Submitted by “Btrfs holds the promise of giving Linux many enterprise class file system features similar to ZFS but with even more features and better performance. In fact, many Linux experts think that btrfs is one of the keys to the future of Linux. While btrfs is not quite ready to be your only file system, it is in the kernel ready for testing and is still undergoing very heavy development. In this article we will introduce the key features of btrfs find out how it compares to existing file systems.”
Benchmarking compression with 0000-data is utterly pointless.
I have high hopes for ButterFS, maybe it will become BetterFS someday
Chris Mason, the project lead, has said that he prefers to pronounce btrfs “Better FS”. I thought sure that would catch on. It’s catchy and positive. How in the world did this “Butter FS” pronunciation become so popular?
To the extent that names make a difference, I’m calling it Better FS, and I hope you do too.
The pronunciation as “butter” comes from trying to pronounce “btr”. In English, and many related languages, the voiced bilabial plosive (“b”) ends in a closed back rounded vowel (“u”).
Dude, common knowledge.
Only pronouncing “btr” sounds more like better than butter. The letter b is usually pronounced “be” not “bu”.
That depends on what type you mean. If you spell it out like b-t-r, it becomes “be”. If you pronounce it “btr”, it becomes “bu”.
No, it becomes BetterFS or BitterFS (bee-tee-are-eff-ess). I dont know how the dutch pronounces B but in English there are no circumstances where the letter B is pronounced “bu”.
Edited 2009-04-23 17:59 UTC
You’re missing Thom’s point as you’re still sounding out each letter rather than trying to read the letters collectively as a word.
Plus you’re not taking dialects into account.
Many northern accents would sound “B” as “Bu” even when reading the letter out individually.
The letter “b” is pronounced as “bee” so BitterFS should be the most appropriate pronounciation…
The letter “b” doesn’t make the “bee” sound. Ask any four year old, It makes the “buh” sound. Or watch some sesame street.
“buh” ….. “Tah” ….. “er”
“buh” … “Tah” … “er”
“buh” .. “Tah” .. “er”
“buh” . “Tah” . “er”
“buh””Tah””er”
“buhTaher”
“butter”
Butter!!!
Wow, which version of English do you speak that splits “butter” into three syllables?
It’s pronounced “buh” “ter”.
But that’s beside the point. The letter “b” has only one sound: “beu” (where the “eu” is pronounced like the “o” in wood). Nowhere do you pronounce “b” as in “buh” (like the “bu” in butt).
btr b= buh,t= tah, r= er
ask a four year old. He’ll set you straight.
I asked my 5-year old and 11-year old cousins to pronounce “btr”. One said “bitter”, the other said “be-tehr” (where the first “e” sounds like the “oo” for wood). I also asked the wife’s 13-year niece, and she said “bee-ter”. None came anywhere close to “butter”.
When I asked them to sound-out the word “butter”, they all came back with 2 syllables. Just like they’re supposed to … since the word is only two syllables: “but” “er” or “buh” “ter”.
no, thom is right, why are you arguing this, did you think everyone said butterfs because of a signal from god instead?
The pronunciation issue is not worth a long back and forth flame war. But…
Dude, my native language is English. And I disagree. “Better FS” is just as natural a pronunciation. It’s the pronunciation preferred by the filesystem’s author, which is significant. If you said that Thom Holwerda, as a proper noun, should be pronounced “Baked Beans”, that carries some weight, despite anyone else’s opinion as to whether it makes sense or not. (Common knowledge.) And “Better FS” is better PR, to boot.
Edited 2009-04-22 19:23 UTC
You completely missed the point of my post.
You asked why the pronunciation “butter” stuck instead of “better” – not which of the two is more fitting or preferred by the author, or what makes more sense.
I’m simply stating that a very logical reason why “butter” sticks (…) is because of the phonology of English; ask any random English-speaking person in the UK or the US to pronounce “btr” as one word (instead of spelling it out), they’ll say something like “butr”, simply because the phonology of English leads to people adding a closed back rounded vowel (“u”) between a voiced bilabial plosive (“b”) and a voicless alveolar plosive (“t”).
Why? It’s more logical if you look at the shape of the mouth and tongue going from “b” to “t”. Adding an “e” between the “b” and a “t” requires slightly more effort. Phonology is fun, isn’t it?
My remark about common knowledge was obviously in jest.
At first, yes. After I posted, it occurred to me that sans the “Common knowledge” remark, you might have been sincerely answering the question I posed.
Although I will reiterate that I find “Better” to be the easier and more natural pronunciation. (As well as the more marketable one, and the one preferred by the author.) However, people with different native languages find various different pronunciations of English words to be more natural. So I suppose that a vowel-less sequence of 3 letters might have a wide range of interpretation internationally.
On a slightly tangential topic, I had a friend, years ago, who used to correct me when I used the word “pronunciation”, insisting that it was pronounced “pronounciation”. It was so ironic, that I only ever tried once to set him straight. He didn’t believe me.
Edited 2009-04-22 20:20 UTC
Why do people keep saying “butter sticks”?
It’s “sticks of butter,” i.e. I need a couple sticks of butter for this cake I’m baking.
I was using sticks as a verb…
Dude, my native language is English. And I disagree. “Better FS” is just as natural a pronunciation
I agree, both ButterFS and BetterFS are logical and natural pronunciations for English speakers.
For me (I happen to be from Finland and we say “B” similarly to as you’d pronounce “be” in “beta”) BetterFuss sounds natural
Common knowledge that Btr phonetically sounds like butter, in English? Or do you mean UK English?
How about they just named it BTFS ala Binary-tree Filesystem? Too UPS for them?
Nah, that will become ButtFS.
Dude, common knowledge.
The first name you hear is generally the way you say it for a long time – old habits are hard to break.
I was thinking it was called ButterFS, because it was so flexible on what you can add in to it (modular and always changing shape). It’s still years away from even coming close to how stable and debugged ZFS is today, but it does look attractive. If it can be integrated alongside ZFS in multiple OS’s, that would be interesting.
ZFS has been out now for 4 years after it was introduced to OpenSolaris in Nov ’05. It’ll be interesting to see how quickly it comes along with the open community working on it versus on how Sun did with their developers. Running with the GPL license should pull in a larger pool of developers I would think as well. Only time will tell!
PS: http://lwn.net/Articles/267896/ shows Zach Brown of Oracle saying it’s “Butter FS”. And according to wikipedia:
Valerie Henson. (2008-01-31). http://mirror.linux.org.au/pub/linux.conf.au/2008/Thu/mel8-262.ogg Chunkfs: Fast file system check and repair. Retrieved on 2008-02-05. Event occurs at 18m 49s. “It’s called Butter FS or B-tree FS, but all the cool kids say Butter FS”
Edited 2009-04-22 18:45 UTC
Zack and Valerie, as much as I respect them, are not the primary authors. I’m not certain whether or not they are even contributors. Zack may work at Oracle, but the only btrfs developer I have ever heard from, or heard mentioned, is Chris Mason. (Although he *has* thanked Zack, on occasion, for “his ideas”.) If Chris, the author, says “Better”, I say “Better”.
Edited 2009-04-22 19:40 UTC
Try marketing it under the “Butt-her FS” brand, that would do it for geeks
Kochise
Most people I know at the LUGs here call it BitterFS. I guess that may stick with Sun buy out.
-Ad
Chris has confirmed that btrfs is still strategically important to Oracle. Perhaps the more burning question concerns the fate of ZooFS.
Edited 2009-04-22 21:01 UTC
I doubt whether Oracle have given it any thought at all. The people who will end up making the decisions probably won’t even know what ZFS is. However, by buying Sun Oracle have committed themselves to supporting Solaris and ZFS for those who are using it, despite its decline over the past decade, so it’ll be around for some time yet. Just as people aren’t going to reformat everything to get ZFS, those using ZFS aren’t going to reformat everything when btrfs comes along.
… is more than just measuring sequential write and read speed.
There are a damn lot of aspects to consider
* How long does it take to create/delete files/directories? It could be quite annoying deleting a large directory tree. Take a gcc 4.x.x directory tree how many minutes do you wait until it is deleted?
* How much does the performance degrade if more than one actions are performed simultaniously? I assume, BTRFS is designed with server workloads in mind.
* How does the performance degrade over the time? Does the FS fragment? Does reusing a former allocated block imply a serious performance penalty? Are FS maintainance operations (e.g. defragmentation) automatically performed during idle times?
* Does the FS guarantee QoS (quality of service)? Streaming applications depend on guaranteed transfer rates.
…
pica
I don’t know much about the CDDL license for ZFS, but it seems counterproductive for the BTRFS and ZFS teams to be working in parallel on similar things. It would be nice if they could share code, ideas, etc. Isn’t this one of the ideas behind open source software?
not necessarily
I would imagine that CDDL was chosen by Sun to prevent inclusion inside the Linux kernel. And contrary to what most people think, licenses, open or proprietary, matter. A -lot-.
Whether or not brtfs is a duplication of Sun’s work is irrelevant, as any attempt to stick the ZFS inside an incompatible (license wise) platform will most likely end with a -massive- law-suite.
– Gilboa
How is that not a “rampant layering violation” or at the very least, a violation of the “do one thing only and do it well” Unix mantra? They use a single command, the one used to format the filesystem, to manage everything, including all the storage devices? All the storage handling is integrated into the filesystem?
What happened to all the plans they had to coordinate with md/dm to manage the storage side of things, and keeping the filesystem just a filesystem?
Looks like the filesystem *is* the storage system. Even their FAQ mentions that it’s a layering violation. And where’s the division of labour that is the hallmark of Unix systems (small programs that do 1 thing only and do it well)? This is one mega-app that manages everything, from the disks, to the RAID layout, to formatting the filesystem, to managing all of the above.
Having to do the storage management at fs creation time still leaves in the old “1 volume, 1 fs” days of old. The beauty of pooled storage setups is you don’t have to worry about things like this (partitioning, volume management, assigning disk space to specific filesystems, etc).
At least with ZFS, you have a clear separation of storage management from volume/filesystem management (zpool <–> zvol <–> zfs). You can even use ZFS with other filesystems (UFS/swap support on zvols is native to FreeBSD/Solaris, any fs is supported on zvols exported via iSCSI).
It’s interesting how so many people look at btrfs as “storage done right, done the Linux-way” when in reality, it’s methods are worse (or at least no better) than ZFS’.
Maybe behind the scenes, it’s all nicely layered … but then so is ZFS.