structures -- and even then, you'd only be safe until something else starts writing.
I put up with the performance hit from ext3 and ext4 because I want the reliability in the filesystem instead of having to build it into every part of every application.
Too late.
All the journaling guarantees is that if you lose power, you won't have to fsck -- you'll get a filesystem which is internally consistent. Oh, and it also guarantees that you won't see circular directory entries, or an entire directory falling off the face of the planet, and other nastiness.
Whether it's consistent with respect to your application is completely outside the scope of the FS journaling, and is the responsibility of your application. Put it in a library, use a database, whatever -- but it's not the filesystem's fault that you failed to read the spec, nor is it very smart of you to code to ext3 instead of POSIX.
Application developers reasonably expect that writes to the disk which happen far apart in time will happen in order.
First: Where is this in the spec? At all?
Second: It's not "far apart in time". It's within a few fractions of a second.
It seems dead obvious, at least to me, that the update of the directory entry should be deferred until after ext4 flushes that part of the file written prior to the change in the directory entry.
Most uses of the filesystem, this really doesn't matter.
Let me put it this way -- why are you assuming it's file write, then a directory entry change? That makes sense in this case, but suppose I do this:
cd / /var/lock/whatever # (and aquire the lock, somehow) /etc /usr/bin ... # or better, crawl it deliberately /var/lock/whatever
tar -xjpf foo.tar.bz2 #creates foo/
touch
ln foo/etc/foo
ln foo/usr/bin/*
rm -rf foo
rm
Now, since most of those are directory operations, not actual file writes, suppose those hit disk out of order. That means I might have foo/usr/bin/* installed, but not foo/etc/foo. But the 'rm' might've claimed etc/foo, so I can't exactly roll the transaction forward by crawling foo, I have to start over.
Never mind the fact that while I'm doing this, the rest of the filesystem is in an inconsistent state, with respect to my contrived package manager.
So what this means is, with the way the POSIX spec is now, if you want to rely on stuff like that, you have to assume every directory entry hits disk in order -- or you have to flush to disk yourself at a few critical points.
Now, if you can call 'sync' immediately after installing everything, but before blowing away the installation folder ('foo'), I believe the above scheme actually works. In fact, it's more efficient that way, because the filesystem gets to reorder the directory structure before flushing. It's not as efficient as it could be, but it's better.
The issue that FS authors, well any authors of any system programs/tools/etc need to understand is that your tool being usable is the #1 important thing.
Part of usability is performance. This is a significant performance improvement.
So, if you do something that really screws that over, well then you probably did it wrong. Doesn't matter if you fully documented it, doesn't matter if it technically "follows the spec" what matters is that it isn't usable.
The real problem here is that application developers were relying on a hack that happened to work on ext3, but not everywhere else.
Let me ask you this -- should XFS change the way it behaves because of this? EXT4 is doing exactly what XFS has done for decades.
I mean I could write a spec for a file system that says "No write is guaranteed to be written to disk until the OS is shut down, everything can be cached in RAM for an indefinite amount of time." However that'd be real flaky and lead to data loss.
No, that's actually precisely what the spec says, with one exception: You can guarantee it to be written to disk by calling fsync.
I'd give these guys more credit if I was aware of any other major OS/FS combo that did shit like this, but I'm not.
Only because you haven't looked.
In fact, there's a mount option to turn this behavior on in ext3.
The "bad design" goes deeper than that.
XFS does the exact same thing, for what it's worth.