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

This is the problem!

My hardware is described with hard numbers.

I buy a drive marked as 300 GB. For as long as I can remember, *B has referred to bytes. I am unsure when it changed. I would be very surprised to learn that everything in computing shifted to being base-10 while I was on the can.

The actual capacity is 279 gigabytes. 21 gigabytes are gone not because of how I formatted the drive or how the OS otherwise deals with it, but because harddrives are now being sold using SI-esque units while the hardware works in powers of 2. While the concept of converting units is not either foreign or difficult, I don't like doing it unless there is a reason for me to do so.

I want to be very clear on this, the only confusion is about what I am trying to say.

My drive was sold to me as 300 GB. Since all my software and my O/S treats things in powers of 2, I believe it is reasonable to expect 300 gigabytes. In actuality, it is 279. Why? Look.

It may seem trivial to you, but I view it as a redefinition of convention which pays NO heed to the reason why the convention existed in the first place. Imagine if overnight, natural log was redefined to be base 4. Or if mpg was redefined to be meters per gallon. Wouldn't that bother you just a little?

So yes, I would and do feel a little bit cheated.

I was just stumbling around. And the "couple of days" only involved a few hours each at her apartment, where I was welcome for short periods of time (another completely unrelated story).

Because that's the useful definition of a useful quantity, duh.

David Ames

Bloody hell, you are right.

My computer still uses the *ibibyte representations, though.

On 10/12/2005 the venerable snipped-for-privacy@aol.compost etched in runes:

. .

Does anyone know if Colossal Adventure is still out there somewhere? I still use the magic words 'Xyzzy' & 'Plugh' when I need a temporary filename. Sooo many people ask 'What's that mean?' when they see them.

In sci.physics, Rich Grise

wrote >> "Eric Gisse" wrote:

pse36 mmx fxsr syscall mmxext 3dnowext 3dnow

OK, standard dumb questions time.

My system is slightly faster than yours (1.4 GHz Athlon, 2802 bogoMips) and I can't say it's sluggish -- though with Quake4 it could be faster, and that's probably because I need a better video card (it's a BT5500 nVidia-based affair, and at that it's an improvement over the ATI 9000 that was originally therein).

I'll admit I do wonder.

There's often a lack of information that you might think is important. It would be nice to see some measure of battery performance on battery packages, for instance. But no hint is given on the difference between e.g. "heavy duty", "alkaline", and "digital", besides some meaningless jingle like "longer life".

But the average consumer doesn't plunk down $3000 on a digital SLR.

Perhaps. More like ensuring all bits are in a row.

/BAH

pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow

It isn't bloatware that is the problem.

It has to do with how computing service is measured. You keep assuming that more Hertz implies better performance. It does not if all the OS does is run the null job. The null job was our way to keep a CPU ticking while it waited for something useful to do. If resident OS code increases

Note that the terms I use are old-fashioned and DEC-biased. The guys are trying to correct my terminology to reflect today's usages but I seem to not do substitutions easily these days.

Sorry, hardware doesn't impress me. What does impress me is how the OS delivers computing services to the user without forcing the user to wrestle with it.

Do a mind exercise that counts all instructions to deliver this letter A from my TTY to your TTY screen. You have to include all the instructions that "pass on" the post from one server to the next. Compare a count that uses pure newsgroup software with the count that involves a web-based presentation. Include all instructions of user application software plus (and this is the important part) all the instructions than each and every OS, which merely passes the comm packets thru, executes.

If you're really curious, these aren't bogus mips. They are bookkeeping mips, setup mips, scheduling mips, and clocking mips. Think about how a computer knows what to do when (the biz calls this timing). Hint: Tick, tock, tick, tock. Why do you think Hertz is a metric?

/BAH

The term "nybble" came from a humorous article in Datamation. I never expected anybody to take it seriously.

A byte is a string of consecutive bits, and could cross word boundaries. Some computers, e.g., IBM 7030, had hardware to support bytes crossing words, while others, e.g., PDP-6/PDP-10, did not.

My guess is that it doesn't have enough memory to hold 5 M pixels but it does have enough memory to hold a compressed file that expands to 5 M pixels. It's another way to make you believe that your getting more than they're actually giving you.

Ah! That isn't the way software developers work. They use a modified form of the Scientific Method. I've never been able to describe this but essentially each investigation starts from scratch where a science investigation starts from the last success.

Oh, you were distracted. Solving computing problems required no distractions. My yoga teacher remarked how easily programmers could do certain yoga postures easily because some require not-thinking. Developers do this as automatically as breathing. Not-thinking in yoga is removing the chattering you hear internally.

/BAH

Good grief, absolutely not. Video terminals increased production by reducing the time waiting for the mechanical TTY innards to adjust themselves.

Why did you choose Python? Or is it what you have on-line right now?

I was banned from using ASR33s because I broke them from typing too fast.

/BAH

That's the ADVENTure I know.

/BAH

On our machine achitecture (PDP-10), I could define my byte sizes with instructions. I could have 0-36 bits/byte.

This is a flip answer. Lawyers use spellcheckers. What's worse is they probably use Micshit's spellcheckers. When I was executrix, the lawyer said that Micshit often changed the sense of a sentence by erasing or inserting a "not". I told him about file comparing. The procedures they used when working on contracts were horrifying.

/BAH

You are correct, the average person thinks in base 10. Even IBM tried to make computers which did computations in base 10 and used "Binary Coded Decimal" but eventually, as computers became cheap enough for "humans" to buy for themselves, the world of Binary had to merge with the world of Decimal. No, most people really don't care that 1 kbyte is really 1024 bytes instead of 1000. Most people don't even know the differences between a byte and a character.

What people do care about is that they can create documents that other people can read, without having to worry about it crashing in the middle of the document. They want others to be able to read that document without having to pay hundreds or even thousands of dollars for the software that lets others view the documents. They want to be able to find the document when someone asks about it - a week, a month, or even a decade later. They want to make sure that others can read the document later.

Microsoft has done a great job of making computers friendly and easy to use - - for a price. The Microsoft's greatest contribution to the industry is Clippy the dancing paper clip. He may be annoying to experienced users, but all of those helps are a great gift to the people who have never used computers before.

Open Source has done a great job of making computers friendy, easy to use, secure, reliable, and at a much lower price. For those more experienced computer users who want things to work faster, better, and with less effort, and more flexability and capability, at a lower price, then Linux is a very good choice for them.

It is described using computing biz terminology and metrics, not science metrics.

That wouldn't be a diversion; it would address your mistaken assumptions about how the computing biz defines its nomenclature.

And bytes were never defined. They still aren't. A byte could be a single bit or an unknown set of bits. A byte is generally a copy of the settings of n contiguous bits.

It was base-10 when salemen were talking to potential customers who thought in base-10. It was base-8 when machine architectures used octal addressing. It was base-2 when digital architectures were used. It was base-16 if the addressing was IBM-based.

I don't believe that you did an original formatting of that disk. In my day, we called the first time for DECtapes certifying. I don't remember what the word was for virgin disk packs. This initialization takes up room.

You have been told that this is not SI.

You are given a maximum number. There is no way you will ever get more bits out that pack. That is all you are given. There were customers who spent enormous amounts of money and human wallclock time to squeeze out every useful bit they could when this media cost more than your house does.

Now count the f****ng bits. Bytes are not precise enough. Do you insist that the milk you buy is measured to the accuracy of a nanogallon?

Possibly. Have you also considered that you may also be confused about how your physical disk looks and works? Have you ever looked physically at those bits and how they are arranged across the platters?

That is your maximum.

I've been trying to tell you why but have not managed to be able to do the writeup well enough.

This is not a redefinition of a convention. The convention was always done this way. And I'm older than you are.

You are trying to do a crank job. I'm about done here.

/BAH