>There will always be preferred customers and I suppose a lot of these reservations are made in person, face to face and way in advance.
To be a preferred customer, come back a second time. All the starred restaurants have known when I've come back a second time and made a show of appearing to care about it.
FWIW, I recommend La Toque in Napa. 1 star, deserves 2.
I don't live near one either. I had to travel to Vegas. I don't expect super dupa dining to be on the vacation plans of a high proportion of people, but then you won't find me on a rock climbing expedition. It's worth working on your food phobias first, or you may be wasting money.
Run the numbers on high power (I.E. higher resolution) ultrasounds vs. autism and ultrasounds look a hell of a lot more likely to be a cause than the MMR vaccine.
The most I ever paid for a meal was $700 per head for a 16 course tasting menu at a 3 star restaurant. I booked 6 weeks ahead. It was money well spent.
I'm not delighted with the Flaimbait mod on this one. There appear to be some thin skinned Linux advocates who aren't familiar with deploying real time software in life critical situations.
I think it got hijacked and commercialized. There's probably great scope for a free, open source RTOS that is feature rich only when it needs to be. It just requires an unfeasible amount of effort by a multitude of talented, unpaid volunteers.
It's more than the scheduler. It's the memory management, driver interaction, IO semantics and probably a bunch of other things. If you don't build it RT from the ground up, you don't get RT at the application layer.
Linux, after all these years on Android, with every faster CPUs, still suffers from uncontrolled pauses and interruptions in user interaction that clearly are to do with scheduling and not to do with the available CPU resources.
This is because its scheduling tries to serve all purposes from database servers to desktop games. This is why it is unsuited for a car UI. QNX is fine.
If you had a 1979 copy of Wizardry on an Apple ][ floppy disk, you could images the contents. But if you wrote them back to a disk and tried to run it, it would fail.
This is because as a means of copy protection, Wizardry used track arcing. Part of a track was written on a track. Another partial track was written half a head-step away. The timing of the writes was synchronized so the partial tracks didn't overwrite. Anyone doing a naive read and write, or even a not-so-naive scan of the half tracks would fail, because they would get the timing of the writes necessary to prevent collision and to meet the consistency checks in the program.
Obviously people reverse engineered this and wrote adaptive copy programs that you could direct to do the right thing, but how is an archivist going to know that?
If you can get this level of deviousness on a primitive floppy disk, I imagine that there is plenty of deviousness to go around on other formats.
Getting stuff to work is one rather important aspect of getting stuff done. Also those Phi threads are big honking general purpose threads with lots of cache and ALU resources, not a highly strung state machine hanging of a matrix multiplier.
There is a small subset of problems that map to parallel threads of SIMD operations. Try optimizing IC layout on a GPU, or evaluating the biases in a crypto function by running the probabilities backwards through the gates. Those are not a problems for GPUs, but they are real world problems that need addresses and take a bucket load of CPU.
E.G, with a 90% chance of getting a red, the terrorist mastermind would need to send 7 terrorists through to get a 52% chance of one of them getting through unsearched.
Any terrorist with a simple grasp of binomial theory could work out the number of terrorists to send through the gate necessary to achieve a 90% confidence that one of them gets through with the bomb, given only the relative probability of red vs. green.
So we must prevent binomial theory getting in the hands of terrorists.
Slashdot Mirror
In my experience, the people who appreciate the best restaurants are usually pretty good cooks themselves.
It's a foody thing.
>There will always be preferred customers and I suppose a lot of these reservations are made in person, face to face and way in advance.
To be a preferred customer, come back a second time. All the starred restaurants have known when I've come back a second time and made a show of appearing to care about it.
FWIW, I recommend La Toque in Napa. 1 star, deserves 2.
This.
That's how I scored a table at Joel Robuchon. Opentable said it was booked. I called and it wasn't a problem.
I don't live near one either. I had to travel to Vegas. I don't expect super dupa dining to be on the vacation plans of a high proportion of people, but then you won't find me on a rock climbing expedition.
It's worth working on your food phobias first, or you may be wasting money.
Run the numbers on high power (I.E. higher resolution) ultrasounds vs. autism and ultrasounds look a hell of a lot more likely to be a cause than the MMR vaccine.
The most I ever paid for a meal was $700 per head for a 16 course tasting menu at a 3 star restaurant. I booked 6 weeks ahead. It was money well spent.
My priorities may differ from yours.
I know, but it doesn't stop a driver screwing you over.
There's a reason smart phones run the app processor on linux and the comms processor on a real time kernel.
This is how cable TV companies work as well.
I'm not delighted with the Flaimbait mod on this one. There appear to be some thin skinned Linux advocates who aren't familiar with deploying real time software in life critical situations.
I think it got hijacked and commercialized. There's probably great scope for a free, open source RTOS that is feature rich only when it needs to be. It just requires an unfeasible amount of effort by a multitude of talented, unpaid volunteers.
It's more than the scheduler. It's the memory management, driver interaction, IO semantics and probably a bunch of other things. If you don't build it RT from the ground up, you don't get RT at the application layer.
Linux, after all these years on Android, with every faster CPUs, still suffers from uncontrolled pauses and interruptions in user interaction that clearly are to do with scheduling and not to do with the available CPU resources.
This is because its scheduling tries to serve all purposes from database servers to desktop games. This is why it is unsuited for a car UI. QNX is fine.
If you had a 1979 copy of Wizardry on an Apple ][ floppy disk, you could images the contents. But if you wrote them back to a disk and tried to run it, it would fail.
This is because as a means of copy protection, Wizardry used track arcing. Part of a track was written on a track. Another partial track was written half a head-step away. The timing of the writes was synchronized so the partial tracks didn't overwrite. Anyone doing a naive read and write, or even a not-so-naive scan of the half tracks would fail, because they would get the timing of the writes necessary to prevent collision and to meet the consistency checks in the program.
Obviously people reverse engineered this and wrote adaptive copy programs that you could direct to do the right thing, but how is an archivist going to know that?
If you can get this level of deviousness on a primitive floppy disk, I imagine that there is plenty of deviousness to go around on other formats.
Keep the media.
I find a switch is an effective way of modulating a light source.
So a crunchy exterior with a soft squishy middle.
Yup, idiocy was what I was aiming for in my original post and predicates are ideal CYAs.
No he wouldn't. He'd dead.
Actually It's Blaise Pascal's theorem.
Could you simply extend the math to Markov chains to deal with this?
I haven't told them about the other transition they haven't noticed.
I program in gates. The dark side runs on my gates.
Getting stuff to work is one rather important aspect of getting stuff done. Also those Phi threads are big honking general purpose threads with lots of cache and ALU resources, not a highly strung state machine hanging of a matrix multiplier.
There is a small subset of problems that map to parallel threads of SIMD operations. Try optimizing IC layout on a GPU, or evaluating the biases in a crypto function by running the probabilities backwards through the gates. Those are not a problems for GPUs, but they are real world problems that need addresses and take a bucket load of CPU.
What makes you think the TSA would act rationally?
E.G, with a 90% chance of getting a red, the terrorist mastermind would need to send 7 terrorists through to get a 52% chance of one of them getting through unsearched.
Any terrorist with a simple grasp of binomial theory could work out the number of terrorists to send through the gate necessary to achieve a 90% confidence that one of them gets through with the bomb, given only the relative probability of red vs. green.
So we must prevent binomial theory getting in the hands of terrorists.