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

http://www.neowin.net/news/ubuntu-implements-units-policy-will-switch-to-base-10-units-in-future-release

by 0xdeadc0de • 16 years ago

46 51 16 years ago

I don't really care whether base-10 or base-2 is used to measure data as long as it's standard for all hardware and software.

wmf• 16 years ago

It's way too late for that. Networks and hard disks have been using base 10 and software has been using base 2 for decades. IMO Ubuntu has magnified the confusion now that some software uses base 2 and some uses base 10.

barnaby• 16 years ago

I'm amazed at the depth Canonical are going into in order to make the user experience for Ubuntu as simple and stunning as possible. It's this kind of attention to detail is making Ubuntu the number one OS out there.

macco• 16 years ago

I think it is great that they are caring about the non-technical user. Why should be a kilo = 1024? It is 1000 - everybody knows this. :)

derefr5• 16 years ago

And then, when one of those non-technical users buys a "500GB hard drive", they'll actually be "getting" 500GB [according to their computer], instead of only 465.6. That oughta make everyone's lives a bit less anxietal.

AngryParsley3• 16 years ago

Compared to OS X or Windows, I'd say Ubuntu gets a lot of details wrong. The window button positioning was a recent example, but it's by no means the first. Below are some bugs that I have experienced while using 9.10 or 10.04 beta. I also noted the year they were filed.

Lots of redundant fonts: https://bugs.launchpad.net/ubuntu/+bug/18666 (2005)

Copy-pasting text is broken in some apps. https://bugs.launchpad.net/ubuntu/+bug/11334 (late 2004)

No way to auto-arrange desktop icons. https://bugs.launchpad.net/nautilus/+bug/20284 (2005, https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/1260... is related)

No ability to change individual screensaver settings. https://bugs.launchpad.net/ubuntu/+source/gnome-screensaver/... (2005)

Can't read some DVDs without using sudo/root. https://bugs.launchpad.net/ubuntu/+source/util-linux/+bug/10... (late 2004)

Setting volume to 0 doesn't mute speakers. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/16454 (2005)

Ubuntu is improving (9.10 finally fixed a bunch of problems with Intel graphics), but the user experience details are definitely a weak spot. Considering how old these bugs are, I don't expect the situation to improve any time soon.

plesn2• 16 years ago

On the contrary, I feel that they are just beginning to get interested in the user experience: boot process, themes, app installation, IM integration, scanning app, etc... I have high expectations for 10.10 on this side.

hartror• 16 years ago

Sorry but I have to strongly disagree with you there. The user experience in Ubuntu is strong and consistent. Their philosophy makes sense even for power users. Users can find the all the options they normally need and power users can dig into things to customise further.

I've never hit any of these bugs (and have been using Ubuntu full time since defecting from Fedora in '05) and most of them have very good reasons for not being fixed or are a work in progress and none of them are a user experience nightmare.

jrockway3• 16 years ago

I like how you post this comment in every Ubuntu thread. Why not sit down and fix those bugs if they bother you so much?

theschwa• 16 years ago

Out of curiosity, do you have another distro that you feel does a better job of this, or do you feel this is simply true for most distros?

Alejandro• 16 years ago

IMHO, this is the kind of attention to detail that could make Ubuntu the number one OS someday.

nfnaaron• 16 years ago

Number one distro, maybe, but no Linux distro will ever be the number one OS.

It's not Mac, for those people.

It doesn't run Outlook or MS Office, for those people. OpenOffice is not MS Office, no matter how good it is. Anything run within wine is merely a niche within a niche.

And it doesn't run whatever is the current set of MS enterprise apps and frameworks, for those people.

Linux will be a widely used server OS and a niche desktop OS, for the foreseeable future. iPads and smart phones will obliterate the desktop market before we ever see The Year of the Linux Desktop.

klodolph2• 16 years ago

Yes, yes, YES! I loathe the 1024-byte kilobyte.

phob• 16 years ago

Why?

klodolph2• 16 years ago

Because I sometimes want to do math. Lets say I want to put 6h33m of video on a 4.7 GB DVD. What bitrate do I choose? Ah, 4.7e9*8 bit / 23580 s = 1.59 Mbit/s. Now what happens if we use powers of two? I have to type a lot more into my calculator.

nova• 16 years ago

KB does exist, it's a Kelvin-byte.

axod3• 16 years ago

This is like the half metrication of the UK. You end up with 2 systems, and twice as much confusion.

Base 2 is far more logical and useful IMHO :/

Wouldn't it be so much simpler if humans had evolved to have 16 fingers instead of 10...

AngryParsley3• 16 years ago

We should use imperial units since they're more natural/intuitive/(insert rationalization here):

1 InchByte = 2.54 bytes = approximately pi bits (Edit: oops, it's about 20 bits. I divided where I should have multiplied.)

1 FootByte = 30.5 bytes

1 YardByte = 91.4 bytes

1 RodByte = 503 bytes

1 MileByte = 160,934 bytes

1 AUByte = 14.7 terabytes

1 LightYearByte = 841 petabytes

So 1 megabyte is almost exactly 2^15 footbytes. (1MB = 6.2 milebytes = 10,941 yardbytes = 32,786 footbytes)

Fun fact I learned while doing these calculations: There are 63,239 AUs in one light-year and 63,360 inches in one mile. So if you want to get a better idea of interstellar distances, imagine everything scaled down. E.G. if Earth was 1 inch from the Sun then Alpha Centauri would be 4 miles away.

axod3• 16 years ago

Using measurements appropriate to the thing you're measuring makes far more sense than trying to standardize everything into some cludgy unuseful system just so it's "standard".

FWIW, if you're basing things on imperial, you would more likely base your measurements on useful measurements that come up frequently in the domain you're measuring - file sizes etc.

For instance, you'd probably have some unit of measurement that pretty much equates to an average mp3 file of a song - say 3MB. Call it an mp3byte... etc

Then you could say "Yeah this file upload is 10mp3bytes", and people would know what you're talking about.

geocar• 16 years ago

That doesn't make any sense. You need to be using imperial volumes.

1 tablespoonbyte is approximately 36 bits (I just counted, with couscous, the universal measure), which makes a pintbyte is 1152 bytes- about a kilobyte, but also the number of pixels across on my screen.

It also gives us a gallonbyte - just shy of 10 kilobytes at 9216 bytes. That's even more convenient, because it's square root is 96.

Of course, having a hogsheadbyte at 442368- well, I think even you will find it difficult to argue the usefulness of finally being able to use a multiple of 442368.

Finally, we've got the tunbyte: 2322432. Nobody will ever need more than about 28% of a tunbyte of memory, so you see it scales out at just the most convenient point for us here.

chadaustin• 16 years ago

Maybe it's my physics background, but I never understood the "base 2 is more logical" argument. The only time I care about file sizes is "how many files can I fit on this {DVD,external drive,iPhone}?"

It's so much easier to translate GB to MB by multiplying by 1000 and vice versa.

axod3• 16 years ago

Easier if you think in base 10 I guess. Computers, and many programmers, don't though.

chanux• 16 years ago

Yeah I never really felt comfortable with this base 10 thing. Maybe because I believe base 2 representation is the way to express what it actually is.

On the other hand base 2 should've been confusing average Joe for a long time now and the operating systems that want themselves to be average Joe understandable.

rdtsc• 16 years ago

> On the other hand base 2 should've been confusing average Joe for a long time now

Base 2 became not just confusing but seemingly 'fraudulent' when it comes to buying terabyte sized storage media. A user would expect to get 1TiB but is actually getting 1TB.

   1TB  = 1000^4 = 1000000000000B
   1TiB = 1024^4 = 1099511627776B

That is quite a bit of a difference.

Am I the only computer geek who thinks that base 10 is more logical and consistent? If a byte (1B) is considered a unit, then 1KB should be 1000B, 1MB=1000^2B, 1GB=1000^3B, 1TB=1000^4B and so on?

jrockway3• 16 years ago

This is interesting, as we are finally moving away from storage that uses base-10 measurement (rotating disks). SSDs and flash drives are mostly base-2 devices.

pieter• 16 years ago

Where did you get that from? Both Intel and OCZ, probably the most popular makers of SSD's, use base-10 in their product advertising. HDD or SSD doesn't matter from a consumer perspective.

http://download.intel.com/design/flash/nand/mainstream/32229...

http://www.ocztechnology.com/products/solid_state_drives/ocz...

jrockway3• 16 years ago

The physical hardware involved, rather. You can write whatever you want on the box, and you can disable as many sectors in firmware as you want. Doesn't mean that NAND flash doesn't come in power-of-two units.

Batsu• 16 years ago

You can use whatever base you want. Realistically, end users don't even know when they're low on space anyway. Even if they do, it's a system of "deleting unused files" not finding the closest file size you can.

CrLf• 16 years ago

Snow Leopard uses base-10 units in some places, base-2 units in other places.

It isn't really confusing because when I go check how much free space I have on my hard-drive I'm just looking for a rough value, it can be base-10, base-2 or just "lots", "enough" or "almost out". The same happens to files and directories, I just want to know if they are "huge", "big" or "reasonable".

The problem is that the "old way" of reporting units in base-2 served the same purpose in exactly the same way. But this gratuitous change brings an heterogenous environment that causes problems when I want do do scripts or somesuch technical stuff where the exact amounts _do_ matter, and now have to worry in which base they are (and often this isn't obvious).

Consistency was dropped for nothing.

As for Ubuntu, I think this is a "me too" move that serves no purpose. But hey, I don't use Ubuntu, so I couldn't care less.

m0tive• 16 years ago

Surely this won't actually effect you when you want to do a script. If you want the size of a file, 'ls' (and most commands I think) gives you it in bytes, not KiB or KB or whatever...

mpk• 16 years ago

How will this affect upstream compatibility with Debian?

Raphael• 16 years ago

Shouldn't be a problem, as the internal representation of size is still bytes.

patio11• 16 years ago

I think 90% of users no longer need to know how large their files are, and the other 10% are not far off. (For that matter, I think 90% of users don't need to know that "files" exist.)

DrJokepu2• 16 years ago

I think that will pretty much mean that I'll be a Fedora user in the future. I understand that many people prefer base 10 file sizes for various reasons, I don't however so that change would drive me mad really quickly.

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

Related Stories