Slashdot Mirror
Linux Kernel 2.6.6 Released
maradong writes "The new Linux Kernel 2.6.6 has been released just 2 hours ago. The Patch from version 2.6.5 to 2.6.6, which can be downloaded on kernel.org measures 2.4MiB and the Changelog can be found at the known place."
Wow, this is the first time I see a Slashdot news using the right unit, i.e. MiB = Mebibytes = 2^20 bytes = 1 048 576 bytes, as opposed to MB = Megabytes = 10^6 bytes = 1 000 000 bytes.
Congrats !
theefer
Note that this breaks the loading of Nvidia modules.
Rather annoying since Nvidia knew this issue was coming.
The fix is to back a patch out, but it's a bad idea.
Stay at 2.6.5 if you use Nvidias drivers, for now.
-- Note: If you don't agree with me, don't bother replying. I won't read it.
Or Men In Black?
Free Firefox news reader.
So that's two places where I've seen it used.
All I want is a secure system where it's easy to do anything I want. Is that too much to ask ~~ Randall Munroe
I'm not trolling here, it is just my knowledge of Linux is limited. It seems like updates to the kernel get released all the time. How is this way of fixing bugs any different than the microsoft update? Vash
I actually got the file approx. two hours ago, I guess I was lucky.
There's some new stuff in there that should fix cpufreq for my laptop, so I couldn't be happier. It's building right now. On my 450 Celeron/64MB.
I'll report back... tomorrow.*sigh*
toresbe
KernelTrap has more information about the 2.6.6 release. Looks like lots of good stuff was merged! Laptop mode, CFQ, ...
A lot of changes went in this release, and from what I read on several mailing lists, there are some regressions. For example it seems the IDE cache flush at shutdown fix is causing trouble for some people. I think I will wait for the next release...
Check out the now merged laptop mode. Allows you to really save that battery. It is also good on my home server that uses hostap - there is not too much to write on disk, so I'll set the timeout to something like once a week...
A definite must for laptop users that want a little more operating hours from their batteries.
Does anybody know if the bttv video card issue, the one that would freeze the machine when capturing from a bttv card under heavy system load, is resolved?
I'm not lazy asking about this here, it's just that I looked everywhere in the changelogs and I can't see anything about it, yet the problem is known. Perhaps the problem went away as another was fixed? Anybody has any experience on this?
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
I'm assuming it's entirely a kernel issue as cat /dev/input/mouse0 or whatever produces nothing when I play with the wheel, but it does for everything else.
[libata] Add driver for SiS 964/180 SATA.
[libata sata_sis] minor cleanups
Anyone using sis964/sata? is it working ok? any major distros you can recommend? (stuck with WinXPPro on my new machine....)
I have downloaded 2.6.5 on a remote experimental 386 server connected with 14kbps modem, and I've just finished a painful (interactive, due to very little memory) multiple-day building process... *sigh*
Sincerely,
Pan Tarhei Hosé, PhD.
"Homo sum et cogito ergo odi profanum vulgus et libido."
This is not as much bugfixing as it is improving the kernel.
Like writing better code, better memoryhandling, adding new features, improved hardwaresupport and the like.
And unlike Windows Update, you don't have to update the kernel if you don't want to. Very little software do require specific kernel-versions, as opposed to Microsoft where almost everything seems to have kernel tie-ins.
Hope this answers your question.
Not Buzzword 2.0 compliant. Please speak english.
and so far it looks prett=20 ]} $}1}&..}=3Dr}'}"}[NO CARRIER]
"It's the smell! If there is such a thing." Agent Smith - The Matrix
Foelisted for daring to suggest that the bastard known as MiB (along with KiB, GiB, TiB...) is somehow more correct or "better" than the well-known, universally accepted, industry standard MB (kB, GB, TB...).
As another poster has already mentioned, MiB is just a made-up atrocity (it's not even a real ISO standard!) which noone needs or wants.
Quality, performance, value; you get only two, and you don't always get to pick.
Windows XP here. No problems. Scrolling is a snap.
Can't help you on the build time, but this will save you time on the download, seeing as you already have the 2.6.5 source;
http://kernel.org/pub/linux/kernel/v2.6/
patch-2.6.6.bz2 09-May-2004 20:18 2.4M
No,
Linux updates = good
Microsoft updates = good
Whatever keeps those crappy windows worms at bay is great. The problem with windows updates is:
1) They don't happen often enough
2) They break things
3) They change license while you're not looking
If you're still having problems, I can break it down into even simpler terms.
OK. Adding him as a foe might be a bit extreme.
I for once, if I had moderator points that is, would just mod him troll.
I have never ever seen anyone use the XiB-units in a proffessional manner.
I have, however, noticed how manufactorers of harddrives (not RAM) use this artificially generated confusion as an excuse for delivering smaller drives than their advertising/labeling says.
The faster the XiBs get sent back to hell (where they came from), the faster we can get our this false adveritsing banned.
Someone please mod mister XiB troll!
/I'll probably get burn't for that
Not Buzzword 2.0 compliant. Please speak english.
Hrm, according to this MiB and its bastard relatives have made it into valid SI units.
Why are we letting vendors of hard disks re-scale the units of measurement so that their products appear larger by having bigger numbers on the box, its madness.
Personally I think we should redefine an inch as half a centimeter so we can all go out and score bigtime tonite.
Well-known and universally accepted? In all sciences and technological applications except for IT, MB (and kB, GB, TB, ...) are accepted to mean 10^6 (and 10^3, 10^9, 10^12, ...). Lots of scientists and engineers all over the world use them in that meaning, and they have been doing it before someone overloaded them for different use in computer contexts.
This sig under construction. Please check back later.
Not to mention that support for NTFS doesn't compile. This should be another rc version, not minor version bump...
Yeah... because Win2k SP2 didn't break any drivers at all...
If I lived in this strange world that a lot of slashdotters do where hardware apparently works easily and reliably with Windows, I would have never switched to Linux. But, in my world, Windows never loads the right drivers, and loses or breaks the drivers once you install them.
All's true that is mistrusted
Does anyone know what kernel SUSE 9.1 will contain?
It has 2.6.4-52-default when installed from CD media.
+++ UGUCAUCGUAUUUCU
It'll only break if you choose the new CONFIG_4KSTACKS option (use 4Kb for kernel stacks instead of 8Kb under the Kernel hacking menu of menuconfig). Leave that option unchecked and it should work just fine (I'm using 2.6.6-rc3-mm1 right now with NVIDIA's driver).
I wish that I could just patch the bits of the kernel that are important to me, and not the whole lot in one go.
/kernel/drivers/net/eci100 and nothing else.. (preferably from kernel configuration).
I would be far more lightly to test betas if I could download driver and filing system updates that relate to me instead of the whole kernel which may have new less stable featuers, my build times would be lower and my system would be more stable.
It would also make it easier to upgrade everything except the broken Nvidia bits....
Oh, I wish i could just patch
thank God the internet isn't a human right.
Last week someone from nVidia finally stepped up to the plate on lkml and told us all the real problem with the apic hangs. They'd told the BIOS writers long ago, but from what I can tell, only Shuttle had done anything about it. So they finally released the same info to the Linux community.
Hours after the information was released, the first patch followed. A little feedback and tweaking, and it's into the mainline kernel in less than a week. Kudos to Ross Dickson, et al, for all the work they'd done trying to fix this problem, prior to the official informatino release.
Does anyone know if the patch for either forceDeth or the 3com 2nd adapter on some nForce2 boards is in the mainline kernel, yet?
The living have better things to do than to continue hating the dead.
Has this been solved yet or must I keep running 2.6.3 with nforce2-apic and nforce2-disconnect-quick patches?
Apparently, the tradeoff is that your CPU runs hotter with the patches.
This redefinition of standard terms can be made to work both ways...
I recommend that we hereby agree to redefine the term "crap" to mean "really rather good".
Now, feel free to feel to describe any hard-drive manufacturer's product as "crap", without fearing a lawsuit.
I had mod points yesterday, I regret having wasted them now. But marking someone "foe" automatically gives them a permanent -1 from my POV, so it's just as good. :)
Quality, performance, value; you get only two, and you don't always get to pick.
You should read you own links:
It is important to recognize that the new prefixes for binary multiples are not part of the International System of Units (SI), the modern metric system.
...
Faced with this reality, the IEEE Standards Board decided that IEEE standards will use the conventional, internationally adopted, definitions of the SI prefixes. Mega will mean 1 000 000, except that the base-two definition may be used (if such usage is explicitly pointed out on a case-by-case basis) until such time that prefixes for binary multiples are adopted by an appropriate standards body.
Hopefully, it will remain that no "appropriate standards body" adopts this ridiculous notation!
Except decimal, that is
Working for necessity's mother.
Troll1: 'Linux developers aren't as diligent about testing the code before they release it.'
Troll2: That's because linux users can fix any faults they find and send them to the developers.
Troll3 (me): And because I always install bugzilla on my Linux desktop, I am sure to track each and every kernel bug I find in day to day usage with documentation and reproducable results. (ROTFL) Umm... what do you think Microsoft is doing with the Automated Crash Reports that they receive?
I'm still shocked that Mibibabyboobybytes has been accepted as a "standard!"
How many thousands of titles (possibly billions of books) have been written based on the FACT that Megabytes and Kilobytes, et al, have all been BASE-2 from the initial concept?
The ONLY people in the entire industry who considers MB/KB/et al to be in base-10 are the hard drive manufacturers, and that's just so they can claim their 230GB drives are 250GB!
You don't go out and buy a 536.89MB stick of RAM, you buy a 512MB stick!
Your video card doesn't have 134.22MB of video RAM, it has 128MB!
I don't know why, I should be used to it by now, but the "standards bodies" still blow my mind with their utter stupidity.
- Preferences: Solaris 10 (servers), Ubuntu (desktops), Solaris 11 (personal servers) -
oh when will they add good NTFS write support?:/
Granted, I know that is not the case, but 2.6.x is not even 6 months old ( 2.6.0 released December 18th, 2003) at this rate of release are we looking at 2.8 in September? This just seems crazy to me. I thought that's what the "odd" numbered kernels were for, testing. At this current rate of release it sure feels like the supposed "stable" kernels are the ones being tested on.
This isn't meant to be a troll or a flame just an observation. Many of the distros have finally gotten around 2.6, but it sure seems like the kernel devs have given the distro devs a very rapid moving target to hit. I still see all to often recomendations here and other places telling people to use 2.4x for "mission critical" use. Why?
Why is 2.6 not as reliable as 2.4 was?
Why are people in this thread commenting about all the things 2.6.6 breaks?
Why does an even kernel need to break *ANYTHING* isn't that what dev kernels are for?
I love to see progress as much as anyone, heck, I run Gentoo. I just wonder if the Kernel needs to be treated with a bit more care. Would you buy a car from an auto maker, who every month changed the engine in their car?
Introducing Microsoft Vacuum 1.0 The first Microsoft product that doesn't suck.
If you wanted an SI unit of information, it would be more sensible to use 10 bits as the basic unit (or even one bit), rather than a byte (which is actually not even a fixed unit, but is usually read as 'octet'). Attempting to graft MB = 10^6 bytes is at least as arbitary (even more so, IMHO) than defining MB = 2^20 bytes.
Maybe you should use GENTOO LINUX!!
This must be some kind of an English thing. Because not only they sound ridiculous in English, they sound more ridiculous in Finnish. Not mentioning that there are no Finnish counterparts for kibis and mebis (etc.). Probably same with many other languages too. Not mentioning that byte is shortened as t (comes from Finnish word "tavu") in Finnish. So how many Tits in Pit would you like?
Miles, pounds, stones, inches, boiling water at 273.. Take a hint
There are 10 kinds of people, those who understand the metric system and thos...
Unable to read configuration file '/bigassraid/htdig//conf/14229.conf'
Geocrawler error message.
Youve Slashdotted Kernel.org!
You Bastards!
I think they should make a short version of the change log with a summery of what changed. Does anybody know of such a thing?
kernel.org seems slashdotted from here. Good job direct-linking to it in the story.
Mirror to the rescue!
http://wuarchive.wustl.edu/pub/linux/kernel/v2.6
I just downloaded the patch and am successfully running 2.6.6. :-)
Is it me or was the boot time considerably faster than before? Almost blinked and missed it. Anyone else found that?
Now I just have to clean out init.d.
Nintendo used to use the bit as their standard unit, but that was just another example of marketing: it let them claim the Nintendo 64 had 256 Mb cartridges (i.e. 32 MB). Similarly modem manufacturers; how many people think their 56 kb/s modem ought to be transferring data at more than 7 kB/s? (Or rather, 4 kB/s, since I've never seen a 56k modem work any faster than a 33k modem...)
Visual Studio IDE = for gays. Vim + nmake = for straights.
In SI, the value of M, k, G etc. are not dependent on the unit it prefixes. True, 'byte' is not a standard SI unit, but consistency is still a Good Thing. And orthoganility too: it means the unit defines what you are talking about, and the prefix indicates how many of them you've got. It means that if you could fit 42 bytes per meter on a fictive tape, you can fit 42 kilobyte on a kilometer and not 41.015625 kilobyte per kilometer.
Is the SI prefixes are not useful for a speficif purpose, fine, don't use them. But don't take them and give them another meaning. If you want to use your own prefixes, go ahead and use new names for them.
Now, I agree kibi, mibi etc. sound pretty lame; perhaps someone should come up with better names, but we should stop using M when we don't mean 10^6.
This sig under construction. Please check back later.
HD manufacturers have no excuse for their confusing use of 10's power prefixes as the basic access unit of a disk is a 2's power of a byte.
Networking, OTOH, does have valid excuses. Eg. 64kbps for ISDN means 64000bps, which comes from 8kHz 8b width, and 10Mbps ethernet is indeed 10 million bits per second, not roughly 4.9% more than that - again because the electrical engineers didn't think in 2's power when assigning bandwidth, but rather in conventional 10's powers.
Not only hard drive manufacturers, in telecomunications (AFAIK) 1Kb = 1000 bits. You might think this is silly, but if nothing had been done, you might have technical texts where sometimes 1MB = 1000 Bytes and sometimes 1MB = 1024 Bytes. You can ask the NASA what happens when you mess with that kind of things (!). Not only that, you might know that 1MB=1000B, when refering to a Hard Disk, and you might know that 1MB=1024 when refering to computer memory. But you can't expect everyone to know what a MB is in _every_ context.
You might argue that 1KB has always been 1024B (I don't know if that's really the case), but then, "those standard bodies" can argue that 1Kx has always been 1000x. In fact, I would be interested in knowing why someone decided to call 1024B a KB.
Just for the statistics/archives:
kernel.org has 6 gigs of RAM and a 250 MBits/s connection. We have slashdotted it for over an hour.
I wonder when we'll beat that. *g*
Now if only the US would follow the rest of the world and use metric.
Wow, glad I don't have any mod points this morning, because clearly the "offtopic" choice must be missing.
Not everyone who uses Linux is a kernel hacker, especially nowadays. And yes, there are sites out there that give rundowns of what has changed. But wouldn't it be nice to have an *official* release statement that outlines the changes made? It seems logical to me that the people managing the changes would be able to articulate this the best. I think it would go a long way in making Linux seem a bit more mature.
My beliefs do not require that you agree with them.
Have you considered comedy as a rewarding career? :-)
Use ISO 8601 dates [YYYY-MM-DD]
There is only one reason for this: to make it like the metric prefixes.
Recalleth this olde giokke:
An amateur thinks one kilobyte is 1000 bytes. A computer scientist thinks one kilometre is 1024 metres.
So, the standards bodies are trying to change it to be in line with their prefixes. However, trying to change existing terms to mean something else is not a good idea.
It's the first time most anyone saw it.
Mebibytes was created in 1998 by the International Electrotechnical Commission.
In otherwords burrocrats invented this term and it's only just now being used.
And Meg?
Kilo, Mega, Giga, Tera etc were just jargon for the sake of jargon when the computer industry actually needed the binary counterpart to hundred, thousand, million, billion. So the worthless redundency got reused to prevent confusion.
It's an approprate time for addoption.
The computer industry over the years has toyed with swapping the binary and decmal count for a long time.
The Commodore 64 with 64k of ram. It didn't have 65536 bytes of ram but 65025 bytes of ram. ALMOST 64ki.
And then there is the hard disk industry using Million count instead of Mebi count.
Far too often today when people say Mega they actually mean base million.
However please:
It's all new terminology created a handful of years ago by burrocrats who (IMAO) had tree trunks stuck up certen places becouse some industry terminology was using terms still used in the electronics industry (but not in the real world) and if it continued long enough they might actually have to use ENGLISH and risk being understandable by normal people.
I don't actually exist.
[PATCH] USB: usbcore blinkenlights
The per-port LEDs on the most USB 2.0 hubs are programmable. And the USB spec describes some ways to use them, blinking to alert users about hardware (amber) or software (green) problems.
This patch is the infrastructure for that blinking. And if you should happen to "modprobe usbcore blinkenlights", the LEDs will cycle through all the ports ... which is not a USB-standard mode, but it can certainly handy be handy as a system heartbeat visible across the room.
Das ist goot, ja!
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
Linux Updates = Rare kernel issues cause reboot
Windows Updates = All Updates cause reboot.
In point of fact the Commodore 64 did have 64k(65536bytes) of ram. There were however rom memory spaces that covered over the ram. So reading a byte gave you the rom value and writuing wrote to the rom in hose areas....you could also set bytes to "swap out" the rom chips. This would allow you to read the ram "un der" the rom address spaces. Quite a nifty approach since there was memory addressing issues over 64k for that chip. :)
Just out of curiosity, what's it going to take to get to kernel version 3.0? Honestly, what changes, additions, etc have to be incorporated until they call it Kernel 3.0?
That declaration would carry some weight if your foe list wasn't that long.
99 foes, many of them with decent reputations. Did you have an unhappy childhood or something?
Use ISO 8601 dates [YYYY-MM-DD]
> you might have technical texts where sometimes 1MB = 1000 Bytes and sometimes 1MB = 1024 Bytes
For the love of god, we should all hope that such confusion is limited solely to the poster.
/ \
\ / ASCII ribbon campaign for peace
x
/ \
funny motherfucker
Just so you know, "8 centimeters" won't exactly endear you to medical or scientific women.
Windows XP. (And a shower).
Ahem
wget -c
A geek downloads a long file and you assume he didn't shower?
If you use wget you don't need to watch the download (and yes there is a Windows port).
With his computer tied up and not needing any user supervison what is a geek to do?
Play video games? Computers tied up dorkwad.
(Well not really but it's Linux.. games? If it were Windows it'd be tied up)
Surf the web (bandwith eaten up.. downloading)
Umm girl friend (hahaha I'm so funny)
Shower....
Odds are he towled off and logged into slashdot.
Now if he were using Windows XP he'd probably be showing between the BSOD and reboots.
As far as I can remember, binary units (bits & bytes) have always been packed by power of 10 (2^10, 2^20, ...).
We are speaking of machines working with base 2 : not base 10 or 7 but 2.
The size of the memory of my ZX81 was 1KB or 1024 bytes and not 1.024KB
The size of the memory of my CPC464 was 64KB or 65536 bytes and not 65.536KB
The size of the memory of my main personal computer is 1GB or 1073741824 bytes and not 1.073741824GB
Tell me if I'm wrong but AFAIK that power of ten stupidity is related to mass storage using power of ten to sell you a 111GB disk as being a 120GB.
When I will be able to store 120 times the raw image of my memory in my 120GB Maxtor/Hitashi/Seagate disk, I will agree it's a 120GB.
I don't think computer science needs those foolish names and unit changes to ensure complexity in the units. It is not a commercial game.
KISS is the rule.
Irrelevant news and morons using moderation to mod down what they disagree on. 2018 resolution: so long.
In telecommunications, it is standard practice to use 10-based prefixes. 10Mbps means 10^6 bits per second, and 56 kb/s means 56000 bits per second. In serial communications, it is normal to use a startbit and a stopbit ber byte, so you use effectively 10 bits to transmit one byte.
BTW, I've seen 56kbps modems working at 44 and even 48 kbps. The speed you get depends on the quality of the phone line. Also remember that those modems work assymetrically: the 56 kbps is only downstream. Upstream is the same speed as a 33k modem.
This sig under construction. Please check back later.
Damn i just love my 1.95Mb dsl line... Or would you call it a 2Mb? :)
The thing is, IMO, that we DO have a problem - we dont allways talk about the same units and sometimes it mattes!
When using MiB - NOONE (that knows what they are talking about) is in doubt what you mean - but if you say MB - noone really knows.
How kan a term that clears confusion EVER be a bad thing? The problem is people like you to insist on using terms that confuses - hopefully, in 10 years noone would be confused when you say 2MB and means 2000000B...
burrocrats
;)
Rulers of the underground tunnels?
The Commodore 64 with 64k of ram. It didn't have 65536 bytes of ram but 65025 bytes of ram. ALMOST 64ki.
Wha? The C64 had a full 65536 address space #0000 to #FFFF. Of course some of it was ROM and some of it was RAM and some of it was system RAM and some of it was user RAM (and some of it was reserved for cartridge ROM).
Personally I am not going to start using mebibytes for things that have been traditionally measured in megabytes. I'm just used to the fact that hard drives are actually smaller than advertised. As for transmission rates on networks... I don't know. I don't care how the device itself is rated, but we should be measuring transfer rates by how much memory the data would fill (i.e. amounts measured in base 2).
Powers of ten work great for people who are doing math on their fingers, but powers of two work much better for computers.
I do not have a signature
I don't think computer science needs those foolish names and unit changes to ensure complexity in the units. It is not a commercial game.
Computer science started by changing the names (the meaning of the names, actually). In order to reduce complexity, we need to undo that change.
KISS is the rule.
Exactly.
What is the simplest:
- k equals 1000, Ki equals 1024
or
- k equals 1000 in all sciences, except in computer science where it means 1024, most of the time. If followed by 'B' it mostly means 1024, when followed by 'b' it means 1024 when talking about memory sizes and 1000 when talking about transmission speeds. It all depends on the context.
This sig under construction. Please check back later.
You're not even knowledgable enough to present your
own case, otherwise you would know that the traditional
term for 1024 bytes is "KB", not "kB". Us old farts in this
business knew that the capital K was what differed
1000 from 1024. Alas, that trick couldn't be used
for mega and giga, thus the whole system gradually
degraded over the years. Thus I have grudingly come
to accept the new MiB, KiB, GiB norms.
9.1 Pro. has 2.6 (but with the amount of releases, who knows what build it is =\) http://www.suse.com/us/private/products/suse_linux /prof/index.html
Regards, NITZER
I have no way of knowing.
Network speeds have always been done in decimal. 10base{5,2,T} = 10 Mb = 10,000,000 bits per second. And Ethernet (in its 10base5 Thicknet variant) is old, dating from 1972. It's not just greedy hard disk manyfacturers.
I don't have a problem with disambiguating them. I just wish the names weren't as stupid. (MiB is okay, but mebibyte?!)
Any news as to when Linus will stop working
on 2.6.x and start the 2.7 branch?
Except that vendors are already doing such a thing and getting away with it, with the excuse that 1000MB is a perfectly valid value for 1GB. Take a look at your drive size reported on the box and the size reported by software, and you'll see why the new prefix was conceived in the first place.
The reasoning behind it is sound, but at the moment it's not practical for the same reason the US is still stuck with inches and all those other horrid, unintuitive units. Chances are, it'll stay that way indefinitely, because people are so resistant to change.
I wouldn't go that far. Not entirely.
I'd say you had a point saying that noone wants these changes. I certinaly don't want or particularly like them. But I can see that if they aren't needed, it's not for much longer.
They aren't "more correct" and they aren't "better". But what they are is clearer. And, like it or not, it's getting to a point where that clarity is needed.
Firstly there's the two types of manufacturer. For whatever reason the HDD manufacturers prefer to use the 10^n meanings. Maybe it's so that they can swipe more money on misleading advertising. Maybe it's some sort of tradition. Maybe it's both - a tradition that just so happens to benefit them. But they're not going to chance.
For memory-manufacturers the reason is clear. When dealing in binary (and unless something happened overnight, memory is still working on digital signals) then you can only really work to the power of two. So they're going to continue using the 2^10n notation.
Secondly you get everyone else. Whether professionals in other disciplines, or merely Joe Average taught in school (or whatever) that kilofoo is always 1000 foo, and megafoo is always 1000000 foo, they're going to have assumptions about what the prefix means that in any other context would be right but in this case would (or may... - damn HDD labels) be wrong.
And even then, if you need to refer to "one thousand bytes" then how else could you shorten it?
Back when computers where still specialist then it wasn't too much of a problem. But now computers are so prevalent that the potential for confusion is too high.
I'd love to get everyone else to change. To me "one megabyte" is "1024 x 1024 bytes" and always will be. But getting every other SI prefix to change to make way for one is unlikely.
Personally I don't "read" KiB/MiB/whatever any differently. My brain still "hears" it as kilo-/mega- or whatever. Probably always will - those "bibibibibi" bits trip me up. But when I see it (or even write it) I know with 100% certainty that the 2^10n is meant (often mentally interpreting it as "binary megabytes" or whatever...). If it's not there, I always wonder. On products it oftenleaves me always searching for the small print to be totally certain of what is meant.
Like it or not, the confusion is there. And something has to be done to reduce it. And, unfortunately, we're the ones in the minority side of the prefix-usage.
Tell that to the hard-drive manufacturers.
TiggsThey don't accept it, or use in in the industry. They may be wrong, but unfortunately they're not exactly helping things any. It means it's a part of the ocmputing industry that's muddying up the waters internally.
Tiggs
"120 chars should be enough for everyone..."
less /proc/config.gz | grep 4KSTACKS
CONFIG_4KSTACKS=y
glxinfo | grep direct
direct rendering: Yes
direct rendering: Yes
Working just dandy here. Using the nVidia 5336 drivers.
Look, I know someone's asked before why there's not a summary in the changelog already, but here's what I want:
Would someone please try and summarize the top 10 changes here? At damned near 500kb, the changelog is bigger than some compiled kernels!
Please?
This is a point that is often missed.
Even if you can't stand the "new" prefixes, you immediately know the value referred to. You may think that whoever wrote it needs a swift kick up the rear, but you understand their meaning.
TiggsIf you see (or use) the "traditional" prefixes , you wonder. Even if you are always adamant that "megabyte means 1024 bytes" there will always be a worry about whether the person you're talking to is thinking in the same terms.
Tiggs
"120 chars should be enough for everyone..."
From the looks of the changelog, the bug that prevents compilation of an SMP kernel for the SPARC32 (sun4m) architecture doesn't appear to have been fixed.
I guess this means I have to wait some more before trying out the kernel.
"I won't mod you down - I feel the need to call you a twit explicitly, rather than by implication."
a linux *kernel* update is not comparable to a windows update.
a linux *distro* is the whole usable system, like windows. so an update from a linux distro such as red hat, suse, debian, or mandrake, is comparable to a windows update.
one interesting thing about linux is that a lot of users like to upgrade the individual components rather than wait for their distro provider to do it. as far as i know, there's no comparable micro-update channel for windows.
Aah. Nail, head. Hit.
It's controversial, it's quite probably needed, yet it's given names that sound so childish that it's simply going to inflame people against adopting them.
Maybe if they'd tried coming up with terms that actually sounded a little more serious then they wouldn't be quite so hotly contested.
TiggsTiggs
"120 chars should be enough for everyone..."
Are you the guy who makes this joke whenever a new kernel is released? It gets funnier every time I read it.
Not only the NVIDIA drivers, but also the closed-source ATI drivers (the fglrx module, to be exact) will not function when 4k stacks are enabled.
:)
:)
So ATI will have to update their drivers for Fedora and other new distributions as well.
The NVIDIA Nforce2 APIC patches work perfectly with kernel 2.6.6-rc3 (which still seems to be using 8k stacks by default and is therefore still compatible with the binary drivers).
Linux just keeps getting better and better.
"Oooh, does that mean we get to kick some puffy white mad zionist butt?"
this even beats ...but does it run Lunix?
and
Imagine a beowulf cluster of those.
you are teh sux0r!
Actually, using the "bit" rather than "byte" makes particular sense when you consider the way architectures keep doubling in word size. Use of the byte made sence when everything was based around an 8-bit architecture, but now that we are moving to wider and wider word sizes it is becoming increasingly inappropriate as a baseline unit.
IMHO, use individual bits and the Mib/Mb system for binary/metric. It's the inevitable conclusion of the "big numbers are better" logic, and it makes sense from an SI perspective as well as a CS one, so it would just make everyone's life simpler if we just got there sooner and moved everything over to such measurement systems - otherwise we will go through the headaches of dozens of changes and differing conventions to get where we're going anyways.
Besides that, for real measurements it would make more sense to go by words (not bytes) and stick to real even binary numbers like 2^2=4, 2^2^2 = 2^4 = 16, 2^2^2^2 = 2^16 = 256, 2^2^2^2^2 = 2^256 = 65536, and so on, rather than the bastard scales of 2^3 (byte) 2^3*2^10 (kilobyte), 2^3*2^20 (megabyte) etc.
Am I the only person still having major problems with the Adaptec 7xxx drivers? They broke in 2.6 and they still aren't working on one of my machines. They work on all the others, though, so there's something subtly not quite right...
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
3com 3c59x is the option to select under Network>3Com
Troll1:
Most Linux developers don't have the hardware to do decient regression testing, they relly on users. Don't believe me, go ask them.
Microsoft relies, in part, on vendors writing there own drivers and certifying there hardware against Windows, I'm sure they have more regression testing facilities too.
Troll2:
Like it takes a genius to debug code, am I a genius, you don't seem to think so?
If people spent as much time fixing kernel bugs as Windows users spend re-installing and de-working Windows the kernel would be a lovely place to be.
Troll3:
Can't say I've ever seen one of those, are Microsoft spying on me again?
And I do note down any bug or missing feature I find in an application, but then I do do development work and am concerned with usability.
thank God the internet isn't a human right.
Because their motivations aside (profit, of course), the units make much more sense as they use them. The old computer science terms are inappropriate and grow more inappropriate every day, considering that the boundary between computer science and other things has been blurred for a long time and will only become more so.
At least the hard drive manufacturers have a little asterick next to every size they report and a footnote that states their meaning plainly.
Pop quiz: which definition applies to the following situations?
Mb = Megabits
MB = Megabytes
mB = Mega bytes (1000 bytes)
Cheers
I've heard for some time about how IBM is supporting Linux. But the 2.6.5->2.6.6 changelog really drove the point home, to me. It's amazing to see how much stevef@stevef95.austin.ibm.com has contributed, probably on IBM's nickel. :-) Keep it up, IBM.
Raj Against the Machine! http://social-butterfly.appspot.com/
if loading them as modules doesn't work then trying compiling them into your kernel, the only reason I can think of where this would be a bad thing is if you need a small kernel. My ethernet card refused to work until I compiled support into my 2.6 kernel, rather than loading as a module.
Ok well, firstly you probably meant "asymmetrically", not "assymetrically"...and uh...no...your wrong.
Dialup is NOT symmetric, but IS asymmetric.
/* sig */
I have tried both the nvidia nvidia driver and the Xfree86 nv driver. Both will not work with my Dell 2000FP flat panel under 2.6.6 or 2.6.5. It seemed to work under 2.6.4. Also, I am not enabling the 4K stack change. All else the same. 2.4.26 working fine
Dunno. Whenever I see the term Mebibytes I read it as "maybe bytes", after which I'm always slightly confused as to exactly how much bytes I'm getting. "Oh, you mean I may get some bytes?? Gee, and how much is that then?". It may be bytes, it may be not :-P
MB, as in 2 based, has always worked in the computer field. The only culprits are harddisk manufacturers since they abused it to claim higher capacities, and network engineers, since they refer to signal speeds (which are expressed in Hz). The latter is an exception but understandable if you know the source of the term; the former is inexcusable.
The fact that they are trying to redefine MB to be the new 10 based value isn't helping matters; they should have made a new term for that as well. By looking at a document or webpage I cannot tell off-hand whether they mean 10^6 or 2^10 when I see "MB"; they should have called it "MdB" or something ('d' for decimal), not just "MB".
"Fix it? It has been disintegrated, by definition it cannot be fixed!" - Gru in Despicable Me.
Ok well, firstly you probably meant "asymmetrically", not "assymetrically"...and uh...no...your wrong.
You're right.
Dialup is NOT symmetric, but IS asymmetric.
Isn't that what I said?
This sig under construction. Please check back later.
doh, my appologies...I misread your post ;)
/* sig */
Anyone not using a mirror whenever possible is the real bastard.
(and use a patch when it's smaller than the full kernel source).
(lookup ketchup to manage kernel patches).
The speed you get depends on the quality of the phone line. Also remember that those modems work assymetrically: the 56 kbps is only downstream. Upstream is the same speed as a 33k modem.
If anything I think the old V.34+ modems were more reliable. My old local dialup ISP used to have banks of analog lines coming into the building that were hooked up the USR V.34+ Couriers. Those babies were rock solid (even if it was a bitch to manage 450+ analog lines coming into your building). With compression enabled I could download some stuff at 8-10kB/sec. I always got 33600 baud connections and never got dropped.
Then they upgraded to digital V.90 boxes and the quality went to hell. Would typically connect at 45333 but the actual downloads were slower then downloads (compressed or not) on the V.34+ modems. The connections themselves seemed to slow down over time -- one time I downloaded a Slackware ISO -- by the end of the download the modem had slowed down to about 1kB/s. With the V.34+ modems I could push 3.5-4kB/s until hell froze over. They never had random disconnects (the V.90 bank dropped me all the time).
And don't even get me started on the upload performance of the v.90 modems. Even turning off v.90 and connecting at v.34+ speeds wouldn't help. I canceled my account and got one from the phone company -- wasn't worth paying $19.25 to have the same lousy service that I could get from the phone company for $9.95.
In my experience v.90 was nothing but marketing hype. I hated it.
I want peace on earth and goodwill toward man.
We are the United States Government! We don't do that sort of thing.
The M prefix has been a standard for centuries. Mi is not yet an Iso standard although it may be one day.
Mi is a stop-gap (standardised by the IEC) to save people who traditionally (and erroneousley) misuse M to mean 1 048 576 from translating all their current figures (so they can just change the prefix to Mi without having to alter the number). Ki is also useful in situations were something actually is an exact number of Ki (for instance due to FS's or HDD's splitting streams into 1024 or 512 bytes).
However, using MB is actually far preferable to MiB , because it makes more sense, and it is much easier to divide or mutiply by 1000 than 1024.
[When outputting the size of a user's file in an OS or sthg, it is, of course, a good idea to use both because users often need the *i form for some applications. This also makes it clear to users who have used certain OS's that erroneously use K to mean 1024 (i.e.: MSW) that K means K and not Ki.]
Joe Llywelyn Griffith Blakesley
[This post is in the public domain (copyright-free) unless otherwise stated]
I'm working on a box with the above combination and everything works fine.
When a production FreeBSD release comes out, you can rest assured it's rock-solid. They generally don't put anything out unless it actually works 99% of the time.
There's a reason Linux and its development is seen as sort of a little joke. The attitude of Linux development seems to be, if you don't understand how it works, hack it in anyway, let it out in the wild and see what happens. I still remember the memory manager hell of the 2.4 series. And with 2.6 suddenly came the transition to an incomplete udev system that only recently became workable to the point that distros didn't need ridiculous device tarballs to work around its shortcomings, though I've seen few actually switch over yet.
This is why Linux is such a moving target, and why it doesn't see widespread support from commercial software developers. Despite its use in several mission-critical situations, it is still very much an amateur effort in other areas. BSD doesn't try to be anything more than it is, and what it does, it does very well--and makes sure it works.
If Gentoo's Portage was ported to FreeBSD, I would switch completely and never look back (I prefer Portage to ports).
I can't believe it. After upgrading my USB Webcam didn't work anymore. I find out the pwc Driver is now removed :(
Shit, downgrade again...
An amateur thinks one kilobyte is 1000 bytes. A computer scientist thinks one kilometre is 1024 metres.
I do understand their reasoning though, but as you (also) pointed out, it's too late for that - they should have thought of that several decades ago.
- Preferences: Solaris 10 (servers), Ubuntu (desktops), Solaris 11 (personal servers) -
Foe-listed for being so closed-minded that you won't even consider listening to people who have ever spoken an opposing view. Worse yet - being proud of it. (Moderators - +4? You should be ashamed of yourselves.)
By the way, please add me to your foe list. I consider those 99 people to be good company.
Except, of course, for the fact that you're making the prefixes even more confusing, since a small m means milli.
Yes, I agree that the concept of a millibyte is just plain silly, but the point remains.
(And let's not even get into conversations: "Pardon me, when you said 'megabyte', did you mean mega byte or megabyte?")
"I won't mod you down - I feel the need to call you a twit explicitly, rather than by implication."
kernel.org has a ton of bandwidth.
Do you REALLY think that all the books that have been written on the subject will have been updated in ONLY 10 years?!?
Life must be nice on your planet.
As far as confusion goes, I haven't yet met anyone who knows anything about networking who thinks that network speeds are rated in base-2.
It has been long accepted, though equally long protested, that hard drive manufacturers list in terms of speed and not size.
Do you really think that Maxtor, Seagate, Western Digital, et al will suddenly be honest about their drive sizes and list them with this new silly term?
You know... that's part of it - if they had to come up with another term, why make it sound like the person has had too much novacaine?
If I start using that word in front of customers, they'd think I need hospitalization!
As far as computer memory, program usage, space allocation, etc.... it's all base-2. Always has been, and it's impossible to change considering the way computers work.
Changing the name several decades after common use would be like changing the word "airplane" to "jackalope" and "car" to "snipe".
No matter which way you look at it, it's a dumb idea.
- Preferences: Solaris 10 (servers), Ubuntu (desktops), Solaris 11 (personal servers) -
It is, however, an NIST standard.
Mega and all its cousins are units in BASE 10. Please accept this and move on. in 1998 the SI accepted mebi and cousins as valid units for binary measurements.
It's true, i was taught at college that 1k=1024 bytes but it's not the case anymore. Hard drive manufacturers got what they wanted
Open Source Java Web Forum with LDAP authentication
Yes, as soon as you see MiB you know they're talking about 1000000 bytes because Mi is obviously short for million.
Thank goodness they finally got a clear TLA!
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
What is the simplest?
Yeah but you're forgetting the heart of the matter - Kibi, Mebi et al sound crap.
I think you got it backwards, MiB is 1000000 bytes (thinki Mi llion B ytes), while MB (Mega Bytes) is 1024X1024 bytes, and Mb (Mega bytes) is the disk manufacturers confusing nomenclature for 1000000 bytes.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
(dmcbride@sco.com)
Added proprietary Unix System V lines of code
From the reply threads, it looks like we're on the verge of another Religious War.
There's only one thing that can be done now.
Brother Maynard! Bring out the Holy Hand Grenade!
(chanting)
Die Jesu domine,
Dona eis requiem.
Die Jesu domine,
Dona eis requiem.
Mebibytes is easier to pronounce if you've got a mouthful of biscuits (cookies for you Americans! :-)
I think... that they'll EAT YOUR BRAAANE!
Brother Maynard, where art thou? (take two)
From the reply threads, it looks like we're on the verge of another Religious War.
There's only one thing that can be done now.
Brother Maynard! Bring out the Holy Hand Grenade!
(Take that, you mebi/mibi/mobi/wobblybytes!)
Look it's very simple. Bytes are measured in base-2 units, everything else is base-10.
It's bits (small 'b'), so it's 100 x 10^6 bits per second. Which is also 12.5 Million bytes per second, or roughly 11.9MBps.
It's bytes (capital 'B'), so that's 100 x 2^10 bytes per second.
Getting a little silly now. That's 1 x 10^9 Hz (cycles per second).
Ah, this is where all the supposed controversy comes from. The hard drive manufacturers want to use base-10 units so that their drives sound larger than they really are. Everywhere else, base-2 units are used for measurements of bytes (including your file browser).
The only two problems I can see is hard disk manufacturers wanting their drives to sound larger, and marketroids getting the capitalization wrong (bits/Bytes, milli/Mega, ...). None of that is the fault of the units/suffixes, the people who made them, or the people using them. Get over it.
2.6.x is not a decimal
so 2.6.10 > 2.6.9
otherwise you'd get some stupid update scheme infinitly approaching 2.7 (i.e. 2.6.99999)
The Neo-Bohemian Techno-Socialist
Here's a suggestion: perhaps your ISP was simply overloaded. Having all those people downloading web pages and software is a fair bit of bandwidth. Perhaps your ISP (or the people they're connected to) didn't have the bandwidth required. Also, maybe your ISP was "shaping" the connection. This would certainly apply to long downloads like your Slackware ISO download. It started off fast but over time the ISP router(s) gave you lower and lower priority to your connection. ISP's rarely have enough bandwidth to accomodate all their customers downloading at the same time. They have to balance things carefully or people get upset at slow web pages and move to another ISP. So they give higher priority to short fast connections like those required to download web pages and all the little images on them. Long-standing connections are given a lower priority in the packet queue, and maybe even throttled over time.
Here's a suggestion: perhaps your ISP was simply overloaded. Having all those people downloading web pages and software is a fair bit of bandwidth
On a dialup? I've spent seven years in the ISP business working for a variety of small businesses and telcos and I never was asked (or suggested) to throttle dialup users. At the ISP in question we had no mechanisms in place to limit bandwidth open to dialup users -- we also had two T1s that (until we started offering a fixed wireless service) never even approached capacity.
No, it was my experience that v.90 had a lower error tolerance then v.34+. All modems will drop to a lower baud rate to try and save the connection before giving up on it -- it just seemed that v.90 did this way more often and to a much bigger extent then v.34+ ever did (over the exact same phone lines). On decent modems there's a command that you can use right after a connection to view these stats -- but it's been years since I've used my hayes command set :)
I want peace on earth and goodwill toward man.
We are the United States Government! We don't do that sort of thing.
What we need to do is declare (with force of law) that for purposes of disk storage, 1,000,000,000 bytes is called a costly-byte.
Kilobytes = 1024 bytes. Can be abbreviated to kB.
Not kilo bytes = 1000 bytes.
There we go, all sorted. It isn't kilo, it's kilobytes. Completely different.
Ahem. *cof*orthogonality*cof*
You are close...
When one person says "folder" and another says "directory" the two people sometimes get confused. It's rare, but I've seen it happen.
With Linux and Windows, there is a similar confusion. Modules under Linux -- serve the same purpose and are largely in the same parts of the OS -- as drivers do under Windows.
As for Nvidia, they have installation software that is not too hard to use. It will install the NVidia kernel module for the current kernel. I'm sure you've used similar ones for Windows graphics drivers.
The kernel policies are clear, and do not cause problems for the (quick guesstimate) ~4,000 other 'drivers' bundled with the base Linux kernel. NVidia has chosen not to follow those policies so like the problematic non-WHQL Windows drivers, they can suffer similar problems.
That said, for the most part, I've been very happy with the NVidia cards I have. They were very flaky 2 years ago. Now, they work well...even with the 2.6.x kernels.
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
And then you see posts about how horrible (for example) religion is - causes people to kill each other for no reason.
And here we have people determining whether another is a friend or foe depending on the terminology he uses for specifying a certain number.
There's definitely no hope for us..
The Vmware 4 modules also don't compile any more.
Tell that to the hard-drive manufacturers. They don't accept it, or use in in the industry.
Surely however, you know that's just for marketing reasons. Using MB = 10^6 bytes lets them market hard drives as being bigger, sometimes a LOT bigger, than they really are. Even Windows will use the correct definition when reporting disk sizes...
As a comparison, consider one of the recent 200 GB hard drives.
200*1024^3 - 200*10^9 = 14748364800, 14748364800/1024^3 = 13.74 GB.
So basically, the hard drive manufacturers are cheating customers out of a lot of pr0n storage by using the wrong units... Then again, HDD space is so cheap nowadays, most people could care less.
Anyway, my point is, what the hard drive manufacturers/marketers use has fuck nowt to do with what's "correct" or standardized. If it were the other way around, you can be sure they'd still pick whatever unit made their hard drives seem the biggest...
Quality, performance, value; you get only two, and you don't always get to pick.
Your rule is nice and simple, but also wrong.
Consider RAM capacity (which I forgot to put in the quiz). If you look on an individual chip of a DIMM, you'll frequently see the capacity written in Mb. That means 2^20 bits. (It might also be a little higher than you'd expect; they might include the ECC bits in that count. I don't remember for sure.)
The hard drive manufacturers want to use base-10 units so that their drives sound larger than they really are. Everywhere else, base-2 units are used for measurements of bytes (including your file browser).
If you ever happen upon a really old Windows system, look more closely. IIRC there was at least one place that said MB and meant powers of 10. It's "obviously" a bug (and long since corrected), but I can easily see how it happened.
Another quiz item I forgot: floppy disks. A "floppy megabyte" (as in 1.44 MB) is 1000 * 1024 bytes.
OK, you convinced me.
But why stop there? Let's travel that roat some further; from now on:
Kilometer = 1609 meter. Can be abbreviated to km. Not kilo meter = 1000 meter.
Kilogram = 453.6 gram. Can be abbreviated to kg. Not kilo gram = 1000 gram.
Kilojoule = 4186.8 joule. Can be abbreviated to kJ. Not kilo joule = 1000 joule.
Kilowatt = 745.7 watt. Can be abbreviated to kW. Not kilo watt = 1000 watt.
This sig under construction. Please check back later.
I agree that it absolutely makes sense in communication technologies.
Imagine if you were working with a baseband digital transmission at 10Kbit/s yet you'd have to oscillate at 1024 Hz...
Finding that your new network card does 100000000 bits/s instead of 102400000 is a lot less inconvenient.
Well. All kernel updates require reboots in Linux and far from all Windows updates require reboot.
I admit that most installation program say you should reboot etc. However, it is mostly unnessesary. Often it is good enough to reload your shell (log out/in again).
Well, what would be better?
Maybe the only available solution is to find a set which actually work and don't sound shit.
Whether we like it or not, there needs to be some sort of differentiation here. And capitalisation doesn't work. (Anything screwed over by a change in typeface isn't really appropriate) But as we've seen, coming up with an alternative that sounds like a joke doesn't really work.
Kibibyte, Mebibyte, Gibibyte - I can hardly type them without sniggering. Let alone trying to say them.
So, what would work as a better prefix for the 2^10n series?
Tiggs
"120 chars should be enough for everyone..."
Ok, RAM chips are probably a bit of a special case. Since manufacturers want to group them into bytes it makes sense to use the same units. And it's not usually something that the general public come across, so I guess it's ok.
The "1.44MB" (and the rare "2.88MB") floppies are a real doozy. They were twice the size of 720KB floppies, but "1440KB" was too much of a mouthfull. The result, as you note, is a mixture of the base-2 and base-10 units.
2.6.6, the Kernel of the Beast
In otherwords burrocrats invented this term and it's only just now being used.
So it was created by a bunch of asses, then...I thought as much. :-P
Ok, RAM chips are probably a bit of a special case. Since manufacturers want to group them into bytes it makes sense to use the same units. And it's not usually something that the general public come across, so I guess it's ok.
It's not the only such place. It's just the first one I thought of.
Here's one going the other way (using powers of two with bytes). SCSI transfer rates are typically given in MBps. They mean 8 or 16 bytes per clock at N MHz. Since we've already established MHz means powers of 10, that means the MBps here is in powers of 10 also.
You're right in the telecommunication work 1 Kb/s = 1000 b/s.
I know: I'm working at Alcatel and I was very surprised the first time I saw, so the bitrate is 1.024 Kb/s and I asked ok so this Kb is obviously 1000b why not use Kib/s instead and I was answered that Kibi and Mibi were not really accepted in the telecom world..
Please someone mod the parent up, he clearly shows that the distinction between Kilo and Kibi must be clearly used..
can't we have a linux kernel thats released that doesn't need a new release every few weeks? After all MS has a new version of windows out every 3 years or so. And doesn't need to constantly update their kernel for it. Its a rarity for a windows kernel to be updated. Since not so many functions are integrated into it as the linux kernel has. They're all external and therefore when a problem with them is found, the kernel doesn't need updating, just the outside component. Sure makes it easier on me and on MS instead of having to compile a new kernel, they just compile the component that needs updating.
My Gawd WTF...
Seems to me that storage (RAM, HD space, USB drives) uses 2^10, while speed measurements use 10^3. Originally this was true without exception (perhaps back when lying about using true megabytes wasn't all that beneficial).
I propose KiddieByte, MangleByte and JiggleByte for kB, MB, and GB respectively.
(No, I don't pronounce giga "jigga", but when I see GB I do think "soft G" for some reason)
We are speaking about memory units of size always being a power of 2 (yeah, you can sum banks to have something else like 768, whatever ...)
... let's say that. ... true (since they are "using" it).
...
When you think about other units like weight or lenght you forget that they are not coming by 8, 256, 512 or 1024.
When your hard disk is divided in sectors (512 bytes) gathered in blocks of 4096 bytes, what will you say ? Your disk is using blocks of 4.096KB or 4KiB ?
Okay, let's assume you use K/M/G for 10^3 increments.
Let's assume you use those cheesy-named units for 2^10 increments.
What's the use of the K/M/G names when you buy a memory module ?
"Hi, I would like to by an 1.073741824GB memory module" (Be very carefull, if you ask 1GB, the vendor could sell you 1GiB with 73741824 defective/missing bytes)
Ah, yes, you can use the "other" unit.
Mmmh, ok
What could be the use for the "old" widely-used-since-years units ?
Ah, yes, if you buy an hard disk
Or if you program something and use exactly one multiple of 10^3 bytes as an allocation unit (not often, to say the least) and say it in a dinner (difficult, but hey, who says it wasn't ?)
Right
That would mean that with their new meaning, the "old"/wusyu would be pretty useless.
Ok
So we would have GB and GiB.
How do you say that ? Gee-bee ? Gee-ai-bee ? Wonderful. I wish you luck.
I for one, will still use the units I know since 20 years now.
The fact "common people" have trouble thinking that for memory units we (scientists) are using K=1024 instead K=1000 for technical reasons means nothing to me.
Irrelevant news and morons using moderation to mod down what they disagree on. 2018 resolution: so long.
What you do not seems to understand is that there is a reason for that 2^10.
The numbers you have used have not meaning at all.
Irrelevant news and morons using moderation to mod down what they disagree on. 2018 resolution: so long.
Wikileaks, no DNS
Apparently mice work in the USB port but not in the PS/2 port. This was discussed on the sydney LUG recently. I can definitely say it works in the USB port.
Well, I decided to install 2.6.6 and give it a try.
It compiled well, I tried to change as few settings as possible.
But, unfortunately, it couldn't cope with my CD/RW drive and complained about the interrupts (really, this problem should already be fixed in a stable kernel...)
There had been absolutely no problems with 2.6.5, but now the drive just wasn't working. It couldn't even eject.
I tried almost everything:) Finally I decided not to let ACPI deal with the interrupts (some forums gave me the idea).
And voila! pci=noacpi saved the day (night)!
But I don't know still, whether it also broke any useful features. I'll keep trying.
Reason 1: K 2.6 is really a beta test Windows:
What has happened is that we have in part had a major release for BETA tsting that a lot of people have started using (Alpha test was 2.5) so there is a shakeout of problems. The same occurs in Windows with early release users raising bugs tht are fixed in the final release, if and only if they are major enough. Linux is a bit more open to repairing bugs faster so there is a cycle of fast updates. As noted elsewhere it will slow down a lot shortly. Then a new alpha branch (2.7) will be openned and a lot of work will go on there. Alpha branches (2.7) are where angles fear to tread.
Reason 2: Linux kernel = low level stuff + drivers.
Windows has the drivers externalised a fair amount so that the vendor can replace the driver without specific change to Windows. (This is good). Linux does more with imbedding drivers for performance (This is good too for different reasons). Updates to drivers therefore cause more releases (perceived as bad, but not really)
Hope this helps...
After running YOU, it is 2.6.54-3 Art
They do have meaning.
I don't think computer science needs those foolish names and unit changes to ensure complexity in the units. It is not a commercial game.
I just figured it was easier to say "1k" than "1024" and I can't bother to remember "1048576" so it's just "1M".
I seem to recall reading (in The Art by Knuth I think but may be wrong) a suggestion that 1024 be noted by "kk", 1000 by "k", 1048576 by "MM" and, 1000000 by "M".
I think it might be a nice idea but it's still easier to say "one k" than "one k k".
... unfortunately.
GAYEST unit ever named. Except, perhaps the gebibyte (WTF??)
When the term is difficult to pronounce and sounds incredibly stupid to any native English speaker. At least that's why I don't like the unit. The suffix MiB is ok, but I'll call it "megabyte" until someone comes up with an acceptable name.
True story.
Flamebait my arse. The original post spouting crap about mebibytes is the flamebait.
Considering that there's NO SUCK FUCKING THING AS A METER, nobody will care. Go and cry a liter of tears... ...FUCKIN LITER what is up with you americans?
If you didn't get my point, it was that on Linux the only time you have to reboot in order to make changes effective, is if you update the kernel.
In Windows, even though things have gotten slightly better in XP, you'll almost have reboot for every minor change you can imagine.
For instance, a update in your Webbrowser or Mediaplayer. Now that makes sense doesn't it? That was my point.
Maybe my bad for not being clear enough, but I feel kinda insulted when people think I'm that ignorant.
Not Buzzword 2.0 compliant. Please speak english.
Glad you noticed :)
This sig under construction. Please check back later.
Ok, right. Maybe I need an explaination but I fail the see the link.
You are just mixing units are you ?
Anyway, as I already said, those numbers have no meaning. They are arbitrary if you like it better.
The 2^10 and other binary "packing" of units is not arbitrary (not that much). There is a technical reason, as you may already know.
Did I missed your point ?
Irrelevant news and morons using moderation to mod down what they disagree on. 2018 resolution: so long.
Especially when you know that in some languages it could have a bad effect.
Bits are already subject to childish jokes.
I just can't imagine a forum of students being completely still when the professors use that term.
Irrelevant news and morons using moderation to mod down what they disagree on. 2018 resolution: so long.
They could be useful to some group of people, just like k = 1042 can be useful to a group of people. To me, that's not enough to warrant the use of existing prefixes with a different meaning.
In science, when an existing unit is not very useful, people in the field often use different units. That's fine, since they don't use existing names for their new units. They invent their own, so one doesn't have to know each and every possible context to know that something you see might have a meaning different from what you think it is.
This sig under construction. Please check back later.
Sorry, not. M remains 10^6, like Megaton, Megawatt etc. There's too many other places where it's used. The "i" is to signify "binary". B is byte. b is bit.
1KB=1000 bytes.
1KiB=1024 bytes.
1Kb=125 bytes.
1KiB=128 bytes.
Finally you can spit into face of markedroid and say "Yes, this drive is 120 megabytes. But it will fit only 100 gibibits of raw data. After formatting it leaves less than 80. So go screw yourself with your 200 CDs you claim it would fit!"
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
Well, luckily since bit is about the least unit, prefixes for millibits or picobits were free to use so nobody would think 10mB isn't 8000 bits, not 0.08 bit :)
More like changing automobile to car, digital brain to computer, cinematograph to cinema, or TV typewriter to terminal.
The ONLY people in the entire industry who considers MB/KB/et al to be in base-10 are the hard drive manufacturers, and that's just so they can claim their 230GB drives are 250GB!
Do they consider kilobytes to be 10^3 bytes? I only ask because in the old days of floppy disks you had 720K floppies. A "K" was a good old 1024 byte kilobyte. (In fact now I think about it, nobody ever used to say kbytes because of the possible confusion. A "K" was a specific unit of memory nothing to do with SI). Anyway, somebody thought of a way of doubling the amount of storage on a floppy to 1440K. These became known as 1.44 Megabyte disks. Think carefully in terms of actual bytes. The actual amount of storage is 1.44 x 1000 x 1024 bytes.
All I want is a secure system where it's easy to do anything I want. Is that too much to ask ~~ Randall Munroe