The way I read the summary was that the browser would maintain a "virtual window" inside of the real window. The real window could have any size; it is the size of the virtual window which would be quantised to 100px steps, and the gap between the real window and the virtual window would be the "letterbox".
Well, I'm surprised that someone only "discovered" this 4 days ago.
With this being a/8, I'd have thought it was quite well known before last Friday. In fact, I'm sure people have been well aware of it for quite a while now!
I remember the fight between one compiler vendor who had sizeof(int)==2 and another who had sizeof(int)==4 on the same CPU, the same OS. The first one argued that the external data bus was 16 bits, so it was a 16 bit machine, 16 bit ints. The other said the register file is 32 bits wide, so 32 bit machine.
Roads should be funded entirely by people who use them.
And what about all the other people that don't use the roads directly, but still benefit from them.
How is your food delivered to the supermarket? How does a fire engine get to your house when it's burning? How does the ambulance get to you when you're dying?
I'm sure there would be some drawbacks to a scheme like this, but what's the difference between a "web number" and a "telephone number".
Re:How do you know "how fast" a clockless system i
on
Clockless Computing
·
· Score: 1
While I agree that you still use a FLOPS benchmark, there is still the issue that FLOPS, MIPS etc are not terribly good benchmarks. Especially for modern processor architectures which have seperate integer, FP and vector ALUs (G4). In the vector ALU case, it can process 4 FP ops per clock tick, 4 32 bit ops per clock, 8 16 bit ops per tick, or 16 8 bit ops per tick. How to you combine all these possible scenarios into one benchmark?
You might have multiple parallel load/store units. How do you benchmark these?
Also, how do you deal with issues such as cache. With cache now being on the processor die, should it be considered when creating a benchmark for a given processor?
The purpose of any tool, whether it's a hammer, a TiVo or Perl, is to enable its user to do something.
Good analogy, but you still have to learn how to use a tool before it enables you to get something done. A hammer may be one of the simpler tools, but what if you hold it by the heavy end, and try poking a nail with the stick?
What about more complicated tools? You still have to learn how to use them.
If you go to a car dealership to test-drive a car, do you expect to have to spend five minutes reading the manual in order to figure out where the ignition is, and how to operate the seatbelts?
If you've never driven a car before, then yes you would be expected to read the manual. (Oh wait, you have to learn to drive a car, so your analogy is useless). People buying their first computer will never have set one up before, so should be expected to read the manual. When they buy their second computer, they'll know what to do without it.
Slashdot Mirror
The way I read the summary was that the browser would maintain a "virtual window" inside of the real window. The real window could have any size; it is the size of the virtual window which would be quantised to 100px steps, and the gap between the real window and the virtual window would be the "letterbox".
I'm intrigued by the concept of a changeable ringer (maybe that's too much of a local Scottish slang term?)
I thought the name of HAL in 2001 was derived from "Heuristic Algorithmic Learning"?
And it's pretty much a large part of "Manna" by Marshall Brain (well worth a read, by the way)...
http://marshallbrain.com/manna1.htm
Definitely not opening a link to "pee-nas".org from work!
Bueller. Bueller. Bueller.
Mind you, he never got caught.
> "I get more than a feeling"
> "Boston"
Very good.
Ding, ding, ding - geography lesson required!
Well, I'm surprised that someone only "discovered" this 4 days ago.
With this being a /8, I'd have thought it was quite well known before last Friday. In fact, I'm sure people have been well aware of it for quite a while now!
Which bit of "scotland" in the URL you posted did you not understand?
The RFID tag in the bog...
I don't know about your country, but toilets here in the UK aren't that technologically advanced yet.
Considering back in 1992 I paid US$300 for an 80 Gig drive, and RAM was $50 a megabyte...
Bloody hell, that was cheap.
Oh, wait you meant 80 Meg drive didn't you...
...All browsers are named after cars.
OK then, from now on, Mozilla will be known as "Persephone"
I remember the fight between one compiler vendor who had sizeof(int)==2 and another who had sizeof(int)==4 on the same CPU, the same OS. The first one argued that the external data bus was 16 bits, so it was a 16 bit machine, 16 bit ints. The other said the register file is 32 bits wide, so 32 bit machine.
What about the Analog Devices SHARC, where: -
sizeof(long) = sizeof(short) = sizeof(char) = 1,
but they are all actually 32 bits.
Roads should be funded entirely by people who use them.
And what about all the other people that don't use the roads directly, but still benefit from them.
How is your food delivered to the supermarket? How does a fire engine get to your house when it's burning? How does the ambulance get to you when you're dying?
My unix box is on 27/7, that's unlimited.
:)
Now that's seriously unlimited
You run that on a redundant system that you move to live after it has been testes right?
:)
I guess you really put your balls on the line if you don't
Its 2048bit encryption. Thats 2 megabits.
*cough*.. Kilobits!! *cough*
kibibits
Just stick in your bookmarks.
(Or on a homepage, like in the old days).
I'm sure there would be some drawbacks to a scheme like this, but what's the difference between a "web number" and a "telephone number".
While I agree that you still use a FLOPS benchmark, there is still the issue that FLOPS, MIPS etc are not terribly good benchmarks. Especially for modern processor architectures which have seperate integer, FP and vector ALUs (G4). In the vector ALU case, it can process 4 FP ops per clock tick, 4 32 bit ops per clock, 8 16 bit ops per tick, or 16 8 bit ops per tick. How to you combine all these possible scenarios into one benchmark?
You might have multiple parallel load/store units. How do you benchmark these?
Also, how do you deal with issues such as cache. With cache now being on the processor die, should it be considered when creating a benchmark for a given processor?
But when I brought all of this to my friend's attention, they just said that nobody uses Mozilla and blew me off.
He blew you off. My god, you must have some very close friends.
The purpose of any tool, whether it's a hammer, a TiVo or Perl, is to enable its user to do something.
Good analogy, but you still have to learn how to use a tool before it enables you to get something done. A hammer may be one of the simpler tools, but what if you hold it by the heavy end, and try poking a nail with the stick?
What about more complicated tools? You still have to learn how to use them.
If you go to a car dealership to test-drive a car, do you expect to have to spend five minutes reading the manual in order to figure out where the ignition is, and how to operate the seatbelts?
If you've never driven a car before, then yes you would be expected to read the manual. (Oh wait, you have to learn to drive a car, so your analogy is useless). People buying their first computer will never have set one up before, so should be expected to read the manual. When they buy their second computer, they'll know what to do without it.