backtop

Print E-mail del.icio.us 205 comment(s) - last by lbaow.. on Nov 14 at 8:13 PM
Seagate CEO Bill Watkins has more than a few things to worry about: Seagate purchasers are entitled to free software or 5% cash back

With Cho v. Seagate Technology Holdings, Inc. wrapping up, Seagate customers who purchased a hard drive in OEM or retail packaging between the dates of March 22, 2001 and September 26, 2007 are eligible to receive one of two different settlement benefits per hard drive purchased:
  • The “cash benefit” is available to customers who purchased a Seagate hard drive between March 22, 2001 and January 1, 2006, and entitles customers to receive 5% cash back on the amount they paid for the hard drive, before taxes. Claims for a cash benefit must be filed using the mail-in form, which is available on the settlement web site.
  • The “software benefit” is available to customers who purchased a Seagate hard drive between the dates of March 22, 2001 and September 26, 2007, and entitles claimants to a free copy of the Seagate Software Suite, which retails for $40. Claims for a software benefit can be filled out online.

In both cases, customers must have purchased the hard drive as a discrete unit, as hard drives shipped with pre-built computers are not eligible under the proposed terms.

Cho v. Seagate was filed in April of 2005 by Sara Cho over claims that Seagate falsely advertised the capacity of their hard drives, overstating it by 7%. The nature of these claims lies in the difference between a gigabyte (1,000,000,000 bytes of 1 GB) and a giga binary byte (1,073,741,824 bytes or 1 GiB), as the abbreviation of “GB” is often used for both.

Seagate has denied and continued to deny both the false advertising claims and the fact that it has harmed anyone, and as of yet the courts have not ruled on the merits of the case.

Under the proposed terms of the settlement, Seagate will be required to:

  • Make “certain disclosures” about the nature of its hard drives’ storage capacity.
  • Reimburse affected customers with the benefits listed above.
  • Pay $1.8 million in attorney’s fees to the plaintiff’s counsel.

The final hearing date to approve the settlement will take place on February 7, 2008 at the San Francisco Superior Court.

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
5% vs. 7%
By mdogs444 on 10/23/2007 8:25:42 PM , Rating: 1
Just as a honest question -

How is that they only have to give back 5% when they took you for 7%? Shouldnt they be required to give you back 7% if they overstated 7%?

I know pricing of hard drives are close, even when they are 20, 50, 100 gigs difference, but its seems that it would make more sense to give back 7%.

Just my opinion.




RE: 5% vs. 7%
By Adonlude on 10/23/2007 8:40:05 PM , Rating: 2
Depreciation in the cost per GB over the last few years perhaps?


RE: 5% vs. 7%
By Justin Case on 10/23/2007 8:42:36 PM , Rating: 3
They didn't "take you" for anything. "Giga" / "G" means 10^9 and their "80 GB" drives do have 80 x 10^9 bytes. The problem is that most software (wrongly) considers that "Giga" / "G" stands for 2^30 (which is an incorrect use of a standard SI magnitude prefix).

http://en.wikipedia.org/wiki/Giga

Seagate is probably doing this because they've concluded that it's cheaper than paying the lawyers, even if they win (which I'm pretty sure they would).

If they had claimed their drives had 80 GiB and it turned out they had 80 GB, then they would have been fooling people. As it is, they were the ones being accurate; it's the operating system that reports the wrong size (smaller than it should). An 80 GB drive from Seagate (or any other manufacturer) does have 80,000,000,000 bytes (actually, they usually have a little bit more than that).

If these people want to sue anyone, they should be suing Microsoft (& others) for reporting the drive size in GiB and calling it GB.


RE: 5% vs. 7%
By TomZ on 10/23/2007 8:58:13 PM , Rating: 5
Justin, you should be smart enough to realize that the computer industry repurposed SI prefixes many years ago and has used powers-of-two to measure memory for at least 35-40 years. Your statement that "software is wrong" is itself wrong since it adheres to the classic definitions instead of the newer "binary" units that were much more recently defined.

Furthermore, the "binary" units have not been fully embraced by the computer industry. For example, semiconductor memory is always measured in powers-of-two. That is because their manufactured size is powers-of-two, just as HDDs are.


RE: 5% vs. 7%
By Justin Case on 10/23/2007 9:35:37 PM , Rating: 2
The "newer binary units" (I take it you mean the IEC binary prefixes, not units, since the units are the actual bytes) are "Ki" / "Mi" / "Gi" / etc.. Not "k" / "M" / "G" / etc., which are standard SI magnitude prefixes, based on powers of ten.

Seagate is absolutely correct when stating that their 80,000,000,000 byte drive has 80 GB. Sorry but just because some lazy programmers decided that 1024 was "close enough" to 1000 to call it a "kilobyte", that does not change the meaning of the SI magnitude prefixes used in just about every area of science and engineering.

1 kg isn't 1024 grams, 1 MW is not 1048576 watts and 1 GHz is not 1073741824 hertz.

1 GB = 10^9 B = 1,000,000,000 B
1 GiB = 2^30 B = 1,073,741,824 B

...just as...

1 GHz = 10^9 Hz = 1,000,000,000 Hz
1 GiHz = 2^30 Hz = 1,073,741,824 Hz

Software in general (and operating systems in particular) should either start to report the real size of drives, or change the prefixes from SI decimal to IEC binary.

The "computer industry" did not (and cannot) "repurpose" standard SI prefixes. Network connection speeds, interface speeds and media bitrates, for example, are typically measured using the correct SI multiples. The same goes for hard drives. It's only RAM and software reporting of file sizes / disk space that uses binary multiples, and it's about time that was fixed.

If Windows starts reporting years as having 255 days, does that mean calendar makers should get sued?


RE: 5% vs. 7%
By TomZ on 10/23/2007 9:56:02 PM , Rating: 5
quote:
Sorry but just because some lazy programmers decided that 1024 was "close enough" to 1000 to call it a "kilobyte", that does not change the meaning of the SI magnitude prefixes used in just about every area of science and engineering.

No, not true - it was done this way from the beginning of computer science (in the 1950's or 1960's, maybe earlier) because memory was always manufactured in powers-of-two sizes. Therefore it became convenient to have, e.g., KB = 1024 bytes instead of 1000 bytes. I'm stating facts here, not opinions - open any CS/CE/EE textbook written and you'll see the same basic definition.
quote:
The "computer industry" did not (and cannot) "repurpose" standard SI prefixes. Network connection speeds, interface speeds and media bitrates, for example, are typically measured using the correct SI multiples. The same goes for hard drives. It's only RAM and software reporting of file sizes / disk space that uses binary multiples, and it's about time that was fixed.

Again, oh really, then where did the original definitions come from, before IEC? Also, you don't have a good explanation as to why RAM uses powers-of-two while other things like bitrates do not. The reason is that RAM, like HDDs, are manufactured in increments of powers-of-two, so it is most convenient to express their capacities that way. Bitrates, however, are based on powers-of-ten crystals, e.g., 100MHz crystal.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 1:46:50 AM , Rating: 2
Just because something is "convenient" (read, easier to implement) does not make it correct.

And it certainly does not mean that people who use the SI magnitude prefixes correctly (as they are used in every branch of science and engineering) are all of a sudden wrong.

The SI prefix "giga" / "G" stands for "multiplied by 10 raised to the power of nine", period. It doesn't have different meanings depending on what you are measuring, just as the number "ten" doesn't represent different amounts depending on what you are counting.

And you're (also) confusing bitrates (a digital concept) with bandwidth (an analog concept). The bandwidth used by digital interfaces is usually higher than its bitrate.


RE: 5% vs. 7%
By mindless1 on 10/24/07, Rating: 0
RE: 5% vs. 7%
By Justin Case on 10/24/07, Rating: 0
RE: 5% vs. 7%
By TomZ on 10/24/2007 12:04:22 PM , Rating: 1
So what's your excuse...? :o)


RE: 5% vs. 7%
By Proteusza on 10/24/2007 12:10:07 PM , Rating: 2
Does the SI define what a byte is? We know them to be 8 bits (hence the "by eight").

Because, if not, then why should the ISO define how to count bits of bytes?

For an example, the Imperial system has feet and inches, among other things. 12 inches = 1 foot. But foot is not an SI measurement, neither are inches.

I guess, what he is saying, is that because the ISO doesnt define inches (unless they do?) they cant define kilo-inches or mega-inches. They also cant define what a foot is.

I think thats what hes getting at. Myself, I wonder if there is a better body for standardizing computer measurements than the SI.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 1:54:19 PM , Rating: 3
quote:
Does the SI define what a byte is? We know them to be 8 bits (hence the "by eight").


The actual unit of data is the bit; the byte is just a "working" amount that depends on specific system architecture details. Various sizes (from 5 to 12 bits) are used by different devices. Which is why some languages (ex., French) use the word "octet" to describe 8-bit bytes (the size used by the vast majority of personal computers).

Since a bit is simply a binary digit (a two-state entity) and "eight" is a common number, there is no need to "define" either unit (unlike, say, a meter or a foot, which needs to be defined as a physical length).

quote:
why should the ISO define how to count bits of bytes?


They don't. But they do define what numeric notations mean. It doesn't matter what you are counting; "kilo-" always means x10³ (x10^3) and x10³ always means "times one thousand".

Of course, you are free to decide that 10³ actually means (to you) "three hundred and twelve", but don't expect to get a job in engineering.

It would perhaps be simpler to have a standard definition for "byte" (which enforced the 8-bit size), but that would go against the original spirit of the term (if it was meant to be exactly 8 bits long, people would have called it "octet" to begin with; the term "byte" was chosen precisely because its size could vary - like bite sizes vary with the size of the mouth). In systems with parity and other error recovery mechanisms, one byte of data often takes up more than 8 bits on the "physical" medium, but it's still called one byte.

When documenting interaction between different types of systems, where byte size compatibility can be an issue, the term "octet" is frequently used instead of "byte", to mean "8 bits", even in English (ex., RFC / IETF documents).

English and Imperial / American units (which aren't always identical, despite using the same names) were not defined by a single body, they are a collection of "traditional" units from different fields, which have been redefined several times throughout history.

They are older than the ISO and CIPM, so they definitely weren't determined by those bodies, but they are currently (as of 1985) based on SI units (for example, a yard is defined as "exactly 0.9144 meters", a foot is defined as "1/3rd of a yard" and the inch is defined as "1/36th of a yard"). So, in a way, the CIPM does define them. Here:

http://www.statutelaw.gov.uk/content.aspx?activeTe...

None of that, however, alters the fact that the SI magnitude prefixes were already established by the time someone decided to coin the term "kilobyte". And I have no doubt that the person who did that knew perfectly well what "kilo" meant (as is clearly shown by the fact that he picked the power of two closest to the official meaning of "kilo-").

The problem was that the term was originally used merely as an approximation. 1024 bytes is "1 kilobyte" if you round it to the nearest kilobyte, just as 16386 bytes is "16 kilobytes" if you round it to the nearest kilobyte. People in the industry knew it was just an approximation. It was never meant to be extrapolatd to gigabytes and terabytes.

Normal users today do not understand (nor should they be required to understand) why that approximation was used, instead of using the terms "kilo", "mega", etc., as they are used in every other context ("kilowatt", "megahertz", etc., always meaning exact powers of ten).

When someone says that the diameter of the Earth is 8 thousand miles, that doesn't mean a mile is now defined as 1/8000th of the Earth's diameter. It just means that person decided to round the Earth's diameter to the nearest thousand miles, for practical reasons. In fact, a mile is defined as 1760 yards, and one yard is defined as 0.9144 meters (in other words, 1 mile is exactly 1609.344 meters).

quote:
I wonder if there is a better body for standardizing computer measurements than the SI.


What do you mean by "computer measurements"? Unless you are dealing with physical entities (space, time, forces, etc.), numbers are just numbers. You don't need to define the meaning of "1000" for "computers" any more than you need to define the meaning of "1000" for bananas. Information is an abstract concept; it's based on pure maths. The SI (and the CIPM, which defines the SI) has nothing to do with it.

If you mean computing terms in general (not measurements), then, again, those have nothing to do with the SI; they are defined by several different organizations, like the IEEE, ISO and IEC. But all those organizations use the SI magnitude prefixes when writing numbers, and they all agree that calling 1024 bytes "one kilobyte" is wrong . The "binary" prefixes ("Ki", "Mi", etc.) were created by the IEC - they are not part of the SI - to give people an easier alternative.


RE: 5% vs. 7%
By TomZ on 10/24/2007 2:12:51 PM , Rating: 1
You seem pretty smart, but maybe a little clueless about comp sci terminology. The meaning of a "byte" being exactly 8 bits is nearly universal. For processors that address a different number of bits at a time, that unit is typically called a "word," not a "byte."

Also, I've only rarely heard of "octet" used in English to mean the same thing as "byte" = 8 bits. That use is very uncommon in English - I've only seen it used in some of the older Internet protocol definitions.

I didn't get through the rest of your post - too long for me, sorry.


RE: 5% vs. 7%
By MrPoletski on 10/24/2007 2:33:25 PM , Rating: 2
I'm waiting for the base-9/octal number system for defining hard disk capacity lol


RE: 5% vs. 7%
By Justin Case on 10/24/2007 8:31:50 PM , Rating: 2
Waiting? You're about 30 years too late. Been there, done that, unlikely to go back. It's bad enough having to deal with big-endian vs. little-endian bytes, completely different sizes for "ints" (sometimes on the same system) and so on.

If you ask me, we should just measure things in bits (which is exactly what people do with modern media formats).


RE: 5% vs. 7%
By murphyslabrat on 10/24/2007 3:44:45 PM , Rating: 2
"Older protocol definitions" or not, my Intro to Networking instructor calls the bytes of an IP address "Octets".

While uncommon, it is still a valid name.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 4:37:48 PM , Rating: 1
It's not just valid; it's essential when you're dealing with devices that handle information in "chunks" of 9, 10 or 12 bits, for example. And those protocols are very much in use today (as I'm sure you know but TomZ above apparently doesn't).


RE: 5% vs. 7%
By TomZ on 10/24/2007 4:51:21 PM , Rating: 1
quote:
And those protocols are very much in use today (as I'm sure you know but TomZ above apparently doesn't).

Wow, you really do have a short memory, don't you? Here's what I said just a couple of posts up:

I've only seen it used in some of the older Internet protocol definitions

And, duh, those are probably the protocols the OP is studying in his/her class, right? E.g., TCP/IP, etc.?

Where else do you the term "octet" used in industry?

Actually, I'm surprised you aren't all upset that Unix uses the term "octet" in a different way, meaning a set of 3 bits for security permissions. After all, you obviously can't seem to comprehend the idea that a given term could have different meanings in different contexts.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 8:21:26 PM , Rating: 2
Actually, I read what you wrote (not that there's much point, except for either a bit of a chuckle or amazement at how some people are able to post and post and post about things they don't understand).

There's nothing "older" about those protocols. IPv6, for example, is so new it's not even implemented by most of the internet yet:

http://www.faqs.org/rfcs/rfc2460.html

The word "octet" appears more than 90 times just in that document. And you can find it in pretty much any internet protocol description (IMAP, SMTP, HTTP, etc.), as well as technical literature for networking hardware, and so on.

The term "octet" has nothing to do with "the industry". It's a numeric term which simply means "group of eight", just as a "quartet" is a group of four. PC system architecture is based on bit octets (groups of eight bits). Since other devices handle data in bytes of 5, 6, 7, 9, 10, 11 or 12 bits, whenever different systems need to interoperate, 8-bit bytes are specifically described as "(bit) octets" or "8-bit characters".

Using the word "octet" to describe UNIX permission masks is actually incorrect (not that that stops some people), since there is no "set of eight"; there is a set of four values (ie, a quartet), each of which can take one of eight possible values (which is, itself, a bit mask for three flags). For that reason, the correct term is octal (meaning "with eight possibilities"), not octet (which would mean "a group of eight things").

http://en.wikipedia.org/wiki/File_system_permissio...

So if you call UNIX permission masks "octets", congratulations, you're wrong again.

In fact, in UNIX (as anywhere else in the IT world), the term "octet" by itself usually refers to bit octets, also known as 8-bit characters, or 8-bit bytes. For example:

http://www.mail-archive.com/unicode@unicode.org/ms...

quote:
Unix filenames consist of an arbitrary sequence of octets, excluding 0x00 and 0x2F.


Or, of course, the definition of UUEncode (UNIX-to-UNIX encode):

http://www.opengroup.org/pubs/online/7908799/xcu/u...

quote:
The algorithm that is used for lines in between begin and end takes three octets as input and writes four characters of output by splitting the input at six-bit intervals into four octets, containing data in the lower six bits only. These octets are converted to characters by adding a value of 0x20 to each octet, so that each octet is in the range 0x20-0x5f [...]


Again, if you want to pretend that you're familiar with these things, try taking a 10-minute break from your posting marathon and do some basic research. With access to Google, even you can look like an expert (ok, maybe not an expert, but at least vaguely informed).


RE: 5% vs. 7%
By TomZ on 10/24/07, Rating: -1
RE: 5% vs. 7%
By Justin Case on 10/24/2007 10:41:24 PM , Rating: 2
quote:
You are far and away the most arrogant and overconfident person I've ever seen posting on DT.


It probably just seems that way to you because I call you on your bullshit, and (unfortunately for you) I actually know what I'm talking about; I don't just look up Wikipedia articles (or edit them) to pretend I understand things.

As I posted above, the fact that an octet is a "group of eight" does not prevent some (ignorant) people from mixing the term up with "octal". But if you bother to read the full Wikipedia article about file system permissions (which is linked to from the very article you quoted), or if you actually spend 7 years working with UNIX systems (as I have), you will learn that the correct term to describe the numbers used in file system permissions is "octal", not "octet".

Regarding the "dozens of posters correcting me", you must mean your imaginary friends; in this sub-thread, the only other person posting is murphyslabrat, who confirmed what I had explained to you (as will anyone else familiar with basic digital communication protocols).

As to "continuing to post over and over", you must be talking about someone closer to home. In fact, your claims that you are "a software developer" seem pretty unlikely considering the fact that you post around 10 messages a day, just in this forum.

Assuming you need to sleep and eat, that's about one message per hour. Add the time necessary to read the messages you're replying to (though I suspect you don't bother with that, otherwise you would have learned something by now) and the time necessary to research and confirm your answers (yeah, right), and I don't see how you could possibly manage to hold on to a job.

Unless by "software developer" you mean writing bits of Javascript in your parents' basement, of course.


RE: 5% vs. 7%
By croc on 10/26/2007 6:15:31 AM , Rating: 2
Octal generally refers to an 8 bit CPU, such as the Z80. Octet most generally refers to a segment of memory comprised of 8 bits, but is also sometimes used to refer to 'words' in a complex CPU instruction set. i.e, an 8 bit pre-fetch in a 16 bit memory schema would generally be referred to as 'the first octet'.

However, in the case of the above 8 bit CPU, they are interchangeable.

I have never heard either term used in any unix system that I have used to relate to file, ownership, or group permissions.

To quote from one of my favorite websites, ZZEE.com,

2.2. Numeric (octal) representation like "644"
If a numeric representation is used (like in chmod command, for example), then it is in the octal format (with the base of 8), and digits involved are 0 to 7. Octal format is used for the simplicity of understanding: every octal digit combines read, write and execute permissions together. Respective access rights for owner, group and others (in this order) are the last three digits of the numeric file permissions representation. Example: "0644". Here the second digit ("6" in the example) stands for rights of the owner, the third digit ("4" in the example) stands for rights of the group, the fourth digit ("4" in the example) stands for rights of others.

This table shows what numeric values mean:

Octal digit Text equivalent Binary value Meaning
0 --- 000 All types of access are denied
1 --x 001 Execute access is allowed only
2 -w- 010 Write access is allowed only
3 -wx 011 Write and execute access are allowed
4 r-- 100 Read access is allowed only
5 r-x 101 Read and execute access are allowed
6 rw- 110 Read and write access are allowed
7 rwx 111 Everything is allowed

We see that "1" stands for execute only, "2" stands for write only, "4" stands for read only. To combine the permissions you can simply add 1, 2 and 4 to get a needed combination. For instance, to get read and write permissions, you add 4 (read) and 2 (write), thus getting 6 (read and write). To get read and execute permissions, you add 4 (read) and 1 (execute), thus getting 5 (read and execute).


RE: 5% vs. 7%
By winterspan on 10/25/2007 2:12:19 AM , Rating: 2
That has to be the most outrageous claim I have seen on this website. I have been reading DailyTech regularly for a few years now, and each time I scroll down to the comment section I cringe at the thought that this "TomZ" character has once again soiled a perfectly good discussion with his arrogant, abrasive, and woefully misinformed comments.

As for the use of "octet" as a more precise definition of an 8 bit byte, it is widely used in hardware engineering.


RE: 5% vs. 7%
By Alexstarfire on 10/24/2007 2:58:48 PM , Rating: 3
I didn't really feel like reading all of your quite dumb posts to be honest. Sure, Kilo in SI is 1,000, but everything in the computer has been done in the power-of-two notation since computers began. Sure, they probably shouldn't have used SI to begin with, but they did.

The reason why people are mad over it is because if everything else uses the power-of-two notation then HDD manufacturers should as well. They know good and well how the OS will report it.

Would it make you feel better if they just changed the damn notation a bit?


RE: 5% vs. 7%
By Justin Case on 10/24/2007 4:28:46 PM , Rating: 2
quote:
I didn't really feel like reading all of your quite dumb posts


Looks like you should have, maybe you could have avoided publicly displaying your ignornace on the subject.

quote:
The reason why people are mad over it is because if everything else uses the power-of-two notation then HDD manufacturers should as well. They know good and well how the OS will report it.


So you think your "4.7 GB" DVDs hold 5046586573 bytes of data? You think your "3 GHz" CPU runs at 3221225472 cycles per second? You think that your 3 Gb/s SATA interface can transfer 3221225472 bits per second?

Guess what, most fields of IT (and electronics, which includes computer hardware) use the correct x1000 factor. It's only (some) software reporting of RAM and file sizes that uses the incorrect x1024 factor for "kilo" (and progressively wronger factors for mega, giga, etc.).

In fact, all relevant standards organizations specifically discourage or forbid the use of the term "kilo" / "k" to mean "multiplied by 1024". The IEC offers an easy "fix": just use "Ki" (and "Mi", "Gi", etc.) instead, and you don't even need to fix the maths.

If anyone is in violation of the industry standards here it's the software which still uses the term "GB" to mean "1024 x 1024 x 1024 bytes", instead of "one billion bytes". They can either learn to count correctly (1 GB = 1,000,000,000 bytes) or report the size in GiB (1 GiB = 1,073,741,824 bytes).

Seagate is absolutely correct, both in terms of accordance with the rest of the industry and in terms of compliance with the national and international standards organizations.


RE: 5% vs. 7%
By MrSmurf on 10/28/2007 5:10:05 PM , Rating: 2
Justin and Tom, you both have too much time on your hands.

Move on.


RE: 5% vs. 7%
By geddarkstorm on 10/24/2007 12:42:38 PM , Rating: 2
Oh really? Feet is a unit, as are meters, but meters belongs to the SI and feet belongs to the King's system. Units exist in different systems, as well as magnitude prefixes. If you are not using an SI unit, you cannot correctly use an SI prefix. In the same vein, if you use a unit of another system, you can use a prefix that also happens to be in the SI and have it equal something else. Your issue is with language.

Moreover, since when was the SI the all knowing rulers of the universe? I'm sorry to say, but in the computer industry, where Seagate resides, KiloByte and GigaByte refers to powers of 2 ALWAYS. The computer industry does not run on the SI system; the prefixes are not SI prefixes even if they sound the same.

BTW, I'm a biologist and researcher. We use the SI system all the time because it's convenient, but there are many other systems out there. The SI system is not the penultimate, nor do they define everything. As long as you have consistency, there isn't an issue. The computer industry has always been consistent with its usages, Seagate broke that consistency and therefore falsely advertised sizes. You probably don't know why the lawsuit against McDonald's and the temperature of their coffee succeeded, do you? It's because McDonald's heated their coffee above the industry standard; which when you are out industry standard of you are fair game.

Now, Seagate can claim that it's the HDD industry standard to use SI prefixes instead of IT (information technology) prefixes (even if the prefixes use the same word, they are different values! Don't be so easily fooled by appearances or think you know more than you do), and maybe they are right. But, it's also possible the courts will rule that the industry standard is the IT definition for Giga and Kilo, and then Seagate is screwed.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 2:19:18 PM , Rating: 3
There is no such thing as "the King's system" (what king would that be, anyway?). Furthermore:

1. Imperial units are currently based on SI units (ex., one yard is defined as "exactly 0.9144 meters"). As are virtually all other physical units in use today.

2. Magnitude prefixes are merely standard number notations; they can be applied to any number describing a quantity.

3. There is no such thing as "IT prefixes". Nowhere is "kilo-" defined as meaning "multiplied by 1024". In fact, most fields in IT (communications, storage, signaling, etc.) use the prefixes correctly. A lot of modern software also uses the correct terminology (ex., describing 3128 bytes as "3 KiB", and not "3 KB").

4. Computing and engineering terms are defined mainly by the IEC and IEEE. Both of which use the standard SI prefixes, and both of which agree that "one kilobyte" means "1000 bytes". There is absolutely no organization recommending the use of "kilo-" to mean "multiplied by 1024".

5. Seagate did not "break the consistency". Hard drive sizes have always been correctly reported (80,000,000,000 bytes = 80 GB), by all manufacturers.

It's amazing how some people feel the need to post highly inflamed speeches about how others are doing things "right" or "wrong" when they don't have a clue about history or about the underlying standards and terminology. Maybe you, as a biology researcher, could try to explain that phenomenon...


RE: 5% vs. 7%
By geddarkstorm on 10/24/2007 3:00:20 PM , Rating: 2
Oh man, there are many problems with what you said.

1. Calling "Imperial" as "King" is an expression since the kings/queens of england set the rules and changed or reaffirmed the measurement system at their leisure http://www.britannica.com/eb/article-9016522/Briti... . Moreover, this system came FIRST, long before metric; metric was later and a totally different system. Converting between yards and meters does not mean they belong to the same system, just as pounds and liters are of the totally two different systems. Because MATH is universal, you can translate between them all.

2. They can, but they can also be defined however it is wished. Romans used a base 12 system, to them Kilo would be whatever times 12 to the whatever. In computer science, base two system was originally used for whatever reasons (most likely based on bits either being 1 or 0, so one of 2 choices).

3. There isn't, I used that term for the purpose of distinction, and in computer science that is how it's defined, now isn't it? At least that was how it was originally defined.

4. Interesting, as even though you claim that, the convention that it does mean 1024 has been commonly used for decades. Certainly both conventions have been used at once, but in different areas.

5. You think that, as do I. I never said they did, BUT, I don't know what the industry standard is defined as, and it may well be that the courts will rule that it is expected that giga and kilo mean the common definition we see for file sizes and what is typically taught in schools. The courts will decide, as I stated, as they did with McDonald's (even though McDonald's had a nice disclaimer saying how hot their coffee was, they still got burned (heh)). Then they'll basically say if you are right and wrong, till then it's your opinion nothing else.

You love to spread disinformation. The fact kilo is used to mean 1024 bytes in most applications is irrefutable, it is historically so and has been in the sciences as well as in software and the school system. If that is changing, then it is changing. SI didn't always exist, and it brought huge changes with it in the hopes of uniting the different fields of science who were using different measures. That could be happening here, and maybe even the whole kilo is 1024 is an error that happened long ago and propagated: But it still exists, and it is still in the common mindset, so trying to say it's so horrible wrong is in itself wrong. It is what it is and it could change or it could not, we will see! In any case, when it comes to the "bit" and "byte" unit, as far as I've ever seen without exception with the current days software, any kilo or so prefix is fit to the 1024 times system; if new stuff is changing that then it is, but it's still historical.

I also am amazed at how utterly fervent you are about this issue as if someone's holding a gun to your head. Oh my, it's not being used in the proper SI context! The world will EXPLODE! Gees, relax for once.


RE: 5% vs. 7%
By Alexstarfire on 10/24/2007 3:03:38 PM , Rating: 2
Do you enjoy being a jackass or are you truly just trying to piss people off. The Imperial units of measure were NEVER based off of SI. I don't even see how that is possible since SI came after Imperial. Couldn't tell you what king it was, if it even was a king, but it is said to have originated in England from one of its rulers.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 4:00:54 PM , Rating: 2
Do you enjoy displaying your ignorance in public forums, or are you just too lazy to do some basic research?

http://en.wikipedia.org/wiki/United_States_customa...

quote:

U.S. customary units, also known in the United States as English units [...] are defined in terms of SI base units [...] most countries, including the United States, redefined their customary units in terms of SI units like kilogram and meter.


http://www.statutelaw.gov.uk/content.aspx?activeTe...

quote:

The yard or the metre shall be the unit of measurement of length and the pound or the kilogram shall be the unit of measurement of mass by reference to which any measurement involving a measurement of length or mass shall be made [...] the yard shall be 0.9144 metre exactly [...] the pound shall be 0.45359237 kilogram exactly.


Seriously, is this forum full of 12 year olds, or do people just sleep through highschool nowadays? You never had a physics or maths class about units of measurement...?


RE: 5% vs. 7%
By BVT on 10/24/07, Rating: 0
RE: 5% vs. 7%
By highlandsun on 10/26/2007 8:05:51 AM , Rating: 2
quote:

5. Seagate did not "break the consistency". Hard drive sizes have always been correctly reported (80,000,000,000 bytes = 80 GB), by all manufacturers.


If you believe this then you are clearly too young to know what you're talking about. The first hard drive I remember holding was a 5MB Shugart. This is important to note, because these facts *are* true:
1. hard drive sectors have always been 512 bytes

2. disk drives have always been rated in whole number capacities, with a whole number of sectors

3. 5,000,000 / 512 is not a whole number ( = 9765.625, fyi.)

The fact is that historically, hard drives were rated in units of 1024-byte kilobytes, just like other computer memory-related hardware. It was a natural outcome of the hardware layout: There was always an even number of tracks and an even number of sectors per track. And of course the controllers used to manage the drives obviously were based on powers of two, like all other computer hardware.

LBA (Logical Block Addressing) for hard drives was a relatively recent convention. Before that, all computers used disk geometry (heads/tracks/sectors) to operate hard drives. And these parameters were always constrained to 4 bytes (0-255 heads, 0-65535 tracks, 0-255 sectors) and pretty much always used even numbers in each field, which pretty much guaranteed that a drive capacity was an even multiple of 4096 bytes. (Which is also one reason that 4K has been the most common filesystem page size for over 3 decades, because it would always use every sector on a disk with nothing left over wasted.)

It was only when disk capacities finally exceeded the range that's expressible in C/H/S notation, and LBA took over, that the close correspondence of disk capacities with 4K blocks disappeared. Because once you hit 1GB, pretty much any capacity above that divides evenly by 512 anyway, whether it's K-aligned or not.

I wonder how many milliyears you've been around computers, or if you've even been within a kilomile of an original Winchester disk.


RE: 5% vs. 7%
By Justin Case on 10/29/2007 6:23:36 PM , Rating: 2
I've been around computers since before hard drives were invented, so you do the maths (the first hard drive I ever used was 10 MB, though), and I wouldn't even consider the days of manual CHS configuration as "old history".

This is not, however, about sector sizes (which are typically not even measured in kB, let alone GB), it's about the size of drives.

80 GB (for example) is not a power of two (and neither is 80 GiB), so the whole argument about using powers of two is void. Seagate uses the term correctly, so do all other manufacturers, and so do all the relevant standards organisations.


RE: 5% vs. 7%
By Fritzr on 10/30/2007 3:40:28 AM , Rating: 2
The sector size used by the drive's internal OS is a power of 2. There is no particular reason it needs to be, it has nothing to do with the physical characteristics of a drive. The manufacturer is quite free to adopt metric sectors of either 500 or 1000 bytes. However their software engineers would probably be seriously upset over having to write algorithms to handle base 10 math using base 2 units. Having written subroutines to convert binary digits, octal digits, decimal digits and hex digits for a simple assembler I can say that by sharing a multiply by 2 subroutine binary, octal and hex required 9 bytes of code each. Decimal require 22 bytes of code since it cannot be handled as a binary value.

The total number of sectors on the disk is set by technical considerations. First it is decided how many tracks the disk will have. Second it is decided how many sectors can be fitted on each track. You can get more by assigning more sectors to the outer tracks, you get simpler addressing if you assign a fixed number of sectors to each track. The final decision is how many total heads will be used. 1 platter is usually 2 heads. Multiple platters can be single sided, but are often double sided. The top and bottom surfaces of multiple platter drives may be left unused.

When all those design criteria are settled you next determine the capacity of the drive by adding up the number of sectors on "in use" tracks. I phrase it that way since some of the tracks will be assigned as backup that will be invisibly allocated when an "in use" sector is unusable. Under the CHS addressing the sector was simply marked bad and attempts to access it returned an error. Under LBA a backup sector is read/written when a bad sector is accessed.

The number of KB of storage is 1/2 the total sectors free. For a drive of around 80MiByte capacity this will be 80*1024*2 sectors. There is no reason whatsoever that the drive manufacturer cannot use the industry standard definition and offer an 80GB disk drive that has 80GB free. When the file system is setup there may be some file management usage that allocates some sectors to OS use, but the total storage including this OS usage should be 80GB as reported by the file manager on a drive that is advertised as having 80GB formatted capacity.

Another anomaly was that drive size limit set by the CHS addressing. The reason for the limits on the values that could be used was that the drives & the controllers did not agree on the data fields in the CHS value. This meant that the only legal values were the ones that did not use the bits that were disagreed on. This could have been fixed by either the controller manufacturers using the CHS bit field assignments used by the drive manufacturers or by the drive manufacturers using the bit field definitions used by the controllers. This change was never made. The size limits were "fixed" by software that could access sectors with illegal CHS values.


RE: 5% vs. 7%
By mindless1 on 10/24/2007 2:27:16 PM , Rating: 2
Byte is part of the computer industry's binary system. Using a decimal prefix to a binary unit is an invalid expression within this system if you attempt to apply it without the rules of that system, you've missed the whole point.


RE: 5% vs. 7%
By mathew7 on 10/24/2007 2:47:02 AM , Rating: 2
First of all, HDD's have nothing with powers of two. They are completely analog storage. They even have the sector storage much bigger than the 512 bytes we see (ECC & stuff). HDD vs RAM is like CRT vs LCD. Like you use sync periods in CRTs, the HDDs use ECC and positioning algorithms, in contrast with RAM/LCD which use a direct row/column mapping.

Second, what's the next step? Law-suit to DVD-RAM group? You do know that DVD is advertised as 4.7GB, while it actually has 4.4 GiB? Of course, 7% of 4.7GB is MUCH less than 80GB. Also, when you write in the ISO format, you don't have such a big FS overhead.


RE: 5% vs. 7%
By mindless1 on 10/24/2007 4:49:41 AM , Rating: 2
Quite wrong, they are not analog storage. Do you know what digital storage is? It is quite valid to use a varying threshold hi/low value to determine a binary digit.

HDD vs RAM is not like CRT vs LCD. An old audio tape is analog but a hard drive stores DIGITS.

That by definition, is digital. I suppose you'd say a CD is analog too? You'd have to using your logic since just as both mediums have binary digits written, they are read back against a threshold value which is not a hard 0 or 1.

Next step? I say give HDD manufacturers one last chance to start advertising capacity valid for the environment in which the device is used. If they don't, fine them $1000 a drive. If we can make examples of individual citizens then why not a hard drive company?

NOthing against Seagate, they just happened to be the brand du jour, and have an equal opportunity to restate capacity based on the product use, which is the most appropriate context possible.


RE: 5% vs. 7%
By Veraiste on 10/29/2007 5:12:27 PM , Rating: 2
Why not force the software makers to go in line with *everyone* else? No one else uses that horrible approximation. It is solely used by a portion of the software industry in reference to disk space.


RE: 5% vs. 7%
By Proteusza on 10/24/2007 5:01:59 AM , Rating: 2
mindless1 is entirely correct. Just because something is stored with voltage levels, does not mean it is binary.

Hard drives are binary, each position on the platter is a 0 or 1.

Computers have been measuring everything in powers of 2 since they were made, because of the binary system. There was a time when MB stood for 1024 KB, and GB stood for 1024 MB. the SI defines that as incorrect, fair enough, but, for the early days of the computer industry, thats exactly what they meant. Many people are still confused between GB and GiB, and I dont think that either the SI or hard drive manufacturers have done enough to make people aware of the difference.

Just a question, how does the SI force others to adopt its meanings for prefixes? Are there laws in every country forcing SI compliance? Just wondering.


RE: 5% vs. 7%
By Oregonian2 on 10/24/2007 2:27:10 PM , Rating: 2
quote:
Hard drives are binary, each position on the platter is a 0 or 1.


Yes, and the bytes in a 80,000,000 "80GB" hard drive are binary bytes.

It's just the "80G" part that's not binary as such, and never has been for hard drives as long as I remember (been a designer of microprocessor based systems since 1973).

Drives aren't inherently built in binary multiples. The number of tracks on a platter is how many that fit, not something an even binary multiple. The number of bytes on a track is how many that fit, not some even binary multiple. Even the number of heads (platter surfaces) isn't a binary multiple sometimes. The 750GB WD drive I just bought on sale at Fry's has 5 platters, meaning 10 heads. "10" isn't exactly an even binary multiple. They've 3-platter ones too. They once had two platter ones with only three data heads (fourth was used for timing of something, I don't recall). They then just map things from there.

They paying for doing what everybody has done "forever" is IMO silly and stupid. In any case I've been having failures in Seagate China-built drives lately, so my former enthusiasm for them has subsided (and going back to WD).


RE: 5% vs. 7%
By TomZ on 10/24/2007 2:42:12 PM , Rating: 1
I'll admit HDDs could go either way, IMO. Although you're right about tracks/sectors/heads being based on physical geometry, the sector size is typically 512 bytes (never like 500). Therefore, once you multiply things out, you will get a nice even number in "KB = 1024" measure. Allocation of space in the file system data structures is also powers-of-two, since it fundamentally synchronizes with the sector size.

But I can understand that converting the total using the more marketing-friendly base-10 method is not completely invalid. The problem with this approach is that it contradicts the typical use when referring to memory and files in use by nearly all software.


RE: 5% vs. 7%
By Oregonian2 on 10/24/2007 5:36:45 PM , Rating: 2
First of all, yes the disk is probably a multiple of 512, but that doesn't matter much, it could be 11,111 of those. Not a power of two of the for the reasons you and I mention.

Memory is by power of two because of the addressing scheme in hardware IC's where another pin is added (or through multiplexing). An address is a power-of-two binary progression because of that (even though there HAVE been "3K memory chips" made in the past (by AMD I recall) and although the "K" is 1024, the total amount isn't a power-of-two amount). Chips practically speaking need to be power-of-two for reasons of easily (with little to no logic) making blocks of memory larger than one chip.

Software gets power-of-two binary memory because it uses memory that's that way in terms of "supply", but in truth it's not really binary sized either. You can (and do) malloc(sizeof struct foobar) sorts of things all the time and that struct is some random size, not power of two. Also if you look at a .EXE file, it's very likely NOT a power-of-two in size. Software uses memory in random sizes. Could be 17 bytes. Software is RARELY accessing power-of-two memory-wise (unless you want to count reading word-widths, which IMO is a silly argument). The only thing power-of-two is the total memory that's installed, and that's back to the hardware chips of memory. Go and look at the filesize of files on your PC. They're random in filelength (in windows do a "properties" of a file then look at the byte count, it'll be "random". The size listed in a explorer directory listing is either rounded up or is the file allocation size (quantized to the file system's block size). Properties should give the actual size of the file in bytes. They're probably not a power-of-two size, and probably not even a multiple of 1024.

One could, I think, make a stronger case that memory is wrong and should use power-of-ten sorts of numbers.


RE: 5% vs. 7%
By Fritzr on 10/26/2007 10:28:43 AM , Rating: 2
Been done already. Some of the early computer systems used base 10 digits instead of base 2 digits. That is a "bit" on those systems could have any of 10 different values. Just for fun try building a logic table for base 10 Boolean. Instead of not True==False, you now have absolute False, absolute True & 8 levels of partially true in between. Fuzzy logic anyone :P

The final decision to use Base 2 for computers was a hardware decision. The circuits were easier to design and build. Computers that do not use powers of 2 were quite common in the early days. The systems that were easiest to design stayed on. So the 12bit/18bit addressing went out of fashion and the 8bit DATA stuck. You will find 5,6,7,9,10,11,12 etc. bit lengths in data groups when examining the actual communications stream. Parity memory in your computer is 9 bit. 8 bits storage & 1 bit for error correction. You see 8 bits, the computer uses 9. Data transmission when sending 8bits of data often adds a "stop bit", "parity bit" and other comm protocol bits that are filtered out by the receiver. The Baudot standard uses 5 data bits plus the protocol overhead.

HDD sectors are sized in powers of 2 to allow for binary addressing of the DATA. The HDD drive has it's own internal protocols that are invisible to the user. These protocols also write data between the sectors and in some cases between the bytes in the sector. This overhead reduces the capacity of the hard disk from the user's point of view. This is why Commodore was able to format a 1.4MB floppy to store approx 1.6MB of data. The difference represents wasted space & system overhead on the IBM format. Microsoft also used a lower overhead storage that they tried to keep proprietary. There was a lot of frustration when people tried to copy those CAB files. They just wouldn't fit on the same size disk they were copied from :P


RE: 5% vs. 7%
By Fritzr on 10/30/2007 4:11:41 AM , Rating: 2
The number of bytes in a file is arbitrary. The allocated storage is File System defined. Using Microsoft file systems or any of the file systems used by Linux/Unix systems the ALLOCATED storage will always be a multiple of the disk sector size. Exceptions to this rule are files stored in container files (Zip, CAB, etc.) The allocated storage assigned to the container is a multiple of the disk sector size.

The total number of unused sectors marked as allocated is determined by the number of sectors assigned to a file system logical block. For example if the allocation unit of the File System is 4KB then the wasted storage will be 2KB*(the total number of files) The reason is that all allocations for 1 byte up to 4KB will use 4KB storage in this example. Floppy disks use 12 or 16 bit values to address File System blocks. Hard drives use 32 or 64 bit block numbers. The larger the number of bits available for a block address the smaller the File System block can be. The file system block is almost always sector_size*2^n bytes. No technical reason other than making a programmer's life easier.

Memory is measured in powers of 2. However there is no requirement that all addresses be populated with physical memory. The Commodore VIC 20 used 9K total RAM. 8K was accessed by one block of addresses, 508 bytes were also mapped into the video chip address space. Yes that is not a power of 2, it is instead the number of displayable characters. An additional 1K block of 4 bit memory was mapped to the video chip. All of these complications were design decisions.

The Commodore 64 memory map was a nightmare since you could address around 192K of address space using 16bit addressing. This was done by using multiple banks of memory space. You selected the bank & the address. The number of memory banks did not have to be a power of 2, but on the B series it was a power of two. That series also had a 64K address space, with a total of 1024K addressable. By designing hardware that does what the designer intends you can do many things to make life difficult for programmers, including, but not limited to, decimal memory addressing.


RE: 5% vs. 7%
By TomZ on 10/24/2007 8:11:12 AM , Rating: 2
quote:
First of all, HDD's have nothing with powers of two. They are completely analog storage. They even have the sector storage much bigger than the 512 bytes we see (ECC & stuff).

As you said, the storage capacity is 512 bytes = 2^9 bytes. Therefore if you multiply by total number of sectors, you get some number that is pretty conveniently expressed in powers-of two.

But the marketing people are right - HDD capacities are better expressed in powers-of-ten since that gives a perceived capacity 7% more without adding cost.


RE: 5% vs. 7%
By geddarkstorm on 10/24/2007 12:45:27 PM , Rating: 2
HDDs are digital. Either you have your magnetic alignment pointing north or south (which equals a 1 or 0, though I don't know which one is which, and different manufacturers could easily do it either way, it doesn't matter). Analogue is sine wave or anything like that, which means an infinite spectrum--that is not the case.


RE: 5% vs. 7%
By TomZ on 10/23/2007 10:07:06 PM , Rating: 3
Justin, I also think you should read this Wiki article, since I think they got it right:

http://en.wikipedia.org/wiki/Kilobyte

However, as 210 = 1024 ˜ 1000, the established K (for kilo) was early on employed as a convenient "approximate" prefix for memory capacities in multiples of 1024.

the word "kibibyte" is seldom seen in practice ... according to Google test, its usage is less than 0.3%

The overwhelming popularity of the 1024 definition for memory and file sizes means that anyone using "kilobyte" to mean 1000 in these situations is likely to cause confusion. However, it is common to use 1000 when deriving kilobyte measures from quantities which are not based on powers of two, such as bitrates.


RE: 5% vs. 7%
By TomZ on 10/23/07, Rating: -1
RE: 5% vs. 7%
By Justin Case on 10/24/2007 2:28:30 AM , Rating: 4
"Look, if you guys are going to up-rate TomZ's posts and down-rate mine, you might as well go ahead and..."

...tell the ISO and the CIPM, and everyone working in physics, mathematics and engineering that they've been using incorrect terminology for the past 50 years.

The alternative, as I explained above, is that in one particular field (or rather, sub-field), some people found it easier to just pretend that 1024 = 1000, because it saved them some work and made numbers look rounder.

That does not mean that people who can actually count have suddenly become wrong. Seagate (and all other hard drive manufacturers) have always used SI prefixes correctly. So have manufacturers of several other kinds of electronic devices.

Just because your operating system thinks that a drive with 80,000,000,000 bytes has "74.5 GB", that does not mean Seagate (or anyone else) is trying to deceive you. It just means your operating system is a bit confused about what "giga" means (as a consequence of some people starting to use the SI prefixes incorrectly 30 years ago, because it was "convenient" for them).

Much simpler than using the IEC prefixes is to simply do the maths correctly. When displaying the size of a partition (or a file), the OS should simply use the correct magnitude values. If the drive has 60,000,000,000 bytes, then display it as having 60 GB, which is 100% correct, and will make users feel much better than if they saw "55.9 GB" (which normal users assume means 55,900,000,000 bytes).

There is absolutely no reason why this nonsense of "1024 = 1000" should subsist today. Computers should present numbers in a form that's consistent and in accordance to the standards used by normal human beings , not computer programmers from 1980. If you're going to argue that the 1024 factor makes sense due to some irrelevant technical detail that lies several layers below the interface presented to users, you might as well argue that all numbers should be displayed in hexadecimal format, or perhaps binary. After all, that's how they're manipulated internally, right?

And then some people wonder why so many users complain that developers don't have a clue about the real world... sigh...


RE: 5% vs. 7%
By TomZ on 10/24/2007 8:14:17 AM , Rating: 2
quote:
...tell the ISO and the CIPM, and everyone working in physics, mathematics and engineering that they've been using incorrect terminology for the past 50 years.

I'm not saying anything negative about ISO, IEC, etc. I'm just saying that the newer IEC definitions don't have much traction in the industry.

And to say one is right and one is wrong is like saying that Metric units are wrong and Imperial units are right. They are both equally valid.

quote:
And then some people wonder why so many users complain that developers don't have a clue about the real world...

That's an ironic statement coming from you, who are completely ignoring real-world usage for half a century, instead favoring an obscure IEC publication that practically nobody uses.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 12:17:16 PM , Rating: 1
quote:
And to say one is right and one is wrong is like saying that Metric units are wrong and Imperial units are right.


Using this type of argument shows that you are either deliberately dishonest (perhaps to try to save face after your blunders like mixing up GB and GiB, and claiming that "the judge said Seagate broke the law" when the article clearly states that no court has made a decision about this case), or you are completely clueless.

Metric and Imperial units have different names (and, in any case, those are units , not magnitude prefixes).

This is about a small sub-section of the IT industry using pre-existing SI prefixes ( which already had their meaning well established in every field of science, maths and engineering, as well as other fields of IT and electronics) as an approximation. 1024 bytes can be considered as "one kilobyte" in the same way that 1024 metres will usually be described, in everyday speech, as "one kilometer".

Since it wasn't practical to manufacture or address memory in chunks of exactly 1000 bytes, chunks of 1024 bytes were used instead. And since it wasn't practical to say "one point zero two four kilobytes", people just called it "one kilobyte", which is close enough and gets the point across. Anyone working in the field knew it wasn't actually one kilobyte. It was one kilobyte plus 2.4%.

Now computers are used by a lot of normal people, who have no idea where the "1024 ? 1000" approximation comes from, but who (for the most part) do have some basic education in maths and physics. So they know the SI prefixes (from "gigahertz", "kilowatt", etc.), and they know their real meaning (x10^9, x10^3, etc.). On top of that, we are now dealing with giga and terabytes, not kilobytes anymore, so the error is no longer just 2.4%, it's closer to 10% in some cases (because it adds up with each magnitude level).

So it's up to the (few) areas of IT where the 1024 approximation still subsists to fix things. The best way to fix it would simply be to calculate the values using exact numbers, instead of approximations. That means that a 4,000,000 byte file should be listed as having exactly 4 MB, and so on. That is what is done for interface speeds, bitrates, hard drive sizes, and so on.

The IEC "binary" magnitude prefixes ("Ki", "Gi", etc.) were created just as a quick solution for programmers who are either too lazy do to the maths properly, or too concerned about backwards compatibility. So instead of calculating the final value correctly, all they need to do is write "GiB" instead of "GB", and at least they stop being wrong . It's a different matter whether or not users will have any idea what "Gibi" means, but at least they should be able to figure out it does not mean quite the same as the "Giga" they learned at highschool.


RE: 5% vs. 7%
By TomZ on 10/24/2007 12:26:10 PM , Rating: 3
Well, your personal attacks make your argument right, and I've got nearly 20 years of engineering experience in computers, electronics, and software, so I don't think I'm exactly "clueless."

I can see from your post that you are slowly conceding the argument, finally admitting that the industry has always has 1kB = 1024. That's how I learned it in school, just like everyone else with engineering and/or comp sci degrees. And that's what is used all day long by people in the same field. It's not a small minority of people, BTW.

One more point I will make before I ride into the sunset is that 1K = 1024 is not due to programmers being "lazy." I would point out that it takes actually slightly more effort to calculate KB, MB, and GB in terms of powers-of-two than powers-of-ten. Which you would know if you had any clue.

Oh, one more thing, you have yet to explain why IEC prefixes have been a total failure in the industry. Besides Linux, practically nobody uses them anywhere. If they are so "right" as you say, why does practically nobody use them?


RE: 5% vs. 7%
By TomZ on 10/24/2007 12:27:17 PM , Rating: 2
Sorry, typo: your personal attacks make your argument right -> don't make your argument right.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 2:22:59 PM , Rating: 3
quote:
I can see from your post that you are slowly conceding the argument, finally admitting that the industry has always has 1kB = 1024.


Then I guess we can add "not being able to read" to your list of problems.

Hard drive sizes have always been correctly reported, by all manufacturers. A Seagate 80 GB drive has 80,000,000,000 bytes, as does a Western Digital or Hitachi or Samsung drive.


RE: 5% vs. 7%
By TomZ on 10/24/2007 2:33:32 PM , Rating: 1
quote:
Then I guess we can add "not being able to read" to your list of problems.

Fortunately we don't have to rely on your failed interpretation of my reading abilities, since we can simply quote what you said.

Since it wasn't practical to manufacture or address memory in chunks of exactly 1000 bytes, chunks of 1024 bytes were used instead. And since it wasn't practical to say "one point zero two four kilobytes", people just called it "one kilobyte", which is close enough and gets the point across

Which, to probably everyone else besides yourself (you're in denial obviously), means that you understand the origins of the more commonly used meaning of kilobyte, specifically 1024 bytes.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 2:48:56 PM , Rating: 3
I also "understand" why people might say that the Earth's diameter is "8 thousand miles". That does not mean "one mile" is defined as 1/8000th of the Earth's diameter, or that the "8000 miles" approximation should be used when calculating larger units (such as AUs, for example).

And it definitely does not mean that people who use the units correctly should be sued and labelled as "deceptive" (which is what you and several others here are doing in regard to Seagate, which correctly reports the size of its drives, as do all other manufacturers).

You mixed up GB and GiB, drew a lot of wrong conclusions from that, and now you're doing rethorical backflips, trying to change the subject. The subject is Seagate's use of the the term "gigabyte" to mean 1 billion bytes. That usage is consistent with the rest of the industry and (more importantly), it's consistent with well-established standards for numeric and engineering terminology and with the recommendations from all major standards organizations.

P.S. - FYI, I was working with computers back when the memory of most systems could be measured in bytes (not even kilobytes), so I don't need to "understand the origins" of the term; I was there when it originated.


RE: 5% vs. 7%
By geddarkstorm on 10/24/2007 3:06:42 PM , Rating: 2
Maybe so, and Seagate probably is not in the wrong, but the courts will decide if the common perception of the consumer is that GB means this "GiB" unit (which is recent in and of itself, is it not?), and if that's so then they can get burned for it. Should they be? Probably not. But the fact the measurement exists that kilo is 1024 bytes and such and has been typically used in most people's common experience is not precluded by if it's being used in the way that is "correct" or not as seen by the SI. If the entire industry has normalized it by the SI way, then Seagate probably cannot be sued, but it's a dangerous world out there for businesses.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 3:49:46 PM , Rating: 3
But, you see, that is not the common perception of the meaning of the prefix "giga". People learn (or should learn) at school that "giga" means "multipled by one billion (x10^9)". And no one expects "one gigahertz" to mean "1073741824 hertz" (or maybe some people do, it wouldn't surprise me at this point).

Since there was no simple, short term for "1024 bytes", and since the people working in IT all knew that 1000 bytes was an impractical amount, they used to refer to 1024 bytes as "one kilobyte" (which is fine, if you're rounding to the nearest kilobyte). But they knew perfectly well that it was just an approximation, and that "kilo-" means "multiplied by 1000". Most of them were, after all, engineers, and very familiar with numeric terminology.

It's unfortunate (to say the least) that so many people extrapolate from the few cases where the term is still used incorrectly to the ones where it is used correctly (storage, communications, interfaces, processing speed, clock speed), instead of the other way around. I guess it just makes it even more amazing that civilization has managed to come this far.

Why does this happen? Personally I blame operating systems and software in general. If they had fixed the display of file and drive sizes when the error started to get significant (i.e., when we started to talk about "megabytes"), we wouldn't be in this state with some people passionately defending something that is simply wrong. Luckily, some modern software does use the correct terminology, so maybe there's still hope.

Seagate isn't just "right" according to the rest of the storage industry, it is also right according to all relevant standards organizations. It's unfortunate they decided not to go ahead with this case, since there was a chance it would educate some people. But they obviously know that giving away (useless) software will cost them less than paying their lawyers, and it's not like they would stand to profit anything from winning this.

In a way this reminds me of how some "IT journalists" started spreading the (meaningless) term "SATA II", going against the SATA-IO and the hardware manufacturers themselves (in fact, several industry members put up pages explaining why the term was meaningless, and what term(s) people should use instead).

http://www.serialata.org/namingguidelines.asp
http://www.seagate.com/ww/v/index.jsp?locale=en-US...
etc...

But try to explain this in the average 13-year-old-troll-infested internet forum and you get flamed into oblivion ("you must be a noob who doesn't know about SATA II, blah blah blah")...


RE: 5% vs. 7%
By IGoodwin on 10/24/2007 7:04:55 PM , Rating: 2
many years ago, a disctionary would have defined the word 'gay' to mean being happy. Now it has another meaning. Just becasue something has a definition, does not mean common usage cannot change it.

It does not matter how strongly you argue that saying your boss lookeds gay, mand you thought they were happy, becasue that was it's first definition. Usage changes the definition of terms, not the other way around.

As Kilo has been used to refer to 1024 bytes by the computer inductry, it has attached, no matter how strongly you feel to the contrary. Denying the use of that meaning won't make it go away.


RE: 5% vs. 7%
By Justin Case on 10/24/2007 11:05:46 PM , Rating: 3
quote:
many years ago, a disctionary would have defined the word 'gay' to mean being happy.


I did not know that the word "gay" was a mathematical or engineering term, defined by international standards organizations. Who exactly defines it? The Barbra Streisand Foundation?

quote:
Now it has another meaning. Just becasue something has a definition, does not mean common usage cannot change it.


Ignoring for a moment the fact that the prefix "kilo" is a well-defined engineering and mathematical standard term, that changing its meaning would render millions of documents useless, and that standards organizations don't make changes due to "common usage", your statement is still incorrect.

It is not "common usage" to consider that "kilo-" means "multiplied by 1024". On the contrary. Even in areas related to computers, the term is used correctly in most contexts (storage, clock frequencies, interface speeds, etc). It is only in a very restricted context (software display of file and RAM sizes) that the (incorrect) usage appears. And, even there, modern software is (finally!) starting to get it right.

Since that specific area of IT is subject to the same standards as everything else, it is that area that needs to be corrected. Using the term "kilobyte" to refer to 1024 bytes is fine in an informal context, just as it's fine to say that a 1024-watt sound system draws "one kilowatt". It is not fine when making accurate calculations, especially of very large values (where the error will add up).

Going back to the point of this article, Seagate is 100% correct in describing a drive with 80 billion bytes as an 80 GB drive (as is every other manufacturer - they all use the same terminology, in accordance with the IEC, ISO, IEEE and SI/CIPM).


RE: 5% vs. 7%
By Fritzr on 10/26/2007 11:23:41 AM , Rating: 2
quote:
Since that specific area of IT is subject to the same standards as everything else, it is that area that needs to be corrected. Using the term "kilobyte" to refer to 1024 bytes is fine in an informal context, just as it's fine to say that a 1024-watt sound system draws "one kilowatt". It is not fine when making accurate calculations, especially of very large values (where the error will add up).


1024 bytes==1 Kilobyte is standard definition and is correct
1024 watts==1 Kilowatt is your error.

If you find someone defining a Kilowatt as 1024 watts, feel free to slap them down. Electrical Engineering assigns the value of n*10^3 to the metric prefixes and has nothing whatsoever to do with computer storage measurement. The Kilowatt was defined as 1000 watts many years before the terms bit & byte were even thought of, let alone accepted as normal usage. Electrical Engineering also defines the frequency of oscillation that is referenced in CPU clock frequency. Again this is not a measure of storage.

quote:
It is only in a very restricted context (software display of file and RAM sizes) that the (incorrect) usage appears. And, even there, modern software is (finally!) starting to get it right.


Interesting. Here you state that the consistent & normal usage is the 1024 value when referring to storage, but that recent non-standard uses are showing up. Could this older (standard in 2000AD) definition that was & is considered standard possibly be the reason for the lawsuit?


RE: 5% vs. 7%
By Fritzr on 10/26/2007 10:58:21 AM , Rating: 2
If you were there when it originated you must have stayed clear of anything to do with real world applications :)

Floppy disks were measured in 1024 byte Kilobytes
Hard Disks were measured in 1024 Kilobyte Megabytes
Unthinkably large capacities that might exist were
1024 Megabyte Gigabytes
1024 Gigabyte Terabytes
1024 Terabyte Petabytes
In short for COMPUTER STORAGE CAPACITY you substitute 1024 for the 1000 that is used for Metric Measure. Computer capacity measurement is NOT metric. The prefixes are the same, the escalation in magnitude is similar, but the meaning is NOT the same.

If you'd bothered to read the literature & the advertising in the 70's and 80's you'd know that the byte had standardized at 8 bits, the metric prefixes when applied to computer storage referred to multiples of 2^10 instead of 10^3. You'd have also run across the articales discussing the size inflation in the 80s when HDD sizes started being stated in base 10 routinely. It may have been done in the early days, but it was considered fraudulent. At the time Seagate was sued it was still considered to be fraud. Without the footnote warning of the difference it is still considered fraud today.

For a few similar cases. 1 AU is the distance from the Earth to the Sun. The Earth is not 1 AU from the sun. The Fahrenheit temperature system assigns 32 degrees as the freezing point of water and 100 degrees as the normal body temperature of a human. Actual normal temperature is usually given as 98.6F. 1 inch is defined as the lengh of 3 barleycorns. You'll need to sort the grain very carefully to find ones that are 1/3 of an inch, though they will be close. Reality often disagrees with published standards :D


RE: 5% vs. 7%
By Veraiste on 10/29/2007 4:44:42 PM , Rating: 2
However, when these terms were first used, everyone understood them as the approximations they were. As these devices were commercialized, they went back to correct definition. In any case, you just reinforced his argument. The only ones using the incorrect meanings of the prefixes are software developers, likely to the historical ease and speed of bit shifting to change units.


RE: 5% vs. 7%
By Fritzr on 10/26/2007 10:39:42 AM , Rating: 2
Check the box the drive came in ... you will find a footnoted disclaimer to warn you that the drive size is using base 10 measure instead of the industry standard base 2 measure :P

Before somone in marketing decided to inflate their numbers drive manufacturers used the base 2 measure. After one company started using the base 10 measure the others followed along since buyers looked at the 80MB (base 2) and the 83MB (base 10) drives and bought the "Larger" one. It was sales hype not the ISO that made the base 10 values common. It was the use of base 10 values that caused people to take Seagate to court for falsely advertising 80*10^6 to be equal to 80*2^20. The footnote in tiny print on the box is the result of this lawsuit.


RE: 5% vs. 7%
By Justin Case on 10/29/2007 6:29:03 PM , Rating: 2
Apparently you consider that something is "the industry standard" despite the fact that no one in the industry uses it (all hard drive manufacturers consider 1 GB = 1 billion bytes) and despite the fact that all the standard organisations (IEEE, IEC, ISO, CIPM, etc) agree with that usage.

That must be a new definition of "industry standard" that means the exact opposite of what it usually means...


RE: 5% vs. 7%
By Proteusza on 10/24/2007 8:31:45 AM , Rating: 2
So ISO knows more than operating system engineers and all computer pioneers in the last 30 years?

I imagine the ISO would have lots of qualified computer scientists on board, but how many of them were involved in the ground work, where the use of the terms kilobyte, gigabyte etc originated?


RE: 5% vs. 7%
By Justin Case on 10/24/2007 3:12:37 PM , Rating: 2
quote:
So ISO knows more than operating system engineers and all computer pioneers in the last 30 years?


When it comes to defining standards? Yes. That's the whole point of the ISO (and the IEEE, and the IEC - and they all agree on this). In fact, computer-related ISO / IEEE / IEC standards were defined precisely by those "pioneers".

The SI magnitude prefixes were defined before the term "kilobyte" was coined, and the people who coined it were well aware of the fact that it was merely an approximation. No one ever defined (or tried to define) the prefix "kilo" as meaning x1024. 1024 bytes is "one kilobyte" in the same sense that 1024 (or 1025, or 998) meters is "one kilometer". Close enough for colloquial use. That does not mean accurate measurements (especially of very large quantities, where small inaccuracies can add up to a lot) shouldn't be based on the exact, standard value.

Hard drives from all manufacturers use 1 GB = 1 billion bytes. The same goes for optical media sizes, network speeds, interface speeds, CPU speeds, GPU pixel / texel / vertex processing speeds, and so on.

To turn your question around, do the people writing software know more about measuring hard drive capacity than the people making hard drives?

It's precisely to avoid having to answer such questions that there are established (ISO, IEC, etc.) standards. And, in this instance, the guys making the hard drives are the ones who got it right. It's not a matter of opinion, it's a matter of authority and precedence.


RE: 5% vs. 7%
By TomZ on 10/25/2007 8:13:10 PM , Rating: 2
quote:
It's precisely to avoid having to answer such questions that there are established (ISO, IEC, etc.) standards. And, in this instance, the guys making the hard drives are the ones who got it right. It's not a matter of opinion, it's a matter of authority and precedence.

Wrong, as you're well aware, ISO/IEC/etc. in general do not have any "authority" to impose upon software developers or their employers that they use particular standards. Companies producing software are free to adopt or not adopt the IEC prefixes depending on what they feel is in their best interest, unless there are legal or customer requirements that state that compliance is necessary.

Just because you, in your infinite wisdom, believe strongly in the IEC standards, doesn't mean that everyone else shares the same view. Other people do have a right to their own views and judgement, right Justin?


RE: 5% vs. 7%
By Justin Case on 10/29/2007 6:39:57 PM , Rating: 2
Like they said in "Life of Brian", it's your right as a man to be treated as a woman. But just because you have the "right" to bear children, that doesn't mean you can actually do it.

It's your "right" to ignore all the IEC, IEEE and ISO standards and come up with completely new terms, or assign completely different meanings to existing terms.

Just don't expect to ever get a job in any area even remotely related to engineering.


RE: 5% vs. 7%
By Fritzr on 10/30/2007 4:42:48 AM , Rating: 2
Actually in computer related engineering it would probably be fatal to pretend that the common and accepted industry standards don't exist :P

Just try convincing folks that they need to start manufacturing 1Gigabit memory chips that have exactly 10^12 bytes. Of course you can start by retraining all those engineers who insist on using the industry standard of their industry instead of the SI standard :)


RE: 5% vs. 7%
By Justin Case on 10/30/2007 9:23:24 AM , Rating: 2
Straw man argument.

There is no "standard" for "how big the memory chips must be". No one is suggesting forcing manufacturers to change the size of the drives or drive sectors (let alone memory chips, which is not even the subject of this article).

They just need to label them correctly (ie, 2^30 should be labelled as "1 GiB" or "1.07" GB).

Same thing drive manufacturers, interface card manufacturers, optical media manufacturers, network gear manufacturers, video server manufacturers, and so on already do.

Apparently you consider that something is "the industry standard" despite the fact that no one in that industry uses it (all hard drive manufacturers consider 1 GB =