Another controversial change for Ubuntu 10.04: File size policy


Recommended Posts

Not something that will largely affect me. Frankly I dont care. With harddrives in the TB range and probably larger coming in the next few years, should we really complain over a few bytes (giga, kilo, mega, whatever)?

It certainly causes a problem when trying to perform any calculations, especially when you get into the TiB/TB range, the difference is 99,511,627,776 bytes!

Disappointingly even Google gets this one wrong: http://www.google.co.uk/#q=MB+in+kB

I doubt if most users will either know or care about this.

I'm sure if you've dealt with people with no computer background (which btw, is what I'd define to be "most users"), you'll know that we are kinda tired being asked the question "Why is my 1TB hard disk only showing 931 jiggabyte? Should I ask for a refund?".

Why change something after so many years? People need to learn computers work on the base2 structure and not base10.

And all you fighting here need to learn philosophy, and realise that this debate won't achieve anything of use :p

Never used 'KiB' ever and never will don't recognise it because a KB is 1024 bytes etc. Base 10 units are an approximation and not a formal unit in themselves - it's purpose is for apples advertising of how many songs, video and picture your ipod can hold. Occasionally sure it can also be used for people guessing sums in their head but that's it.

Never used 'KiB' ever and never will don't recognise it because a KB is 1024 bytes etc. Base 10 units are an approximation and not a formal unit in themselves - it's purpose is for apples advertising of how many songs, video and picture your ipod can hold. Occasionally sure it can also be used for people guessing sums in their head but that's it.

Huh? That made little to no sense.

You don't recognise KiB, therefore it's wrong? Base 10 units are an approximation? It's used for sums in your head?

The computer world has misused the 1,000 units system for a while - but just because they've done it for some time, doesn't mean it's right. Kilo means 1000, not 1024. This is why "KiB" was created, meaning 1,024 bytes, so that it could have proper terminology without misusing the metric system. (It has bugger all to do with Apple's advertising. WTF?)

Ideally, all systems should sort out their units, and display them correctly. That way users wouldn't get confused and it doesn't necessarily prevent either one being used. Yes, some re-education might be required ... but it's not like people base their life around whether 1 kilobyte is 1024 or 1000 bytes.

Read me: https://wiki.ubuntu.com/UnitsPolicy

Huh? That made little to no sense.

You don't recognise KiB, therefore it's wrong? Base 10 units are an approximation? It's used for sums in your head?

The computer world has misused the 1,000 units system for a while - but just because they've done it for some time, doesn't mean it's right. Kilo means 1000, not 1024. This is why "KiB" was created, meaning 1,024 bytes, so that it could have proper terminology without misusing the metric system. (It has bugger all to do with Apple's advertising. WTF?)

Kilo means a thousand, yes. byte is generally considered 8bits so 2?/2*2*2 = 8 = 1B and 2?⁰ = 1,024 = 1KB. So as it's denoted kilo as the unit prefix 1|xxx with x=byte/bits I never said it was wrong I just don't recognise or use it because the fact is it's an unnecessary term devised for god knows why for something which is as simple as remembering ones tens hundreds thousands in money.

the only differentiation imo necessary is Byte and bit which is simple lower case b = bits B = bytes which is generally accepted.

Kilo means a thousand, yes. byte is generally considered 8bits so 2³/2*2*2 = 8 = 1B and 2¹⁰ = 1,024 = 1KB. So as it's denoted kilo as the unit prefix 1|xxx with x=byte/bits I never said it was wrong I just don't recognise or use it because the fact is it's an unnecessary term devised for god knows why for something which is as simple as remembering ones tens hundreds thousands in money.

the only differentiation imo necessary is Byte and bit which is simple lower case b = bits B = bytes which is generally accepted.

Regardless of how many bits are in a byte ... 1000 bytes is still 1000 bytes, and thus 8000 bits. Kilo = one thousand. How is 2^10 related to the word "kilo"?

Basically, I still fail to see your point. Kilometer = 1000 metres. Kilogram = 1000 grams. These are absolute standards. Computing is the only one that gets it wrong. I've no idea why you think it's a "unnecessary term". What exactly is unnecessary?

The issue currently is kilo being used for two different things. This change is about fixing that.

Regardless of how many bits are in a byte ... 1000 bytes is still 1000 bytes, and 8000 bits. Kilo = one thousand.

Basically, I still fail to see your point. Kilometer = 1000 metres. Kilogram = 1000 grams. These are absolute standards. Computing is the only one that gets it wrong. I've no idea why you think it's a "unnecessary term". What exactly is unnecessary?

The issue currently is kilo being used for two different things. This change is about fixing that.

Ok i'll make it even more simpler. We all know decimal system and stuff. 1.00 is one dollar because it's in first ones after decimal point. 10.00 is tens because now the 1 is in the tens after the decimal. 100.00 now the 1 is in the hundreds and so on. 103.00 This is One hundred and three as there's 1 hundred and 3 ones in front of the decimal, ok.

Now as you move up in units of base units of 10 you can attach SI prefix's kilo mega giga etc metric units are base 10 so it's common usage with them which is what causes this problem but the point is kilo is 1000 yes as I said a KB is 1024 Bytes [024] which is the Byte section in the name so Kilo-Thousand [indicating thousandth digits] + Byte [2¹⁰] = 1024Bytes = 1KB

Now as you move up in units of base units of 10 you can attach SI prefix's kilo mega giga etc metric units are base 10 so it's common usage with them which is what causes this problem but the point is kilo is 1000 yes as I said a KB is 1024 Bytes [024] which is the Byte section in the name so Kilo-Thousand [indicating thousandth digits] + Byte [2¹⁰] = 1024Bytes = 1KB

The point is that a kilobyte is bad terminology for 1024. You can't use it as the basis for your reasoning.

A kilo means one thousand; a thousand bytes is 1000 bytes, not 1024. You're seriously twisting maths to make your thinking work.

If I read you correctly, let's say we have base 3 system using foobars. 3^7 is 2187 foobars. That's the closest we get to 1000 in that system using simple power jumps. So I'm going to define 2187 as a kilofoobar. It doesn't make sense.

Base 2 is just lucky that 1024 lands so close to 1000.

2^20 is 1,048,675 (1 MiB). This is 1024x1024 ... it's still not relating to 1000. The problem just gets worse as you get bigger. 1,073,741,824 (1 GiB) is plainly not a billion of anything.

Basically, a kilobyte should be one thousand bytes. We have a the terminology available to deal with the computing world: a kibibyte. It's not hard to use that instead if needed.

This, like many other issues, will never be resolved to the satisfaction of everyone. Those who oppose the use of "kilo", "mega", etc. to mean anything other than various multiples of 10 will always bitch about it. Those who have used computers for decades before this change will always bitch about it being changed. Kind of like how I personally think the French are annoying to the rest of the world by insisting on calling a byte an "octet", so they have their own abbreviations - Ko, Mo, Go, etc.

This, like many other issues, will never be resolved to the satisfaction of everyone. Those who oppose the use of "kilo", "mega", etc. to mean anything other than various multiples of 10 will always bitch about it. Those who have used computers for decades before this change will always bitch about it being changed. Kind of like how I personally think the French are annoying to the rest of the world by insisting on calling a byte an "octet", so they have their own abbreviations - Ko, Mo, Go, etc.

That can more reasonably be put down to translation differences - it's not changing the definition or anything. It's something used in English anyway: http://en.wikipedia.org/wiki/Octet_%28computing%29

Of course it can be resolved, it's just the computing guys will ultimately "lose". :)

Of course it can be resolved, it's just the computing guys will ultimately "lose". :)

That's why I said it will never be resolved to the satisfaction of everyone. Sort of like the demotion of Pluto. There will always be those who continue to call it a planet because that is what they grew up calling it.

Not something that will largely affect me. Frankly I dont care. With harddrives in the TB range and probably larger coming in the next few years, should we really complain over a few bytes (giga, kilo, mega, whatever)?

in a way yes, you wont see much space difference form your old 20GB drive, but in the 1TB drives you are 'losing' around 70GB. As bigger drives come to the market a lot of people are going to wonder what happened to their hundreds of gigs.

Good to see them finally do this, the kernel and most userland tools have been using the proper suffixes for ages.

It also brings it in line with HD makers (who have always done it properly), and helps the end user (having conflicting numbers = bad)

Exactly. Let mega mean million.

http://en.wikipedia.org/wiki/Mega-

Confirmed in 1960, it comes from the Greek μέγα.

If you want 1048576 then DON'T call it mega but call it something else = Mebi as IEC suggested.

End of confusion and discussion.

  • Like 2

Seeing some of the comments on Slashdot reminded me, that not every computing related subject uses the whole "kilo = 1,024" idea.

Networking for example uses the standard meanings, 1KBps means 1,000 bytes per second, not 1,024 bytes per second.

Isn't that kinda strange, that a network running at 1GBps, can't download "1GB" of data in a second?

Edit: And yet another one I forgot, speed. A 2Ghz CPU doesn't run at 2,147,483,648 Hertz.

Seeing some of the comments on Slashdot reminded me, that not every computing related subject uses the whole "kilo = 1,024" idea.

Networking for example uses the standard meanings, 1KBps means 1,000 bytes per second, not 1,024 bytes per second.

Isn't that kinda strange, that a network running at 1GBps, can't download "1GB" of data in a second?

Edit: And yet another one I forgot, speed. A 2Ghz CPU doesn't run at 2,147,483,648 Hertz.

Your capitalization changes the meaning. Network speeds use 1Gbps which means 1 Giga-bits per seconds. The bit is the lowest unit so 1,000 bits is 1,000 bits. A byte is 8 bits and from then is calculated in Base-2 math. So if you want to convert 1Gbps to 1GBps, you dvide it by 8.

I have never seen this as a major problem for average joes. They simply do not care most of the time.

Your capitalization changes the meaning. Network speeds use 1Gbps which means 1 Giga-bits per seconds. The bit is the lowest unit so 1,000 bits is 1,000 bits. A byte is 8 bits and from then is calculated in Base-2 math. So if you want to convert 1Gbps to 1GBps, you dvide it by 8.

I have never seen this as a major problem for average joes. They simply do not care most of the time.

Uh, that's got pretty much nothing to do with what he was saying :)

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • I sort of agree with you on that. I had a telescope - a real hefty thing, although only around 500CHF - that got me so fascinated about the stars and planets. I would stare for hours, amazed that I could see the craters of the moon or the rings of Saturn in "realtime" (quotations because, y'know, speed of light and whatnot). A friend of mine has a telescope like the one mentioned in the review, and the pictures are amazing to look at. But there is something missing for me. I may as well just go to NASA's website and look through their gallery.
    • As opposed to catching bad press because it's the engine's fault? You didn't really think this through did you? lol. The engine has all the tools a dev needs to optimize their game, Epic isn't gonna hold their hand to make sure they use them. Also, Epic isn't forcing anyone to use Lumen, Nanite or super high resolution meshes and textures. Just because Nanite can render meshes with millions of polygons doesn't mean you should use it like that, in fact, Epic recommends you avoid doing that. Most of the stuttering can easily be fixed by cutting down on unique shaders by using master materials and generating a PSO cache and compiling every shader before loading the game and managing what must be loaded and unloaded at runtime, it's up to the developer to set all that up properly.
    • I think it depends on what you're looking for to do, and the time you have to spare. With my Dwarf 3, I easily spend 3-4 hour sessions; half an hour driving to an un-light polluted place, another half hour unpacking and setting up the smart scope + tripod for equatorial tracking, then 15 more minutes mucking around with settings and shooting calibration frames, spending a few hours shooting, merging with past photo sessions, etc. It's crazy how time flies and I often get home later than I expected. It's something I still need to set aside a good part of an evening to do, all in all. For one session, where you often need like four for best results when it comes to deep space objects. Even with a smart scope like Dwarf 3, regular non-astro photography is still way more approachable to people getting into photography. I find this is a time consuming niche no matter how I go about it. With practice, I can probably begin cutting time here but I think where smart scopes find their home is among people who love to shoot the night sky but don't have the spare time to go deep with the "navigator level" attunement to the night sky itself in addition to everything else. Having said this, _if_ you have even more time to spend on this hobby, it will probably be even more rewarding to do it more by hand and learn the skies and the details of how it all works.
    • I misread the title and thought Teams itself would be redesigned. Imagine having this one as a native WinUI app.
    • Dell, HP PCs ran into endless reboot, BitLocker recovery loops but Windows 11 isn't to blame by Sayan Sen Last month Neowin reported on a major issue on Dell systems wherein a bug in its official support tool was leading to endless blue screen of death (BSOD) and restarts. Following our report, Dell officially acknowledged its SupportAssist-related crash issue, confirming that the culprit is not Microsoft's operating system but rather a faulty version of its own remediation software. In a newly published support advisory, Dell stated that version 5.5.16.0 of Dell SupportAssist Remediation and Alienware SupportAssist Remediation can trigger blue screen errors and unexpected system restarts. The company notes that the problematic component operates independently of the main SupportAssist application, meaning users should not remove the primary SupportAssist software when troubleshooting the issue. According to Dell, the crashes are linked specifically to the SupportAssist Remediation service, which is bundled with SupportAssist OS Recovery Tools, and as such it has since released an updated version, 5.5.16.1, which is said to resolve the problem. Affected users are advised to first verify whether version 5.5.16.0 is installed by checking the Installed Apps section in Windows Settings. If so, Dell recommends updating SupportAssist OS Recovery Tools through either SupportAssist's "Update Software" feature or Dell Command Update. Dell also advises users to back up important data before performing the update and to ensure systems remain connected to power throughout the installation process. If you are still having issues though make sure to report to the Dell support forum. As it turns out though Dell is not the only PC maker currently dealing with update-related headaches as HP is also facing a separate but probably equally frustrating issue involving recent Windows Secure Boot updates that were released with recent Windows 11 Patch Tuesdays. Similar to Dell, HP also put up its own support article where it explains the issue. The company says that affected devices could hit a brick wall when booting as they run into a BitLocker recovery loop after the April 2026 updates. The problem appears to affect systems wherein the new UEFI Secure Boot CA 2023 certificates fail to apply properly. As such affected users will find themselves entering their recovery key over and over again despite the system otherwise functioning normally. HP says such PCs should be updated to the latest available BIOS version and configured with the necessary Secure Boot certificates before installing Microsoft's Windows 11 Patch Tuesday updates. Systems that are already experiencing the problem may require BIOS configuration changes to restore normal boot behavior. Admins can find information regarding that in the support article here on HP's official website.
  • Recent Achievements

    • One Month Later
      DJC50PLUS earned a badge
      One Month Later
    • Week One Done
      DJC50PLUS earned a badge
      Week One Done
    • Proficient
      Eric Biran went up a rank
      Proficient
    • Dedicated
      Conjor earned a badge
      Dedicated
    • Week One Done
      Windows Guy earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      493
    2. 2
      PsYcHoKiLLa
      246
    3. 3
      Steven P.
      72
    4. 4
      +Edouard
      69
    5. 5
      neufuse
      68
  • Tell a friend

    Love Neowin? Tell a friend!