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.
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.
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.
Then the "non-technical" user who complains about this can go fuck themselves, since the Windows machine they're supposed to be switching from does the same thing.
At first it sounds like I'm trolling, but please think this issue through. What crowd is Ubuntu trying to pander to? Are there really people out there "technical" enough to bother caring about a hard drive's size, but not technical enough to figure out that 1000 != 1024?
If you want something that the end user will really care about, ship Ubuntu with a default theme that doesn't look horrible, or include useful codecs and plugins in the default install. "Oh, but those are easy to obtain", you say? So then is it really that much more difficult to spend 5 seconds explaining that it's 1024, not 1000? Not that I care enough to create some stupid "UBUNTU + FLASH LOL" distro, since I use Arch or one of the BSDs, but it just seems stupid that they'd make a silly, superficial change like this and expect a pat on the back.
> What crowd is Ubuntu trying to pander to? Are there really people out there "technical" enough to bother caring about a hard drive's size, but not technical enough to figure out that 1000 != 1024?
They don't care about the hard drive size per se, they just care that all the numbers the computer reports are as impressive as the ones the salesman told them.
> So then is it really that much more difficult to spend 5 seconds explaining that it's 1024, not 1000?
It is when you're being yelled at in the product returns line. It sounds like an excuse and an ass-pull, and it's not something that would calm down a belligerent customer.
Remember that there are two types of people who have Ubuntu—those that got it themselves, and those who had it dropped upon them by a more technical user. For the second group, "superficial" change is the only kind of change they care about, vis. http://www.makemylogobiggercream.com/
Also, as a side-note, the cost of being widely-used by governments and public institutions, and being open-source, is that you have to obey the law, so the option to "include useful codecs and plugins in the default install" just isn't available. (You can't purchase a license for a codec when you literally don't know how many users you have.)
So, since HD manufacturers use inflated numbers the solution is for everyone else to change?
Does this mean we're going to stop using base 2 for things like network speeds as well? What will users think when they have lets say a photo management program, that says that their library is 495 GB (base 2) and then they try to transfer that over to a disk that has 500 gb (base 10) available, only to find 495 doesn't fit in 500. That's going to be pretty confusing.
The fact is the ONLY place I've ever seen base 10 numbers for storage has been on storage media to fool people about filesize. The programs users use have standardized on base 2, and this is just going to be confusing.
I actually don't find this pragmatic at all. The current scenario is that storage numbers are consistent in software (base 2) and consistent in hardware (base 10).
The end result of this is that storage numbers will be inconsistent in software, and consistent in hardware. How is that an improvement?
I don't know where you're getting the "consistent in software" bit. OSX already switched to base 10 seven months ago with 10.6—this is, like with the window caption buttons thing, probably where Ubuntu got the idea. Now the only one doing it differently (that non-technical users are aware of) is Windows—and I'm guessing they'll be right along in W8.
Wow, I had to idea OSX did that, I only use Linux and Windows. Well, I guess things aren't as constant as I thought they were.
I still think the overwhelming majority of software uses base 2, and that base 2 for that reason alone should be the standard. Neither standard is really better than the other, the more popular one should win.
Yes and no, network speeds are base 10 when you buy it and when you spec hardware. Most users interaction with quantifying network speeds, however, is the browser download progress bar, which always uses base 2.
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.
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.
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.
Well, the user experience is more or less the point of Ubuntu, so I'd say they've been working on it since 2004. They've made a lot of progress, certainly, but it's not like they've just turned their attention to these things...
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.
Resolution: WONTFIX
"I don't have any plans to support this. My view is that any screensaver theme that requires configuration is inherently broken."
or are a work in progress
In progress for five years?
and none of them are a user experience nightmare.
I wasn't talking about nightmares, just minor things that are broken. You want nightmares? Try:
2.6.28-11 causes massive data corruption on 64 bit installations https://bugs.launchpad.net/ubuntu/+source/linux/+bug/346691 (March 2009) This one is particularly bad because they went ahead and released Jaunty with this bug. Then, the fixed kernel introduced a new bug that caused FS corruption on systems with Intel disk controllers.
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.
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.
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.
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.
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.
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.
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.
> 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.
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?
Also, the labels are fairly clear; I don't see too much of a problem for smart people to unlearn a bad habit. How many people here think that a kilometer is 1024 meters?
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.
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.
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.
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.
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.
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...
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.)
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.
51 comments
[ 2.9 ms ] story [ 112 ms ] threadAt first it sounds like I'm trolling, but please think this issue through. What crowd is Ubuntu trying to pander to? Are there really people out there "technical" enough to bother caring about a hard drive's size, but not technical enough to figure out that 1000 != 1024?
If you want something that the end user will really care about, ship Ubuntu with a default theme that doesn't look horrible, or include useful codecs and plugins in the default install. "Oh, but those are easy to obtain", you say? So then is it really that much more difficult to spend 5 seconds explaining that it's 1024, not 1000? Not that I care enough to create some stupid "UBUNTU + FLASH LOL" distro, since I use Arch or one of the BSDs, but it just seems stupid that they'd make a silly, superficial change like this and expect a pat on the back.
They don't care about the hard drive size per se, they just care that all the numbers the computer reports are as impressive as the ones the salesman told them.
> So then is it really that much more difficult to spend 5 seconds explaining that it's 1024, not 1000?
It is when you're being yelled at in the product returns line. It sounds like an excuse and an ass-pull, and it's not something that would calm down a belligerent customer.
Remember that there are two types of people who have Ubuntu—those that got it themselves, and those who had it dropped upon them by a more technical user. For the second group, "superficial" change is the only kind of change they care about, vis. http://www.makemylogobiggercream.com/
Also, as a side-note, the cost of being widely-used by governments and public institutions, and being open-source, is that you have to obey the law, so the option to "include useful codecs and plugins in the default install" just isn't available. (You can't purchase a license for a codec when you literally don't know how many users you have.)
So this person is yelling about something they know nothing about? I'm sorry, but that sounds like a sever case of "stupid" to me.
>http://www.makemylogobiggercream.com/
Funny video!
Does this mean we're going to stop using base 2 for things like network speeds as well? What will users think when they have lets say a photo management program, that says that their library is 495 GB (base 2) and then they try to transfer that over to a disk that has 500 gb (base 10) available, only to find 495 doesn't fit in 500. That's going to be pretty confusing.
The fact is the ONLY place I've ever seen base 10 numbers for storage has been on storage media to fool people about filesize. The programs users use have standardized on base 2, and this is just going to be confusing.
The end result of this is that storage numbers will be inconsistent in software, and consistent in hardware. How is that an improvement?
I still think the overwhelming majority of software uses base 2, and that base 2 for that reason alone should be the standard. Neither standard is really better than the other, the more popular one should win.
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.
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.
Lucky for you. Unfortunately, I and many others have.
most of them have very good reasons for not being fixed
Some of them have inane reasons for not being fixed. From https://bugzilla.gnome.org/show_bug.cgi?id=316654
Resolution: WONTFIX "I don't have any plans to support this. My view is that any screensaver theme that requires configuration is inherently broken."
or are a work in progress
In progress for five years?
and none of them are a user experience nightmare.
I wasn't talking about nightmares, just minor things that are broken. You want nightmares? Try:
2.6.28-11 causes massive data corruption on 64 bit installations https://bugs.launchpad.net/ubuntu/+source/linux/+bug/346691 (March 2009) This one is particularly bad because they went ahead and released Jaunty with this bug. Then, the fixed kernel introduced a new bug that caused FS corruption on systems with Intel disk controllers.
ehci_hcd module causes I/O errors in USB 2.0 devices https://bugs.launchpad.net/ubuntu/+source/linux/+bug/88746 (2007)
ath5k driver connection and performance issues https://bugs.launchpad.net/ubuntu/+source/linux/+bug/461419 (2009)
[karmic regression] all network apps / browsers suffer from multi-second delays by default due to IPv6 DNS lookups https://bugs.launchpad.net/ubuntu/+source/eglibc/+bug/417757
I've been hit by all of these except the first one, and I was only saved from that because I stumbled on the bug report before upgrading to Jaunty.
It is very hard to understand how to pander to such people - but Windows is, unfortunately, a good starting point....
There are lots of problems with the ux on ubuntu that makes it problematic for a lot of people.
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.
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...
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.
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.
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.
It's so much easier to translate GB to MB by multiplying by 1000 and vice versa.
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.
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.
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?
http://download.intel.com/design/flash/nand/mainstream/32229...
http://www.ocztechnology.com/products/solid_state_drives/ocz...
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.