Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror

Comment: Re:Misleading Title (Score 1) 336 336

"...officials believe that human error was to blame for the incident, rather than a problem with the robot."

Of course officials at the factory are going to suggest that first... because they're liable for mistakes of the robot, but not for mistakes of the worker. This is standard boilerplate and counts for nothing. Only a proper investigation can determine the truth. More 1%-vs-the-rest, really.

Comment: Re:What is the purpose of regulation? (Score 1) 668 668

Are the words of Bernie Madoff credible? Is the whole point not that he's a liar proven demonstrably to all rational people? Who cares if he's got some rationalization for hurting people. The rules of our society shouldn't be written by and for sociopathic personalities.

Comment: Re:Snake oil is everywhere (Score 5, Insightful) 668 668

"The prior shows a logical certainty, the latter [absence of evidence] is rationalization."

No, the latter is not mere rationalization; it is a logical use of limited resources (like time and money).

People can come up with a billion crazy theories or stories. We don't have time to test all of them or start using all of them by default. Hence, the responsibility falls upon the story-teller or seller to do the test and present evidence before anyone else gives them attention, time, or money in return. That's not rationalization -- it's simply rational.

As I say in my statistics classes: "The null hypothesis gets the benefit of the doubt; the alternative hypothesis has the burden of proof". (Or as Wikipedia puts it: "Rejecting or disproving the null hypothesis... is a central task in the modern practice of science, and gives a precise sense in which a claim is capable of being proven false. The null hypothesis is generally assumed to be true until evidence indicates otherwise.").

https://en.wikipedia.org/wiki/Null_hypothesis

Comment: Nads N. Nads (Score 5, Informative) 290 290

The writer, Nadia Drake (as listed in the byline at Wired.com), doesn't explicate until almost the end of the article: it's not that FB is misinterpreting her actual name as overly exotic, nor is she using a stage or business name, but her account is registered as "Nads N. Nads". She justifies this by saying that her friends commonly call her "Nads" for short and that she also wants to avoid a stalker. That might be justified, but the fact that she buries it near the end of the article, after a whole bunch of support for actual minority and Native American names, makes it feel just a bit self-serving. I would argue that proper journalistic practice would be to front-load this information in the first or second paragraph.

Comment: Re:Idiot (Score 1) 1067 1067

The major difference here is that in the first scenario the water stops in the sink, while in the second scenario it doesn't (so no division actually occurred). Every scenario you can think of will have this same problem; in your second version the material will have not been consumed, and you'll be in a situation of still needing to deal with it or pass it on, which is totally unlike the first case.

Reflect on why this is literally called an "overflow" error.

Comment: Re:x/0 does not equal 0. (Score 1) 1067 1067

There is a system called the extended real number line which does in fact have +INF and -INF as usable values. As you expect, division by zero is still undefined even in that situation.

https://en.wikipedia.org/wiki/Extended_real_number_line

On the other hand, there is an extended complex plane (Reimann sphere) with a single added "point at infinity" for which one can consistently define 1/0 = INF. But this is not the same as any standard computer number format, of course.

https://en.wikipedia.org/wiki/Riemann_sphere

Comment: Re:x/0 does not equal 0. (Score 1) 1067 1067

False.

"The expression 1/0 is not defined either as +INF or INF, because although it is true that whenever f(x) -> 0 for a continuous function f(x) it must be the case that 1/f(x) is eventually contained in every neighborhood of the set {INF, +INF}, it is not true that 1/f(x) must tend to one of these points. An example is f(x) = (sin x)/x (as x goes to infinity). (The modulus |1/f(x)|, nevertheless, does approach +INF.)"

https://en.wikipedia.org/wiki/Extended_real_number_line

Comment: Re:Infinity (max Float64) seems reasonable (Score 1) 1067 1067

"Divide by zero is infinity"

Yes, in the extended complex plane (Reimann sphere); however the "infinity" there is both complex and sign-less and doesn't match any standard computer number format.

No, in the extended real numbers (which has +INF and -INF, similar to standard computer systems), because defining 1/0 = +INF would still cause a contradiction.

"The expression 1/0 is not defined either as +INF or INF, because although it is true that whenever f(x) -> 0 for a continuous function f(x) it must be the case that 1/f(x) is eventually contained in every neighborhood of the set {INF, +INF}, it is not true that 1/f(x) must tend to one of these points. An example is f(x) = (sin x)/x (as x goes to infinity). (The modulus |1/f(x)|, nevertheless, does approach +INF.)"

https://en.wikipedia.org/wiki/Extended_real_number_line
https://en.wikipedia.org/wiki/Riemann_sphere

Comment: Re:Stupidity of Leadership (Score 1) 179 179

"The fact is, factory learning is dead, we just don't know it yet. We have spent the last 250 years in factory schools, built using factory ideas to populate our factories with workers. Today, we need a change in how we educate people, so that they are ready for information jobs. This requires scrapping the 'one size fits all' education model that is clearly dying (NCLB, Common Core etc), and replacing it with student paced education system where each student has a customized curriculum, based on ABILITY and WILLINGNESS to learn."

This argument has been around for what, a century or more now? And individuated software learning systems have been in use for over a half-century. See: the PLATO computer instruction system, starting in 1960. The truth is, the research on how well these systems work in practice has been consistently pretty dismal (I don't have time to get links at the moment, feel free to research). Last week a fellow at MIT and former MS researcher released a book basically saying that after a career of attempts, he's now convinced that technology cannot solve this problem. The most basic beginning instruction requires human guidance, and unfortunately that creates some bandwidth limitation in how many students a given teacher can attend to.

"But no matter how good the design, and despite rigorous tests of impact, I have never seen technology systematically overcome the socio-economic divides that exist in education. Children who are behind need high-quality adult guidance more than anything else. Many people believe that technology “levels the playing field” of learning, but what I’ve discovered is that it does no such thing."

http://www.washingtonpost.com/posteverything/wp/2015/06/04/technology-wont-fix-americas-neediest-schools-it-makes-bad-education-worse/

Comment: Re:There is no such thing as non-empirical science (Score 2) 364 364

You can't use the ancient Greek meaning. In the modern technical context, a hypothesis is something that can be tested. A theory is a larger body of explanations. Look here for the specifics of statistical hypothesis testing in the last hundred years or so. This is basic Stats 101 stuff:

http://en.wikipedia.org/wiki/Statistical_hypothesis_testing

"Just think, with VLSI we can have 100 ENIACS on a chip!" -- Alan Perlis

Working...