Comment Not just accused - confirmed... (Score 1) 196
Spotted this over in the Ars sidebar:
http://arstechnica.com/gaming/2013/10/foxconn-admits-to-pushing-interns-into-overtime-to-build-ps4-parts/
Spotted this over in the Ars sidebar:
http://arstechnica.com/gaming/2013/10/foxconn-admits-to-pushing-interns-into-overtime-to-build-ps4-parts/
Why wouldn't he? An investor isn't trying to time the market that narrowly, so he'll pull as close to a real-time quote he can get confirm its still in the region he was looking for and submit the order.
If you're worried about the nebulous, evil effects of HFT affecting your order price between entry and execution, why enter anything but a limit order?
Actually, if you set a limit price on an order, it should close at the asking price if the asking price is lower.
Then my point still stands - if you enter a limit order, an HFT algorithm can't make money by raising the asking price then selling to you higher than you expect. Firstly, there's a limit price on your order; secondly, how is the algo supposed to raise the asking price without buying something? It makes no sense to buy high just so that you can annoy someone else, because the algo still ends up selling to them for less than it bought.
I think you have it backwards. Unless you are entering an at-market order, your order will be executed at the price which it is originally entered. So an algorithm which "bids up" the price, which it can only really do by entering an order with a higher execution price (just entering high quotes makes no difference if nobody executes), then executes a sell to you at your (lower) bid price, will *lose* money.
Surely though this is entirely dependent on the effects - expected and otherwise - of those extra proteins? The energy argument sounds convincing in its simplicity, but I am doubtful precisely because of that simplicity. We are not dealing with simple organisms in simple environments; the very existence of this article speaks to a complexity and subtlety of interactions which cannot be easily predicted.
From TFA: “The traditional expectation is that any sort of transgene will confer disadvantage in the wild in the absence of selection pressure, because the extra machinery would reduce the fitness,” says Norman Ellstrand, a plant geneticist at the University of California in Riverside.
Well, that seems like a foolish expectation. These modifications aren't already common in the wild, therefore they must be disadvantageous? This seems to be assuming that evolution has already made these plants as fit as they are going to get, and can't possibly be altered in a way that might make them more so (regardless of whether the alteration has any desirable side-effects). To me, it seems pretty stupid to assume that evolution has somehow peaked, for *any* species, given the time scales, diversity and mechanisms involved.
It's not often I come away from an article like this thinking "those stupid scientists, this is clearly wrong because of X", because normally they've been looking at it for a lot longer than I have and there is something (often a whole wealth of things) I don't understand or am not aware of. But in this case - those stupid scientists, this is clearly wrong because evolution will keep going unless we somehow eliminate all natural sources of genetic mutation.
Really, +5 Insightful? I actually thought the summary and headline were quite good. Let's take it point by point:
1. "A genetic-modification technique used widely to make crops herbicide resistant has been shown to confer advantages on a weedy form of rice, even in the absence of the herbicide."
Yep, that seems to be what they're saying - they took genetically-modified rice, cross-bred it with weedy rice, cross-bred the offspring to make a second generation, and found that the resulting plants were fitter than their weedy grandparents, according to several fitness measures.
2. "A common assumption has been that if such herbicide resistance genes manage to make it into weedy or wild relatives, they would be disadvantageous and plants containing them would die out."
Well, yes. From TFA: "“The traditional expectation is that any sort of transgene will confer disadvantage in the wild in the absence of selection pressure, because the extra machinery would reduce the fitness,” says Norman Ellstrand, a plant geneticist at the University of California in Riverside." Seems legit.
3. "But the new study led by Lu Baorong, an ecologist at Fudan University in Shanghai, challenges that view: it shows that a weedy form of the common rice crop, Oryza sativa, gets a significant fitness boost from glyphosate resistance, even when glyphosate is not applied."
This is taken almost word-for-word from TFA, so is also pretty accurate.
4. "The transgenic hybrids had higher rates of photosynthesis, grew more shoots and flowers and produced 48 — 125% more seeds per plant than non-transgenic hybrids — in the absence of glyphosate, the weedkiller they were resistant to."
Yep, these numbers come from TFA. The point being that this "extra fitness" was measured under normal conditions, with no glysophate application; so the weeds are not only glysophate-resistant, but natural selection will operate in their favour.
So yes, this is totally botched; an outrage, I say! Oh, wait...
Wow, what an angry response. MOC itself is one huge hoop, the necessity of which proves my point that Qt hasn't ever really been C++. The slots & signals stuff *can* be done in pure C++, see libsigc++ (as used by the glib and gtk C++ bindings), similar functionality in boost, or the contents of standard . Introspection is overrated IMHO, but it does now exist in the gobject world as well.
I know exactly what sort of hoops have to be jumped through for ABI compatibility, and I am aware of how many libraries piss all over them, which is why I stand by my statement that writing gtk in C made sense.
Have you seen the hoops that older versions* of Qt jumped through to maintain ABI compatibility? It wasn't really C++, it was C-with-classes-plus-a-bunch-of-horrible-macros. Based on when the project started, I think C was the better choice on the grounds of stability, compatibility and portability. If you don't want to do it in C, use the C++ bindings, which avoid (most of) the ABI compatibility issues by virtue of being an OO wrapper around something with a stable ABI.
* I say "older versions" because I don't work actively with Qt, but have read things here and there suggesting the situation is improved in Qt 5.
It's no less valid than the original Google-trends-based premise that interest in GNOME is declining. If anything, the fact that we have a GNOME-focussed article about this, but no KDE-focussed one, shows how bogus it is. TFA actually has a lot more to say, so perhaps the summary would have been better not to call out this one small point.
Mod parent up. If anything, this graph shows that KDE's downward trend is ever-so-slightly steeper, though I'm not inclined to call the difference significant.
Then when you need to scale it up, or introduce some serious data processing, you can just buy servers with larger core counts and... oh wait. It's all single-threaded. Never mind.
I prefer DERP. Delete, Edit, Read and Produce. Users are actually well ahead of the technology curve here - ask any tech support worker, customers have been DERPing for years...
I don't understand. What's so non-portable about a GLSL fragment shader?
I don't use KDE, but I can't help thinking that people who recommend disabling these things - and I've read a lot of similar recommendations - are somewhat missing the point. They're enabled by default because they're supposed to work. The mail client that ships as part of the application suite should not be useless.
I've had some painful experiences with Evolution on GNOME over the years*. I moaned about it, filed bugs, commented on existing bugs, and sometimes just had to give in and use a web-based front-end for email. My underlying feelings are much the same: Evolution is supposed to work, so at no point did I recommend to anyone that they just disable it and move on.
Don't sweep these things under the rug!
* What is it about mail, calendaring & contacts that seems to be so difficult for open-source software to get right?
It is easier to write an incorrect program than understand a correct one.