I'm in the UK where pretty much all lights get a three second yellow (approx) and there's a variable, but non zero, time when all the lights are red.

Drivers get used to this and, in London at least during busy periods, it's very common to have one or two cars cross after the lights have turned red (and that's when the average speed of the traffic probably isn't much more than 10mph so it's easy to stop)

Because cars do this, it's not at all uncommon for the junction to fail to clear before the other way goes green. Rather than getting 5-6 seconds for the cars to clear the junction, you get one or two seconds and sometimes the last cars who jumped the red are left in the middle of the junction when they're now at the back of the queue for the NEXT red light.

One thing that does seem different in the US (this is based on what I've seen in films so maybe not actually correct :-) is that drivers in the US seem to think they have the right to go if the lights are green. In the UK the attitude seems to be that you hit the horn if the cars who have jumped the lights going the other way are blocking you. It's still illegal but slightly less chaotic.

Actually, now I check it's not that simple - I don't know if it has changed recently or I'm just misremembering.

You match: Same day
then within 30 days
then the rest.

and you work out an average cost for each chunk.

http://www.hmrc.gov.uk/cgt/shares/find-cost.htm

Tim.

I don't know about Norway's rules but in the UK, yes.

In the UK capital gains are calculated on a last in first out basis where the asset is fungible - shares, gold things like that.

However, I'm not sure exactly how it would work for an asset like bitcoin that you had mined. In theory the electricity costs should be offsetable when you cash in. When you're just buying and selling it would work like any other share or gold.

Anyone doing serious bitcoin mining now (where electricity costs are going to be a substantial fraction of any notional gain) would be strongly advised to get professional advice - it might make sense to setup a company for the mining.

Skipping bail is criminal contempt of court in the UK

http://www.legislation.gov.uk/ukpga/1976/63

Penalties are severe:

http://www.cps.gov.uk/legal/s_to_u/sentencing_manual/contempt_of_court/

An immediate custodial sentence is the only appropriate sentence to impose upon a person who interferes with the administration of justice, unless the circumstances are wholly exceptional

He would have been (was?) allowed to argue his case that he shouldn't be extradited to Sweden due to the subsequent risk of extradition to the US. I cannot see how he could possibly convince a court that this would have happened and simultaneously show that the same arguments wouldn't have held water when extradition proceedings were in effect.

I can see he might possibly avoid jail in the UK only by surrendering and going to Sweden. But I wouldn't bet that the UK wouldn't want to extradite him back for contempt of court proceedings (or maybe they wouldn't extradite him to Sweden until after he was convicted and sentenced.

Then in the last page they say that even though none of this really matches up, they should still consider that he could be ROBPROF and they should keep an eye on him because his "background contains information inimical to the best interests of the United States" 8-(

That reminds me of the old joke:

Q. Why do the KGB go around in threes?

A. One can read, one can write and one to keep an eye on the two intellectuals.

it takes more effort to be antibiotic resistant than not. That means, absent the use of antibiotics, the resistance will naturally be selected against and fade from the population over time.

Actually, this (often) isn't the case.

It's obvious in theory that antibiotic resistance may or may not have a cost associated - but without any selection pressure, whether the resistance evolves is down to luck. Add the antibiotic and the selection is driven but remove the antibiotic again and the selection pressure doesn't need to be back towards the original state.

What is perhaps more surprising is that reversion to antibiotic susceptibility in the absence of the antibiotic is relatively rare - what actually tends to happen is that there are other mutations driven by the absence of the antibiotic rather than loss of the resistance.

http://www.biomedcentral.com/1471-2148/13/163
http://onlinelibrary.wiley.com/doi/10.1111/evo.12158/abstract
http://www.biomedcentral.com/1741-7007/11/14

The third one is interesting in that it says that sometimes antibiotic resistance can evolve due to a selection pressure unrelated to the antibiotic. If antibiotic resistance was very costly then you wouldn't expect to see this.

Yeah, it used to be optional (provided you never did it - once you'd done it once you were committed forever more) but eventually it became mandatory.

My current card I usually get a "Your card has been enrolled in VbyV and you're being redirected" but provided I allow the "cross site scripting attacks" that are generated it then goes through without any further prompting.

I remember that gner (now eastcoast) only used to work if I disabled javascript just before clicking on the BUY button

Verified by visa only reliably works on a vulnerable version of IE. Anything else and it's completely random whether a particular card/website combination will work.

In the end I changed my credit card to one that doesn't use VbV actually it's Mastercard so securecode (I think) because I got fed up of not being charged, being double charged, getting stuck half way through the process, forgetting my password which I then couldn't reset to something I wouldn't forget because it remembers the last 10^20 passwords, not being able to reset my password at all because it didn't give me the option.

It's particularly bizarre because my card might fail but my girlfriends card might work - for the same account in the same browser session.

SAP upgrades can easily take that long, but SAP can easily run organizations

This isn't my experience.

First SAP is deployed, then it goes though millions of "fixes" that make it about 5 years late.

Finally, everyone is forced to change what were working processes so that SAP can handle it.

Eventually things settle down - some things have improved, some things have got worse but a lot of money has been spent in the process and overall things are no better (I guess the CEOs reports are printed on shinier paper but I've never seen that side of things)

I meant market value at the time of issue. I had assumed that this is what par would be for options but actually par doesn't appear to have any meaning at all for options.

As to the rest, I don't know. You could be right but then it doesn't make sense to say he's getting $77M in compensation. The naive value of what he's getting is, in fact, zero in that case.

What am I missing here? For a company that's doing well, this seems like the perfect way to pay Larry.

His share options are almost certainly offered way below par.

If the shares are $33 and he gets 2M @ $3 then that's a theoretical gain of $60M when he exercises.

A 10% fall in the value of the company means he "only" gets $54M instead - while the investors who bought $2M at $33 lose $6M

I'm not 100% sure of the UK regulations but I think for share options schemes of this size (there are tax exemptions for some cases but I think the cap is £30000) you would be taxed as income on the gain when you exercised the options. (For gains of this size, 50% rather than 28% CGT)

For the US I think it's different and it's a way of avoiding income tax with fairly minimal risk.

The UK sees share option schemes like this as just another form of income, albeit deferred and not guaranteed - so the UK defers the income tax to the point where the gain is actually made.

The original linux kernel bug that was triggered by this bug occurred on x86

should, of course, be

The original linux kernel bug that was triggered by this behaviour occurred on x86

Erm?

I compiled to assembly. How did gcc know what platform I was going to run on?

The original linux kernel bug that was triggered by this bug occurred on x86 - because even on x86 a null pointer dereference doesn't have to cause a crash. You can map memory at address 0. It's just that usually you don't. There was a special case where memory could be mapped there allowing code that was unsafe to be reached because there was neither a crash nor a check for a null pointer.

The C standard says that dereferencing the null pointer is undefined behaviour. Therefore there are two cases the compiler has to consider:

1. The dereference does not involve the null pointer - therefore the check is unnecessary.

2. The code does dereference the null pointer - therefore all bets are off and the compiler is allowed to do anything it likes.

As a result, gcc eliminates the null pointer check as it's a legal optimization for both cases.

You need -O2 optimizations turned on. And you need to be careful that other optimizations don't eliminate the null pointer dereference before the null pointer check elimination.

(614) $ cat test.c
#include <stdio.h>
 
int main(int argc, char* argv[])
{
  char* f = argv[1];
  int a = *f;
  if(!f)
    return 0;
  printf("f=%p a=%d\n", (void*)f, a);
  return 0;
}
 
(615) $ gcc -O2 -S test.c
(616) $ cat test.s
.file "test.c"
.section ".rodata"
.align 8
.LLC0:<br>
.asciz "f=%p a=%d\n"
.section ".text"
.align 4
.global main
.type main, #function
.proc 04
main:
        save %sp, -112, %sp
        ld [%i1+4], %o1
        sethi %hi(.LLC0), %o0
        ldsb [%o1], %o2
        or %o0, %lo(.LLC0), %o0
        call printf, 0
        mov 0, %i0
        return %i7+8
        nop
.size main, .-main
.ident "GCC: (GNU) 4.3.5"

Why doesn't <ecode> format properly?

I haven't heard of any compiler that removes code just because it contains undefined behavior.

GCC does this because it assumes that the code is well formed.

Defererencing the null pointer is undefined behaviour, therefore, if you dereference a pointer, GCC assumes it cannot be null and, therefore, removes later checks that it is null.

Normally a null pointer dereference will cause a crash anyway, but on systems where you can dereference the null pointer this can cause unexpected behaviours:

int a = *p;
if(!p)
      return;

/* Code here is still executed even if p is null */