Forgot your password?
typodupeerror
Ubuntu Data Storage Linux News

Ubuntu Will Switch To Base-10 File Size Units In Future Release 984

Posted by Soulskill
from the stay-above-the-belt dept.
CyberDragon777 writes "Ubuntu's future 10.10 operating system is going to make a small, but contentious change to how file sizes are represented. Like most other operating systems using binary prefixes, Ubuntu currently represents 1 kB (kilobyte) as 1024 bytes (base-2). But starting with 10.10, a switch to SI prefixes (base-10) will denote 1 kB as 1000 bytes, 1 MB as 1000 kB, 1 GB as 1000 MB, and so on."
This discussion has been archived. No new comments can be posted.

Ubuntu Will Switch To Base-10 File Size Units In Future Release

Comments Filter:
  • Interesting (Score:3, Interesting)

    by LBArrettAnderson (655246) on Saturday March 27, 2010 @10:59AM (#31639958)

    I find it interesting that operating systems are headed in this direction, while SSDs are becoming more and more popular, and which (for the most part) use base 2 measurements.

    It looks like both Apple and Ubuntu are trying to get consumers to think that they use less disk space.

  • by Kreychek (264929) on Saturday March 27, 2010 @11:04AM (#31640018)
    I thought the reputable mfgs had already jumped on the bandwagon where they use kibi-/mebi-/etc prefixes to denotes powers of 2? See IEEE 1541 [wikipedia.org]. Following this standard, the change makes sense. Either that or they should have switched to the binary prefixes.
  • Really annoying (Score:5, Interesting)

    by Ma´djeurtam (101190) on Saturday March 27, 2010 @11:12AM (#31640084) Homepage Journal

    I work mostly on OS X and this so-called feature annoys me to no end. I do not know the size of my files anymore, I have to go to the terminal just to know the size of a file (bash hasn't been polluted by this feature).

    I've been using computers for 20+ years and I do _not_ want to change how I think file sizes, especially since I feel that base 10 is the wrong way to count. What's next? Imperial units for us Europeans?

    The most annoying? That nobody has hacked Snow Leopard to restore real units.

  • by hitmark (640295) on Saturday March 27, 2010 @11:24AM (#31640214) Journal

    err, kilo have been a SI prefix since 1795. And mega, giga and tera got defined in 1960.

  • by DMUTPeregrine (612791) on Saturday March 27, 2010 @11:36AM (#31640358) Journal
    We have kibibytes already. They're now using those for the base 2 measurements, instead of kilo.
  • Re:Really annoying (Score:3, Interesting)

    by ceoyoyo (59147) on Saturday March 27, 2010 @11:39AM (#31640388)

    You know, that's the same argument Americans use for not adopting metric.

  • Re:Why? (Score:3, Interesting)

    by Waffle Iron (339739) on Saturday March 27, 2010 @11:41AM (#31640410)

    Why?

    Because any calculation involving a mixture of different prefixes is pointlessly difficult when using pseudo base-2. For example, your file manager reports that a memory stick has 1.20GB free space, and it reports that you have two 613MB files. Will they fit on the disk? If the apps use stupid pseudo base-2 numbers, you have to dig around for a calculator. If instead the apps use base 10 like you've been trained to use since early childhood, the answer is obvious.

    Moreover, contrary to the common misperception, the only large physical quantity in a computer that is usually fixed to a power of 2 is RAM (and virtual memory makes the exact size of RAM totally irrelevant to users). Disk drive sizes, substructures on disk drives, network bandwidth, optical media, filesystems, individual file sizes, bus bandwidth, clock speeds, network packets, etc have no physical relationship to the number 2. (Some filesystems use 512-byte block sizes. So what? That implementation detail is well hiddenfrom end users, and notice that 512 does not match any of the binary prefix sizes either.) Why would you invent an entire new confusing, hard to use pseudo number system (whose "base" idiotically varies depending on the order of magnitude) when the quantities you're measuring don't even have any relationship to it?

  • by gweihir (88907) on Saturday March 27, 2010 @11:46AM (#31640454)

    And even in the metrically challenged USA, SI happens to be the law. Imperial units are only allowed in addition, but prefixes must be SI. If you don;t believe me, look up your own laws before shooting off your mouths.

    The rule actually is that anything measured must use SI prefixes and units (in the US and some other backward countries some historic units may be allowable besides SI units, but prefixes are the law even there) when sold, i.e. the HDD manufacturers are prohibited by local and international law from using base 2 units as the only or main size statement. If they do, that would be fraudulent. The only thing that would save them is that it is permissible to give the customer more than stated.

    So how do RAM sizes come into this? Simple: A RAM size is not a measurement. It is membership in a size class. While a HDD can have an arbitrary size (well, modulo 512, but that is a detail with todays sizes), RAM cannot have other sizes than powers of 2 and hence a statement like 1MB for RAM is a statement of membership in a specific size class and not a measurement. (Incidentially, 1mb is 1 mili-bit, i.e. 1/1000 of a bit. Get this right or be regarded as a moron!)

    I do not understand why so many people cling to a mistake. Grow up!

  • by growse (928427) on Saturday March 27, 2010 @12:11PM (#31640720) Homepage
    It's not a standard in the computer world though. How many bits per second does your gigabit network carry?
  • by Island Admin (1562905) on Saturday March 27, 2010 @12:27PM (#31640872)
    Unfortunately, software is base-2. Hence now we are going to have the following interesting things in documentation:

    Max File Size: 2.199023255552 TB .... As the storage will always be in words, double words, etc. which are based on base-2.
    It does not make sense to me to use KiB and kB, etc as it will not reduce the confusion with the average user.
  • by beelsebob (529313) on Saturday March 27, 2010 @01:05PM (#31641232)

    It wasn't that it made sense, it was that on an old CPU, dividing by 1000 was a ~600 cycle operation – there wasn't an op on the CPU to do it, you had to emulate it in software. Dividing by 1024 meanwhile was a 1 cycle operation - right shift 10 bits.

    There was no "it makes sense to do this wrong", instead it was "this is wrong, but it's so much more efficient, that at the moment, we have to be inaccurate". I'm sure if the people who originally did it realised the quagmire we'd be in these days they'd have never considered getting it wrong.

  • by A nonymous Coward (7548) on Saturday March 27, 2010 @01:41PM (#31641496)

    Anything dealing with bits has always been base-10.

    Really? Memory chips came in decimal units?

    I thought I'd left that behind with the IBM-1620 (20,000 digits) and Univac SS-90 (5000 words).

  • Re:Annoying... (Score:2, Interesting)

    by Anonymous Coward on Saturday March 27, 2010 @02:04PM (#31641688)

    > The 56k modem is 56,000 bits
    Bad example. The 56k modem is really 57600bps. Modem speeds are all multiples of 2400bps: 9600bps, 14.4kbps, 19.2kbps, 28.8kbps, 33.6kbps

  • Re:But its WRONG! (Score:4, Interesting)

    by Smallpond (221300) on Saturday March 27, 2010 @02:07PM (#31641728) Homepage Journal

    In computers, there are critically important reasons why base2 is used (transistors act as electronic switches which have two states: on and off). Wiring buses in computers (address lines, data lines) are base2, with the electricity being on or off. When data is read and written on the hard disk, its done in a binary format. When searching for data on a disk, its either a binary search, or the equivalent of a binary search (divide and conquer) algorithm used to find data, with O(log n) search time. Please don't dumb it down. We don't need to dumb it down. Its not that the computer is to compl-cated, its that the user is too dumb. Users have been suing drive manufacturers who have been lying to people like this. I expect data to be stored this way. Two wrongs don't make a right (it takes three lefts for that), but I digress, don't change it. DON'T!

    FIrst off, young sprat, I've used decimal hardware. So are you going to force me to say 1KB is 1024 even though it makes no sense on that system?

    Second, I've used disk drives for other purposes than computers - data logging using dedicated hardware. Why should I be tied to your computer idiosyncrasies? Mass storage was around before computers settled on binary. It has always used decimal units. You are making a common mistake - assuming that because something is some way now, that it was always that way.

    Third, if base two is so critical to computers, then how many Hertz are in 1 MHz?

    And finally, I'm now working on ternary logic. Storing information as 3^N is already inside some devices you're using, so maybe binary is not so critical after all.

  • by Hal_Porter (817932) on Saturday March 27, 2010 @02:08PM (#31641738)

    Whenever I see the word mebibyte I always hear it in Scots accent. "Don't touch the doggie, it mebibytes".

    This sort of think is useful though, it sorts out the people who learned their computer knowledge from wikipedia thirty minutes ago from the ones who learned it from Byte and Doctor Dobb's Journal thirty years ago.

  • by jolyonr (560227) on Saturday March 27, 2010 @02:34PM (#31641896) Homepage

    Finally they've seen sense and will switch to the metric ten-bit byte.

    I'm fed up of this short-changing every time I use a byte! I'm sure the memory companies will harp on about how it will increase costs by 25% but they've had an easy life for years.

    If God had meant us to program with 8 bits per byte, he'd have given us 8 fingers.

  • by fbjon (692006) on Saturday March 27, 2010 @03:17PM (#31642246) Homepage Journal
    What do you mean never? "Kilo" has always meant 10^3 for HDDs, likewise for mega, giga, etc.

    Moreover, why would it make any more sense to use KiB than KB for things like file sizes other than purely historical reasons? A file is a sequence of bytes of arbitrary length with no grouping whatsoever. It's just a sequence. I wouldn't mind using SI units when dealing with hardware that is already specified in SI units to begin with.

  • by houstonbofh (602064) on Saturday March 27, 2010 @06:04PM (#31643368)
    Talk about revisionist history... The lawsuits over this marketing change are STILL being settled. This whole si MiB crap was a bad attempt to clarify when the numbers were non standard computer usage for marketing reasons. Of course I am old enough to remember doing binary math on my fingers for an algebra test when calculators were not allowed. Really... No calculators. You actually had to understand math. No, really!
  • by Bigjeff5 (1143585) on Saturday March 27, 2010 @07:03PM (#31643720)

    Well, it depends on what you are talking about. The situation is not as clear cut as you depict it.
    1 kb on your disk is usually defined as 1024 bits... but 1 kb/s is usually defined as 1000 bits/second. As an example, a 1.5 Gb/s SATA interface is running with a 1.5 GHz clock, so it will transfer 1500000 bits per second (actually, the number of effective bits will be lower as it uses 8b/10b balancing).

    That's false.

    1 kb/s is defined as 1024 bytes per second. 1 kbps is 1000 bits per second, as is 1 kbit/s. If someone has sold you a 1.5 gb/s SATA interface, they have robbed you blind, as the fastest SATA interfaces in development only run at 6 gbit/s, which is about 40% the speed of your 1.5gb/s interface.

    This [wikipedia.org] might help you out a little bit.

  • by FrozenGeek (1219968) on Saturday March 27, 2010 @07:21PM (#31643784)
    Well, when I was in university, studying computer science (who'd've guessed that a /. contributor studied comp sci?), KB was defined as 2^10 bytes, MB was 2^20, and GB was a pipe dream (hey, I graduated from uni in 1986). So, for us, kB WAS defined as 1024 bytes. That's how ALL of my textbooks, which I still have btw, defined kB.
    Perhaps it was not defined as 1024 bytes everywhere (comp sci types are notorious for having multiple standards), but it was defined as 1024 bytes in a fair number of places.
  • by Glonoinha (587375) on Saturday March 27, 2010 @07:28PM (#31643828) Journal

    Actually 1024 ohms definitely is a perfect 1 kilo ohm resistor, worthy of a gold fourth stripe.
    Citation [wikipedia.org]

  • by Anonymous Coward on Sunday March 28, 2010 @04:43AM (#31646160)

    Only if they're completely in contradiction with each other.

    Cleave.

"Everything should be made as simple as possible, but not simpler." -- Albert Einstein

Working...