I accept that in modern common parlance, "exponentially" just means "grows significantly compared to the increase in its input". Your use fitted that.
However, when the media say things like (and this is *purely* illustrative, not a quote) "Support for the UKIP increased exponentially when Alex Wood stepped down from office, and apologised for his Nazi salute on facebook", then I get all smack-them-in-the-face-with-a-baseball-bat-that's-got-rusty-screws-sticking-out-of-it, and for good reason. Before vs. after does not a meaningful function make.
To be pedantic, the drag increases with the square of the speed, and the power required to overcome that drag increases with the cube of velocity. (F~v^2 and F~P/v, so P~v^3.)
What he's almost certainly saying is that *if you've got the scale on the map*, then *you don't need an in-map straight-line distance measuring feature*, as you can eyeball it. At no point did he say anything contradicting how vital the scale is; you're setting fire to a straw man there.
Yeah, and when I'm travelling on business I'd like to know what kind of areas have high levels of street prosititution, so that I can, you know, avoid them.
That was one of the more amusing cock-ups by google maps/earth for many years:
In Finland, the significantly Swedish-speaking parts were labelled in Finnish, and the Finnish-speaking parts were labelled in Swedish. In Belgium, the French-speaking parts were labelled in Dutch and the Dutch-speaking parts were labelled in French. (E.g. our screenshot maps were useless to the locals in Brussels when asking for directions and pointing, as they didn't recognise any of the street names - they didn't even know where they were on our maps.)
They (google) apparently had the smarts to realise that there might be more than one language in use in a country, but then had the dumbs to not get the selection right.
But he is also commenting on now-Oracle-controlled things he didn't create, such as Solaris. Quite why that was included and others weren't doesn't seem obvious.
Thank you for taking the time to write such a thorough response.
It might be the few good decisions that are made that cast the bad ones in such a harsh light, and I'm paying more attention to the latter rather than the former. I get much of my UK news from the likes of HIGNFY, which are themselves far from unbiased. Bad news travels both further and faster.
I recommenda 4th/5th-order Runge-Kutta extrapolation. It's so stable, I managed to model satelites which would move a quarter of an orbit per time step and not fly off to infinity. Simple 1st/2nd order extrapolation doesn't cut it - you'll end up with two things exceptionally close one tick, and then zipping away unimaginably quickly the next tick.
When I saw 360 degree videos in the headline, I looked at the bag I had just packed for my trip tomorrow, containing my ageing, but decent, video-capable DSLR, and my fish-eye lens. Nothing new about either of those parts, and that's got a 360 degree horizon when looking straight up or down. And my, the perspective...
Oooh, ooh, the UK's gonna grow a spine, after 35 fucking years of Conservative lapdogism (often perpetrated at the hands of those who don't actually call themselves "Conservative", but anyone with any awareness of the political compass (dot org) will attest that New Labour is indistinguishable from Old Conservative)).
Nahhh, not gonna happen.
Are you now going to accuse me of being a man from America too? (Handy hint - don't do that, you'd be very very wrong.)
Is it possible to make a quicksort run "well" in hardware? What are you going to do for the stack, and how big will you make it so that everything still works in the worst case.?
Compare that to a trivial network sort that makes use of the inherent massive parallelism possible in FPGAs.
Is your O(n^(1+eps)) really "running well" next to an O(n^(1/2+eps))?
Anything with a Drake-like equation may have many orders of magnitude error. Firstly, it assumes independence of the various factors, otherwise multiplying probabilities doesn't make sense at all. Whenever I see a Drake equation, my first reaction is to at least turn my bullshit detector to mute, so that I don't get deafened.
That's not true at all - if you want to read a number, then sscanf is your *best* option, as it doesn't have in-band error reporting, the return value is in addition to the read value.
You can use the "return to" exploit paradigm instead (that may be be return-to-libc or something else, all that matters is that you return to somewhere executable.)
Slashdot Mirror
I accept that in modern common parlance, "exponentially" just means "grows significantly compared to the increase in its input". Your use fitted that.
:-(
However, when the media say things like (and this is *purely* illustrative, not a quote) "Support for the UKIP increased exponentially when Alex Wood stepped down from office, and apologised for his Nazi salute on facebook", then I get all smack-them-in-the-face-with-a-baseball-bat-that's-got-rusty-screws-sticking-out-of-it, and for good reason. Before vs. after does not a meaningful function make.
C.f. "decimated"
To be pedantic, the drag increases with the square of the speed, and the power required to overcome that drag increases with the cube of velocity. (F~v^2 and F~P/v, so P~v^3.)
What he's almost certainly saying is that *if you've got the scale on the map*, then *you don't need an in-map straight-line distance measuring feature*, as you can eyeball it. At no point did he say anything contradicting how vital the scale is; you're setting fire to a straw man there.
Yeah, and when I'm travelling on business I'd like to know what kind of areas have high levels of street prosititution, so that I can, you know, avoid them.
> naming
That was one of the more amusing cock-ups by google maps/earth for many years:
In Finland, the significantly Swedish-speaking parts were labelled in Finnish, and the Finnish-speaking parts were labelled in Swedish.
In Belgium, the French-speaking parts were labelled in Dutch and the Dutch-speaking parts were labelled in French. (E.g. our screenshot maps were useless to the locals in Brussels when asking for directions and pointing, as they didn't recognise any of the street names - they didn't even know where they were on our maps.)
They (google) apparently had the smarts to realise that there might be more than one language in use in a country, but then had the dumbs to not get the selection right.
Nope, saying "6" is an arithmetic solution, we need a mathematical one.
6 + Min[t>=0]:(Max[i]:(competence(staff_member_{i}, time 6+t)) <= target_competence)
But he is also commenting on now-Oracle-controlled things he didn't create, such as Solaris. Quite why that was included and others weren't doesn't seem obvious.
I've worked out the optimal time to pull down their coffee machine.
Where will these strategists be without their coffee, eh?
I first encountered in "data flow" languages. And you're right, diminutive words are comparatively superior.
The summaries made it sound like it was just another kind of data flow paradigm. Languages like that are at least into their 4th decade now.
Thank you for taking the time to write such a thorough response.
It might be the few good decisions that are made that cast the bad ones in such a harsh light, and I'm paying more attention to the latter rather than the former. I get much of my UK news from the likes of HIGNFY, which are themselves far from unbiased. Bad news travels both further and faster.
I recommenda 4th/5th-order Runge-Kutta extrapolation. It's so stable, I managed to model satelites which would move a quarter of an orbit per time step and not fly off to infinity. Simple 1st/2nd order extrapolation doesn't cut it - you'll end up with two things exceptionally close one tick, and then zipping away unimaginably quickly the next tick.
Pripyat probably did it quicker - 45000 in a matter of days.
However, they left all the buildings behind. And most stuff they owned.
Exactly.
When I saw 360 degree videos in the headline, I looked at the bag I had just packed for my trip tomorrow, containing my ageing, but decent, video-capable DSLR, and my fish-eye lens. Nothing new about either of those parts, and that's got a 360 degree horizon when looking straight up or down.
And my, the perspective...
Has been as long as I've been aware of politics.
OK, I don't remember much about Callaghan's era, but remember everyone since then.
Oooh, ooh, the UK's gonna grow a spine, after 35 fucking years of Conservative lapdogism (often perpetrated at the hands of those who don't actually call themselves "Conservative", but anyone with any awareness of the political compass (dot org) will attest that New Labour is indistinguishable from Old Conservative)).
Nahhh, not gonna happen.
Are you now going to accuse me of being a man from America too?
(Handy hint - don't do that, you'd be very very wrong.)
Well, I know one whose CEO only knew 4 words...
developers, developers, developers, and developers.
Surprised the nappy never fell down during his monkey dance.
But your definition of "well" appears to include "desperately inefficiently"?
If lots of people think you're using a term badly, then maybe, just maybe, you're at fault - did that ever cross your mind?
Is it possible to make a quicksort run "well" in hardware? What are you going to do for the stack, and how big will you make it so that everything still works in the worst case.?
Compare that to a trivial network sort that makes use of the inherent massive parallelism possible in FPGAs.
Is your O(n^(1+eps)) really "running well" next to an O(n^(1/2+eps))?
That, or having a London A-Z.
Anything with a Drake-like equation may have many orders of magnitude error. Firstly, it assumes independence of the various factors, otherwise multiplying probabilities doesn't make sense at all. Whenever I see a Drake equation, my first reaction is to at least turn my bullshit detector to mute, so that I don't get deafened.
Better to be understood by one AC than by nobody, I guess...
That's not true at all - if you want to read a number, then sscanf is your *best* option, as it doesn't have in-band error reporting, the return value is in addition to the read value.
You can use the "return to" exploit paradigm instead (that may be be return-to-libc or something else, all that matters is that you return to somewhere executable.)
Ooooh, how about the guys behind SE-Linux ("SE" = "securty enhanced" or somesuch)?