It's not that important to me. But for once, I'd love to see fairly exhaustive benchmarks like I used to see for Reiser3. At least, at the time, they felt pretty good.
I imagine Ext4 is actually quite good. It would be nice to be able to see some quantification of how good that was not obviously flawed.
Because disks are buffered, and fsyncing after every call (or forgetting to do so entirely) is silly.
I suppose somebody cares about how well they can expect their 124GB file to stream to disk, but for the rest of us mortals, we care about journalling support (check), a toolset (mostly check), and common-case performance, which in the *nix world involves a lot of reading and writing of small files.
I'd also like to see how these things perform under load, or when multiple benchmarks are running simultaneously.
I still haven't seen sensible benchmarks for ext4 with respect to how large directories scale, interleaved small file read and create, and small-file write with one fsync() at the very end (the only real world case.)
At this point, I have to wonder if the emporer has no clothes, or if the people posting benchmarks are just idiots.
I/went/ through the goddamn 8.01 TEAL pilot, then 8.02 TEAL. It was like chewing on glass. You spend easily 20-30% of the class time fiddling with the stupid response system, and less time getting through the material.
If you look carefully at the picture in TFA, you can see the vitality pouring out of these poor students. They're just awake enough to fiddle with their remote when prompted. Nobody's listening.
It's the professor who makes or breaks the lecture format. Frankly, I would have been sorely deprived if 6.115 had had a different format. The material was dense, but the prof knew how to draw in an audience.
Measuring failure rate in a curved class against an uncurved one where up to 15% of your grade is coming from brain-dead attendance (you can literally be half asleep) doesn't prove that TEAL is effective.
I have apparently managed to get the nomenclature of the various containers all confused in the past four years. That having been said, I still have a low opinion of MKV.
It's sad that in the interim between when OGM arrived, and then MKV bubbled up from the depths of hell, people forgot that OGM had literally everything they needed. And yet for some reason people decided to go with MKV, despite the abortive specification.
Xiph, I wish you had finished this crap years ago.
There's one major difference this analysis doesn't cover. If you're patient, you can get a dell for up to 40% off, and although it's not quite as drastic with Lenovo, the same is true. This macbook will ALWAYS be expensive.
Keep in mind, the reset forging strategy effectively did what you said is OK. They only needed to peek inside for RST.
They could have silently dropped the next 12 packets to and from that host and gotten the same result, and that would still be a horrendous shaping practice.
Correct, but when it comes down to it, any strategy a provider adopts is going to have problems. At best companies really do try to be reasonable, at worst, they find new and interesting ways to shape your traffic and prefer certain types of applications and protocols over others.
I'm not sure there's such a thing as protocol neutral. Different shaping tactics are going to affect different protocol's performance characteristics in different ways.
Yes, I'm sure she was planning on this being her personal private account for non-government use when she named them gov.sarah and gov.palin @ yahoo.com
The science has already been advanced. Dedicate the patents to the public, and let them work toward future advances based on those patents. If someone wants to try and sell you a system based on those patents, they deserve the benefit of competition.
Which is entirely unfair. A lot of people didn't have money for prep courses, or simply didn't have money for college, or didn't want to go to college.
If you're attending a class at a community college, it's because you want to be there, not because it's expected of you.
If you're going into math or engineering, you can expect to eventually take: Calculus Multivariable calculus Diff Eq Linear Algebra Discrete math, automata. Probability Algorithms... maybe some formal crap, too.
Having calculus out of the way's nice, since it's often a prereq for a lot of 'entry level' science and engineering courses.
I also went to an engineering school. I don't ever use calculus and other fancy math in the workplace, but calculus and other fancy math are tremendously useful in understanding many of the modern marvels about us.
As far as summations and differences, this is intuitively true. Vector calculus teaches the intuition for that sort of thing. But without the ability to integrate, you're going to miss out on certain things.
Calculus gives you the power to forget special case solutions and derive as needed in a lot of cases, which is pretty damned awesome.
Slashdot Mirror
Alright! And NOW I feel somewhat excited about ext4. I just wish for a change /. were posting things like this instead of the article in the summary.
It's not that important to me. But for once, I'd love to see fairly exhaustive benchmarks like I used to see for Reiser3. At least, at the time, they felt pretty good.
I imagine Ext4 is actually quite good. It would be nice to be able to see some quantification of how good that was not obviously flawed.
Because disks are buffered, and fsyncing after every call (or forgetting to do so entirely) is silly.
I suppose somebody cares about how well they can expect their 124GB file to stream to disk, but for the rest of us mortals, we care about journalling support (check), a toolset (mostly check), and common-case performance, which in the *nix world involves a lot of reading and writing of small files.
I'd also like to see how these things perform under load, or when multiple benchmarks are running simultaneously.
I still haven't seen sensible benchmarks for ext4 with respect to how large directories scale, interleaved small file read and create, and small-file write with one fsync() at the very end (the only real world case.)
At this point, I have to wonder if the emporer has no clothes, or if the people posting benchmarks are just idiots.
I /went/ through the goddamn 8.01 TEAL pilot, then 8.02 TEAL. It was like chewing on glass. You spend easily 20-30% of the class time fiddling with the stupid response system, and less time getting through the material.
If you look carefully at the picture in TFA, you can see the vitality pouring out of these poor students. They're just awake enough to fiddle with their remote when prompted. Nobody's listening.
It's the professor who makes or breaks the lecture format. Frankly, I would have been sorely deprived if 6.115 had had a different format. The material was dense, but the prof knew how to draw in an audience.
Measuring failure rate in a curved class against an uncurved one where up to 15% of your grade is coming from brain-dead attendance (you can literally be half asleep) doesn't prove that TEAL is effective.
Don't leave your developers out of your design discussions and brainstorming.
That's odd. I didn't see any small-file benchmarks in this article.
If this test had included small-file tests, you would see why reiserfs is relevant.
Anne Coulter just made a buttload of ad money.
Thanks slashdot.
Right, I get that. I just meant, you know. <3, as in an ascii heart.
Maybe it wasn't such a bad slip-up.
that was supposed to be a <3.
These little guys really cheer me up some days.
I have apparently managed to get the nomenclature of the various containers all confused in the past four years. That having been said, I still have a low opinion of MKV.
It's sad that in the interim between when OGM arrived, and then MKV bubbled up from the depths of hell, people forgot that OGM had literally everything they needed. And yet for some reason people decided to go with MKV, despite the abortive specification.
Xiph, I wish you had finished this crap years ago.
It's sad you miss the irony of that statement.
There's one major difference this analysis doesn't cover. If you're patient, you can get a dell for up to 40% off, and although it's not quite as drastic with Lenovo, the same is true. This macbook will ALWAYS be expensive.
Last I checked, MIT had all of 18.*.*.*...
Keep in mind, the reset forging strategy effectively did what you said is OK. They only needed to peek inside for RST.
They could have silently dropped the next 12 packets to and from that host and gotten the same result, and that would still be a horrendous shaping practice.
Correct, but when it comes down to it, any strategy a provider adopts is going to have problems. At best companies really do try to be reasonable, at worst, they find new and interesting ways to shape your traffic and prefer certain types of applications and protocols over others.
I'm not sure there's such a thing as protocol neutral. Different shaping tactics are going to affect different protocol's performance characteristics in different ways.
Yes, I'm sure she was planning on this being her personal private account for non-government use when she named them gov.sarah and gov.palin @ yahoo.com
Patents exist to promote scientific advance.
The science has already been advanced. Dedicate the patents to the public, and let them work toward future advances based on those patents. If someone wants to try and sell you a system based on those patents, they deserve the benefit of competition.
Which is entirely unfair. A lot of people didn't have money for prep courses, or simply didn't have money for college, or didn't want to go to college.
If you're attending a class at a community college, it's because you want to be there, not because it's expected of you.
If you're going into math or engineering, you can expect to eventually take: ... maybe some formal crap, too.
Calculus
Multivariable calculus
Diff Eq
Linear Algebra
Discrete math, automata.
Probability
Algorithms
Having calculus out of the way's nice, since it's often a prereq for a lot of 'entry level' science and engineering courses.
I also went to an engineering school. I don't ever use calculus and other fancy math in the workplace, but calculus and other fancy math are tremendously useful in understanding many of the modern marvels about us.
As far as summations and differences, this is intuitively true. Vector calculus teaches the intuition for that sort of thing. But without the ability to integrate, you're going to miss out on certain things.
Calculus gives you the power to forget special case solutions and derive as needed in a lot of cases, which is pretty damned awesome.