Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×

Comment Re:It's that time... (Score 1) 342

It's important. The times that it's critical are rare. So... add if's it's in the middle of the road you prefer to stop rather than run over it. If it's up-right it's proper to dodge dangerously rather than to hit it. The number of crawling kids in the middle of the road is quite small, but it's larger than the number of infants, so add in something that smoothly increases the probability of human as it's (estimated) weight approaches 90 pounds and decreases it as it exceeds 300 pounds. Or 400. So you have a flattened bell curve with a smooth top.

But really, all this fiddling is just to handle corner cases. Usually you just stop or avoid the thing on the road without wondering much what it is. Only if you can't do either of those do you need the fancy figuring, which is a pain, because that's when you need the fast decision, so you "corner case handler" need to be something simple.
Rule 1: If it's standing up, it's a human. Don't hit, even if you must take damage. (This yields several false positives, but too bad. We need a quick decision.)
Rule 2: Estimate it's weight. (Ouch! That looks like a slow process...so while you're doing it, slow and start dodging.) If it's above 25 pounds, avoid even if you must take damage. (Note that hitting something heavy at a fast speed will damage you no matter what.) Continue slowing and preparing to dodge. If it's following a ball, dodge even if you must take damage.

Sorry, time's up.

This isn't a perfect approach, but it's simple, and doable. The hard step is estimating weight. There is a problem with false positives. A paper mache statue would count as human. But it should handle all common cases. And there should also be a distinction between streets where the traffic is slow and rare and streets where the traffic is fast and common. Freeways are much less likely to have humans walking in the road.

Additionally, there should be a rule about not overdriving your reaction time, especially on slow streets, but nothing can stop a kid from running out right in front of you from between two parked cars. And nobody, neither automaton nor human, can reliably deal with that. Which is why that first rule about "upright" is made to yield a lot of false positives. If you have time, then you can refigure things and perhaps decide that "that's a paper mache statute", so you may start to dodge in a way that will damage yourself, and then refigure to avoid damaging yourself when you, more slowly, decide that such action isn't needed.

Comment Re:It's that time... (Score 2) 342

Why? Just make it so that as far as the machines are concerned Gorillas are a subset of humans. And then keep the actual gorillas away from them.

You've got a reasonable point for more advanced machines, but for now I'd just as soon that they also avoid squashing dogs and cats...or, pretty much anything protoplasmic over, say, 5 pounds. Or 4. Slaugher house machines don't need to be intelligent, and shouldn't be. Not until things are FAR more developed.

And, really, wouldn't you just as soon that your car avoided running over that skunk? So if you adopt a variant of the precautionary principle, you can get most of the advantages without waiting for perfection.

Comment Re:Really ? (Score 1) 256

You need to plate it with teflon.

What bothers me is the idea of that being a colony rather than just an outpost. Where to you get metals? Can you split the CO2 into C + O2 and than use the C for bulk fabrication? It seems as if graphene can be either conductive or insulating, and nanotubes are pretty strong, but now we're talking about a rather extensive fabrication facility just in the initial set-up.

I consider asteroids a much more reasonable habitat. (I'm not sure that Mars is a good choice, but it sure sounds better than Venus.)

Comment Re:Why live there then? (Score 1) 80

Depends. How safe a neighborhood do you want? I believe that the normal asking rent for an apartment in Oakland was around $1500/month a few years ago...but I haven't actually been looking in the last 30 years, so I don't know what neighborhood is implied by that price.

12 * 1,500 = 18,000, so it depends on your other expenses...and whether you want to live that cheaply. OTOH, neighborhood is *VERY* important. And I also don't know what size apartment I'm talking about.

My suspicion, however, is that there was no intention of living in a downtown area, and that commute was as important as cost. Of course, for enough money you can find a sufficiently desireable location in a city, so saying money is the basis isn't incorrect.

Comment Re:Why live there then? (Score 0) 80

Units should be of the appropriate size to what you're measuring. Farenheit is more appropriate for judging room temperature and even cooking temperature when you don't need to be precise enough to get down to fractions of a degree.

The metric system has a lot of value, especially when doing precise measurements. When doing rough measurements at human scale it runs into problems. The meter is about the right size, but centimeter, or even decimeter, isn't a good replacement for foot. And for many purposes centimeter is too small to replace an inch. (When you start using fractions of an inch this advantage disappears.)

One can argue whether a Kilogram or a pound is the more useful general weight, it seems pretty much a draw to me. Ditto for Kilometer and mile. And when being precise metric is the clear winner.

OTOH, for outside temperature, a rough measue (Farenheit) is not only better suited in size, but also in accuracy. You don't get an accurate outside temperature, because it varies too much from place to place. So it's best not to pretend that you do. Which means avoid fractional degrees, whether Farenheit or Centigrade (okay, Celsius, but Centigrade is a better name).

Comment Re:The problem... (Score 1) 195

you shouldn't show speed,

Different people obviously have very diffferent requirements. In the UK, you can be fined huge amounts of money for a small excess of speed, so speed is important. If you are an older driver, the time taken to refocus from the dashboard to infinity (or vice versa) can be very long. However, spotting moose or crazy drivers is probably automatic (not that there are many moose on the road in the UK).

The question is, is it better than people at spotting small children running out from behind stationary buses?

If not, then it should avoid attempting to spot any problem, as it will lead to drivers relying on it the way they rely on satnav instead of knowing how to get anywhere.

The driver should be able to decide what is critical information. Since it is possible for the driver to change frequently, this needs to be set every trip, and probably needs to default to no info at all, as some people wont be familar with it. In which case it will probably rarely be used and can be safely junked. Your Moose May Vary.

Comment Re:Fucking Lawyers (Score 3, Interesting) 181

That would be sufficient to make the APIs stop working. Perhaps you should think again about what information is required and copyrightable.

I think that this decision may mean that Google will need to do something like alphabetize the API. (Customized organization can be copyrighted, but alphabetical order can't.)

Comment Re:Kids don't understand sparse arrays (Score 1) 128

It all depends on what you want to do with your matrices. Various operations have various costs in different sparse matrix formats. The standard ones are COO or coordinate format: a list of triples (i, j, val); DOK or dictionary of keys format: the hashmap you are thinking of; LIL or list of lists format: a list for each row and a list if pairs (j, val) in each list entry; CSR/CSC or compact sparse row/column: an array of indices where each row starts, an array of column indices and an array of values.

COO and DOK are great for changing sparsity structure; LIL is very useful if you have a lot of row-wise (or column-wise) operations, or need to manipulate rows regularly. CSR is great for matrix operations such as multiplication, addition etc. You use what suits your usecase, or change between formats (relatively cheap) as needed.

Comment Re:Welcome! (Score 1) 1083

That is only for tax purposes, more or less.

It is about rather more, and rather more important rights than that.

Marriage is about people being together, not about the government allowing it or not.

And that is only related to religion if you decide you want it to be.

Slashdot Top Deals

Are you having fun yet?

Working...