why do computer scientists say 1KB=1024 bytes?!!

How many programs do you know that require you to use binary, octal or hexadecimal numbers?

None?

Do spreadsheets require binary, octal or hex numbers?

No?

Then geeks _do_ care.

Hi One_Happy_Madman, You asked:

Why do computer scientists say 1KB = 1024 bytes ? ! !

Not using an N_bit address, e.g. using 1,000 instead of 1,024, would waste either memory or CPU cycles, and programmers/electrical_engineers would have to write more code/circuits.

Laymen think in terms of base_10 not base_2, e.g. 1,000 not 1,024, but since when have geeks concerned themselves with the layman ?

Exactly why such important specifications were written with 'octets' instead of 'byte'. When the documents were written, PDPs were still popular and made up a significant portion of the internet.

Over here lawyers are still fond of using mechanical typewriters. But I suspect a lot of the tender documents I've been getting are written in M$ Word.

The clients use the funny new i-units to prevent us from "cheating". But they could have easily specified kB=1024 in the definitions page or an appendix. Except that kb also usually is 1000 bits when talking about transmission medium. So I guess they adopted kiB to be less confusing. Which helps a little because tender documents tend to be vague and confusing anyway.

Some 'modern' CPUs also have BCD instructions. Though I can't recall which, it has been a long time since I've used BCD in my programs. Heck, it has been a long time since I've touched assembly.

It was, and was a usefull term to describe a bucket to put a BCD number into.

No, that's an architectural definition. A "byte" has no real meaning outside a specific architecture.

Exactly. A "could" and "size" are built into the architecture.

I'm not in the EU, I'm in Malaysia.

The CPU you're working on now probably has, because one of the data formats used by the 8087 floating-point coprocessor was "packed decimal": 18 decimal digits, packed two per byte, plus a sign.

Robert Israel snipped-for-privacy@math.ubc.ca Department of Mathematics

I thought a 'byte' is an 'unsigned char'. Is this definition flawed?

OneHappyMadman

"Only a madman would pursue chemistry as a career. The folks who do it right are happy madmen, many of the remainder become alcoholics." - "Uncle Al"

No, it's not bullshit. You had to look at both architectures to find out what they defined a "byte" to be and then do the translation. There is no universal definition of "byte".

What _are_ you rambling on about?

That is why a person in the biz who names things needs to take care. The oddest things, usually the one noun you don't want, become a common term.

Would you put in the word "copyable" in that definition?

/BAH

That's bullshit. How the hell do you think we did cross- architectural specs?

No, that's merely how you get a string of bits from here to there. Your confusion is why I would put "copyable" in the definition. Just because a PDP-10 load/deposit byte instruction didn't cross word boundaries didn't prevent us from concantenating bits and MOVEing them into an location. It just took two instructions instead of one. How do you think we implemented copying bits from/to a f****ng VAX?

/BAH

Ah, in my work, graphic just wasn't what we prep'ed. Most work input is text-based. The guys used those pinups to rest their eyes.

Good grief. Ain't modern tech wonderful? :-)

The economy in your games will be very limited. ;-)

That was so a person could kick the machine if all the other comm didn't come up. JMF and TW used a VT06 (I can never recall what the real name was) to do their stand-alone debugging. I don't remember what kind of hardware tweaking had to be done for that.

That's not a 33. Those glass teletypes were very robust but very expensive. Ours never had glass in them. We had a monitor developer who kept smashing them whenever he lost a wrestling match with the computer. Field Service finally refused to replace new glass. We could always tell where this guy had been working. The glass also had the feature that I could lay my hands on the glass to warm them up.

The iron TTYs were rugged. It's the plastic ones (33) that had a limited mechanical usage. Whenever I had to use one, I slowed my typing speed to about 30 WPM with a 3-4 characters' worth of waiting whenever I hit . I always screw this one up but I think the iron ones with glass were called ASR35s. Or the number was 36; for some strange reason the wrong number keeps popping out of my fingers.

/BAH

Tried? They *did* make and sell such computers. While no current IBM processor does addressing in decimal, the current top of the line machine still has instructions for decimal arithmetic.

IBM still uses binary coded decimal.

Especially the ones that know that it isn't.

FSVO. Since I'm not into S&M, I try to avoid the pit-bull software that you consider friendly.

I have yet to hear a real user say a good word about clippie.

And then there is Baud.

I thought the EU is extremely specific and contrained when tenders are submitted to anything in its juisdiction. So far this pickiness is the only thing keeping Micshit in line.

/BAH

Presumably base 10 BCD was used when applications programs were coded in assembler. The IBM 1620 was a 'pure' BCD computer with variable length fixed point numbers and IBM 360 had a set of instructions for handling BCD computations (may have been an optional extra). AFAIK the original local banking software was written in IBM 360 assembler - all it did was emulate the transaction facilities of ledger machines, together with daily summary and exception printouts for management and branches and statement printouts for customers.

Interestingly the IBM 650 had special currency hardware for accountancy for the British oriented market (1 pound = 20 shillings = 240 pence).

BCD instructions or at least fixed point instructions would be great for accounting and currency formats in spreadsheets!

I thought a large part of your economic improvment plans was to start doing a lot of business with European Union countries? To export anything to them, stuff has to be within specs. They've spec'ed the size of cucumbers for goodness' sake! ISTR the specs for bananas are controversial.

/BAH

Some irrelevant crap.

The reason why computer scientists say 1KB instead of 1024 bytes is because we all know we're speaking base 2 and we're not stupid.

pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow

$ free total used free shared buffers cached Mem: 312860 303920 8940 0 24732 51760

-/+ buffers/cache: 227428 85432 Swap: 1493960 0 1493960

I don't know how to determine that. )-; Kinda looks like "not at the moment", however.

And I've snipped the other stuff that I also don't have answers for. Sorry.

Now you've got me curious - I can exit KDE and start fluxbox, (or maybe even FVWM) and ask "free" again...

"ps ax" is usually kind of instructive as well. :-)

Thanks! Rich

...

That's with KDE running. I've stopped it, and started Fluxbox, and now I get:

$ free total used free shared buffers cached Mem: 312860 151652 161208 0 30844 66484

-/+ buffers/cache: 54324 258536 Swap: 1493960 0 1493960 richgrise@thunderbird:~

Before I started "Pan", and

$ free total used free shared buffers cached Mem: 312860 306412 6448 0 10284 123536

-/+ buffers/cache: 172592 140268 Swap: 1493960 12 1493948 richgrise@thunderbird:~ $

now.

Not that it makes any difference to anybody.

Just for the sake of insufferable pedantry:

$ ps ax PID TTY STAT TIME COMMAND 1 ? S 0:03 init [3] 2 ? S 0:00 [keventd] 3 ? SN 0:00 [ksoftirqd_CPU0] 4 ? S 0:00 [kswapd] 5 ? S 0:00 [bdflush] 6 ? S 0:00 [kupdated] 10 ? S< 0:00 [mdrecoveryd] 11 ? S 0:00 [kreiserfsd] 41 ? S