Forgot your password?

Comment: "Do one thing well" and pipes aren't the same (Score 2) 364

by davecb (#47930861) Attached to: Torvalds: No Opinion On Systemd

"Do one thing well" is how Unix kernel functions are written, and it's just plain a good idea. Systemd probably follows the first principle internally, many programs do.

Creating production systems[1] out of single-purpose commands connected by pipelines is a different principle, and only works if you keep them pretty simple. It's not a prionciple, but it is how a lot of Unix scripts are written, NOT including the shell that glues the parts together, and not including all the more complex programs, like ed or mail. Systemd doesn't follow the second, because it's more like ed than a text transformation like spell.

A more useful question is whether systemd as a whole does one thing, and does it well. About that, one might usefully discuss whether the Unix principle applies.

[Pipelines were patterned after a subset of "production systems" in early AI, which applied transforms to "produce" new things. They're not the kind of production systems you put on a raised floor]

Comment: Re:It doesn't scale (Score 1) 129

by davecb (#47929427) Attached to: FBI Completes New Face Recognition System

You can have 99.999 accuracy, and if the number of comparisons is (N choose 2), then the probability is (N chose 2) * 0.00001, which will be (really huge number * 0.00001) which is (merely huge number).

I don't care how good or bad the implementation is, it has to have more 9's to the right of the decimal than I have zeroes to the left in the number of people, N. That's a known hard problem in computer science (;-))

[And yes, Siemens was getting crappy even then, but that isn't the problem that the FBI has to solve]

Comment: It doesn't scale (Score 1) 129

by davecb (#47926777) Attached to: FBI Completes New Face Recognition System

We do this in Canada too, and it works where the number of people you're trying to recognize is small. The "birthday paradox"* says that if you're comparing 23 people, you have a 50% chance of a match. You have to multiply this by the error rate (usually much less than 2%) of a facial match program to get the false-positive rate, but it's still huge.

The German federal security service tried out Siemen's facial matcher years ago, found it had a low error rate... and was completely useless!

When you had hundreds of criminals to look for in thousands of airport passengers a day, it was directing insane numbers of people to the "capture the terrorist" queue (;-))


Comment: Re:Force of Law (Score 1) 355

by davecb (#47773045) Attached to: Ask Slashdot: What To Do About Repeated Internet Overbilling?
Engage a lawyer familiar with class actions *before* speaking to the police. You're an individual engaging in trial by battle with a huge company, and you need someone with the same degree of hitting power on your side before you start. They can advise on what's most effective in your jurisdiction. In Canada, the fraud squad is effective against enemies of moderate size. I don't see case law from them going after companies the size of a small country (;-))

Comment: Re:First World Problems (Score 5, Informative) 154

by Jahf (#47680769) Attached to: Ask Slashdot: What Recliner For a Software Developer?

To "First World Problems" ... for enough people for it to matter, this isn't a first world problem any more than coding is a first world job (and these days it isn't). As someone with hereditary back issues since my teens that, after 25 years of pain and bad surgeries with severe complications, I am now on disability. I wouldn't have made it 25 years without an extreme ergonomic solution that I paid for out-of-pocket because until recently employers didn't recognize that even "ergo chairs" aren't enough.

For the original question ... you're probably not as bad off as I was. But if you're getting close you have a few solutions depending on how much you can spend (or get expensed). If you're not looking for this level of solution, hopefully the article attracted someone who did.

* Not a recliner, but for helping your back similar to how an exercise ball does but with more support for long coding, I have used a "SwingSeat" at my desk for a decade or more ( This was good enough to get me by until the last round of surgery (which was actually a success) made it too hard to sit upright all day.

* A "Zero G" chair, like the ones by Human Touch (I use a Human Touch Perfect Chair), a bit pricey but not awful. Combined with a laptop table that can raise and angle the laptop you can actually do the vast majority of your work from a completely reclined position.

* For a desk, rather than reclining all the time, I assembled a custom desk from Anthro using their Console line that has a tray that can raise and pivot from sitting to reclining (which actually needs to be raised higher than sitting) to standing (great for those of you who don't have partial leg paralysis, good for me on rare occasions). The desk is HUGE but dang, it does everything. It is the Console with full pole extensions bringing it to almost 6' with monitors on a shelf 8" higher than the shelf that hosts the keyboard tray. Which means the monitors can raise up to standing position as well. I can, if I use a trackpad to avoid mouse problems, use this in combination with the Perfect Chair as a rigged solution similar to the next one.

* I couldn't afford this, or at least I could have before I needed it, but can't now. But if you can, the desks from ErgoQuest are perfect. And some of them are inexpensive enough to be not tooooo awful on the budget (if I had the money I spent on the SwingSeat, Anthro Console and Perfect Chair I could afford an ErgoQuest). You can -sometimes- find these on Ebay but not usually from someone willing to ship and often not for significantly less than having the right one built to your specs.

Comment: Re:Linux Cgroups are a good subset of this (Score 3, Informative) 161

by davecb (#47495245) Attached to: Linux Needs Resource Management For Complex Workloads
The only thing mainframes have that Unix/Linux Resource Managers lack is "goal mode". I can't set a TPS target and have resources automatically allocated to stay at or above the target. I *can* create minimum guarantees for CPU, memory and I/O bandwidth on Linux, BSD and the Unixes. I just have to manage the performance myself, by changing the minimums.
Data Storage

Linux Needs Resource Management For Complex Workloads 161

Posted by Soulskill
from the dirty-job-but-somebody's-gotta-do-it dept.
storagedude writes: Resource management and allocation for complex workloads has been a need for some time in open systems, but no one has ever followed through on making open systems look and behave like an IBM mainframe, writes Henry Newman at Enterprise Storage Forum. Throwing more hardware at the problem is a costly solution that won't work forever, he notes.

Newman writes: "With next-generation technology like non-volatile memories and PCIe SSDs, there are going to be more resources in addition to the CPU that need to be scheduled to make sure everything fits in memory and does not overflow. I think the time has come for Linux – and likely other operating systems – to develop a more robust framework that can address the needs of future hardware and meet the requirements for scheduling resources. This framework is not going to be easy to develop, but it is needed by everything from databases and MapReduce to simple web queries."

Comment: Re:Maybe, maybe not. (Score 4, Informative) 749

by davecb (#47453241) Attached to: Obama Administration Says the World's Servers Are Ours

The criteria is "the company that has the power to demand the data, has to do so if ordered by their country's courts". This probably dates back to the 16th century or earlier. Some time around the Hanseatic League...

A Canadian company with data in Outer Mongolia has to produce the data if it can. If the Outer Mongols prohibit the Canadian company from demanding it normally, the Canadians can't be ordered to produce it, because the data isn't in the Canadian company's control. If they allow it to be demanded normally, a Canadian court can get it. They have to do it via the Mongolian branch, they can't just issue court orders in Mongolia.

Your suspicion is correct: a Canadian company that controls data in the U.S. can indeed be ordered by a Canadian court to produce it .


Comment: Re:Stop throwing good money after bad. (Score 1) 364

by davecb (#47423377) Attached to: The Pentagon's $399 Billion Plane To Nowhere
Canada's government of the day is using it as a money-sink. Our requirements are for a twin-engine, long-range, non-stealthy aircraft with a moderate ground-support capability, such as the F-18 Hornet we now use. They rejected the newer super hornet, and so I fear the entire programme exists only to soak up money they don't want to spend on the priorities of the other parties...

Just because he's dead is no reason to lay off work.