@GP: Instead of a plain complain, it's better to get an interesting discussion to write an explanation of why the post makes no sense, or instead find the good debunking comments and upvote them (there are two or three good comments near the top now).
Also, it's not my area. It's close enough to have a good guess, but in this case for me it's better to let someone else give an accurate reply.
---
Back to this post:
It obviously makes no sense. You have salt water, you extract the water, you have to get rid of the salt. Why waste time reading the details? [There are some interesting technical ideas about new surfaces, more on this later.] Reading the details their brilliant idea is to make salt cubes and sell them. So there is no waste!
When you get rid of the salt using brine, it's easier to transport and dilute the liquid. With solid salt you must scrape it form your high tech surface (without scratching it?!) and now the solid salt is difficult to transport. Also, to sell it you must purify it because it will include nasty things like crabs legs and sea smell.
Once you extracted the 99% of the water, it's difficult to extract the other 1% of the water because it's saturated solution with a low osmotic pressure, vapor pressure and a high boiling temperature. Also, water inside the block of salt is difficult to extract, and you must crush the small blocks.
Salt production is done in big salt lakes areas, where energy is "free". I like to consider it like a huge natural solar panel. You get heat for "free" and dry wind for "free". You must pay for them in an industrial facility. Also, the normal process still requires a lot of manual labor of guys/gals with [mechanical] shovels to makes piles of salt, wait, turn it a few times, wait, turn it a few times, wait, ... and you now have a nasty salt that you still have to purify to be able to sell it.
So they will get salt that is too expensive to sell, and too much of it to flood the market, and if you put it in the garbage can it will be classified as [industrial] waste.
---
The technical part looks interesting, but it's on the bottom of an unrealistic title and first paragraph. The interesting part is about the new surface with nano details and titanium oxide that absorbs Lithium. It sound interesting and they published it so there is some validation of the claim, but after the nonsensical first claims I'd want to take a look at the feasibility details.
---
>> Can we please ban university press releases
> Why?
I work in an university and I expect technical accuracy from the press department of an university. We want people to give us money in exchange of doing real and interesting things. We want people to trust the medical doctors when they give health advice, or a lot of other specialist about other public policies.
A lot of press release of the universities have a lot of exaggerations, burning the trust of the people. Before opening one here, I like to guess what is the real result and what is the bullshit part. I think that a complete ban of university press release here is too much, but I understand why the GP is annoyed.
Statecharts are one of the those concepts that are extremely useful for machine control that it seems crazy to me that they are so little adopted from what I have seen. At most people might know about a flat mealy state machine.
What's better than graphically designing your HSM, simulating it vs all kinds of scenarios to verify it, and auto generating the code?
How useful is this when you are using numbers in a reasonable range, like 10^-12 to 10^12?
Generally I try to scale my numbers to be in this range, whether by picking the right units or scaling constraints and objectives when doing nonlinear programming/ optimization.
The precondition on the link you shared has -1 <= x && x <= 1, so 99 is way outside of that range. But even so, testing for x=1, which is supposed to be inside that range, 0.5 doesn't seem tolerably close to 0.4142.
I have a suspicion that the accuracy number is the mean of accuracies over all valid floats in the range (or something approximating that), which is going to be weighted towards zero where the accuracy is higher, and perhaps where sqrt near 1 has some artefacts.
It is, there's a page in the documentation about how errors are defined. Let me also add: Herbie generally gives the most accurate option it found first, and then the other stuff might be useful for speed (0.5x is way faster than two square roots and a divide!) but it's not as accurate
The input should be the range and the distribution of probability on this range. Intuitively we have a tendency to assume an uniform probability for range [-1, 1] which is not the case if we check every doubles.
You're right I misread the graph. That said though I have played around with Herbie before, trying it out on a few of the more gnarly expressions I had in my code (analytical partial derivatives if equations of motion if launch vehicle in rotating spherical frame) and didn't see much appreciable improvement over the expected range of values, but then again I didn't check every single one.
What would be cool is if you could some how have this kind of analysis done automatically for your whole program where it finds the needle in the haystack expression that can be improved, assuming you gave expected ranges for your variables
Author here. I've got a few papers about this problem (including one in submission), but it is very very hard to do, especially with acceptable overhead. The state of the art is maybe 100x overhead.
The rocket company I worked at designed their orbital rocket in inches and lbm.
Engine flow rates in lbm/s, temperatures in deg Rankine, thrust in lbf. Btu/hour/inch^2/degR heat transfer coefficients.
I don't think they even had a way to do dc-dc voltage step-up and step-down at high power and efficiency, needed semiconductors for that to do high speed switching in buck and boost converters
reply