The purpose of a file system is for peristant storage of data. The purpose of a database is for efficient organized access to data. But all too often we see database groupies advocating the filesystem-as-database, or the filesystem groupies advocating database-as-filesystem.
I use a database when I want to quickly find or correlate some data. "What was the company's net revenue from the GruntMaster2K product for April 2001?" I use a filesystem when I already know where the data is. "find" is a useful command, but I only use it about twice a week, most other times I know where the file is or can trivially browse for it, so the hierarchical structure of the filesystem is more than adequate.
I agree that GConf was a bad example. But more than that, the whole article was "gee whiz" mental masturbation.
The problem with GConf is that it uses very many very small files. The author's solution seems to be to take this to an extreme, and make each individual key a file. Huh? Opening a file is slow. Opening one hundred files is one hundred times as slow.
You know those database programmers who think everything should be done a database, even screensavers? And those XML programmers who think everything should be done with XML, even screensavers? I almost expected this filesystem programmer to come right out and say screensavers should be done in the filesystem...
Re:Data, even metadata, belongs in files, not fs
on
State Of The Filesystem
·
· Score: 2, Insightful
Many detractors of the UNIX security model point to the lack of ACLs and fine-grained security.
You don't throw away the door just because it doesn't have a lock. You simply put a lock on it!
FreeBSD now has ACLs, and it did it without throwing away UFS. It didn't need to replace the "everything is a file" model to do it. It just expanded the available extended attributes. To get ACLs you will certainly need to extend the filesystem, but you don't need to replace it completely.
p.s. FreeBSD also got soft updates without throwing away UFS. In short, you don't need to throw away the old to get the new, something a lot of developers don't seem to understand.
Presumably, on a machine that handles file metadata at the filesystem level, you would abandon this barbaric idea
Why is this idea barbaric? Filesystem metadata should only be for file metadata, and not content metadata. There is a crucial difference. ID3 tags are part of the content of MP3 files.
What is this obsession with reinventing filesystems? Yes, we all want blazingly fast and efficient filesystems that are self tuning. But why this rush to throw out the basic concepts that have been proven to work?
The article is a grabbag of featurettes, but with no coherent model. The classic UNIX filesystem (ufs, ffs, ext) has a simple model: everything is a file. Directories are files. Devices are files. Files are files. 99.99% of the time this is sufficient. It could use some tuning, but there is no reason to replace it with a big accretion of warts. Environment variables as files is an excellent idea. But the example of "gconf keys as files" is a wart, since gconf files are already files. If there is a pressing need for inode-free files (metadata only) then extend the existing model with inode-free files. Simple.
When you do a brainstorming session, you don't take all of the ideas on the chalkboard and roll them up into one new system. Instead you take the best ideas that work together and leave the others behind. Sometimes you even have to leave a good idea on the table. It's great that people are coming up with great ideas. But those ideas need to be fit into a working model that is lucid and elegant.
The FSF groupies keep saying it's free speech not free beer. They keep saying that the GPL doesn't prevent you from selling the software. They keep saying Free Software can be commercial software.
So just sell the software! You'll go broke, but at least you can hang out with the FSF groupies in the unemployment line.
p.s. The FSF itself does NOT say this, only those who think RMS is their savior, but never bothered to read any of his writings.
p.p.s. And of course, if you're the typical slashdot reader, you still won't get it. So let me spell it out: you can't make a living selling Free Software by itself.
Say it ain't so! I've been told now for three years that it was Linux doing all of the special effects for LOTR, and now you tell me that IRIX was part of it!
Does this mean... that... Linux can't... do everything? Noooo!
Yes, monocultures are bad, and since GNU/Linux by definition can't become a monoculture (too many sources of distributions), it's a better deal than any one vendor's products.
While it's certainly better than a single-vendor monoculture, it's still a monoculture. In all probability I still won't get a choice of OS, and will probably have to pay a "Redhat tax". (ever see a Linux preload where they did NOT charge for Linux?)
1. Better tech support from ISPs, hardware vendors, PC OEMs, etc, etc etc.
"Okay bring up the DrakX and..."
"But I'm using Slackware!"
I'm not being facetious here. I use FreeBSD and I once had to lie to an ISP and tell them I was using Redhat in order to get support, because they only supported Windows, Mac and Redhat.
Because Finland under the EU is not a sovereign nation. If the two senators from Wyoming (for example) don't vote for a law, but all the other US states do, Wyoming doesn't get to opt out.
Take over the desktop. then worry about a browser.
What is this obsession you Linux guys have with taking over the desktop? How do I know the AC was a Linux guy? Simply because he was talking about taking over the dekstop!
Monocultures are bad. I don't want one bad situation with another.
For those of you asleep under a rock, let me present this "excuse" in a nutshell:
1) Someone remarks that FreeBSD and BSD/OS dominate the Netcraft uptime rankings.
2) Linux advocate offers the excuse "but the uptime counter thingy rolls over after 497 days, so the Netcraft rankings are not fair."
If you look at this thread from the beginning, you'll see that eht wasn't bitching about the Netcraft rankings, but bitching about the people bitching about Netcraft rankings. And if you had bothered reading his damn post, you would have seen that eht actually read the article, which wasn't about uptime at all!
You just have to take his extreme web-centricity into account. All applications are about data. What can you do with data? Create it, manipulate it, and retrieve it. By using the terms he did, he is arguing that the browser should be the platform (even though he says he's not).
When you're playing Quake you're most certainly not interacting with a database, but you are manipulating data.
How many Open Source project do *real* SQA work? How many do code reviews *before* commit? How many have comprehensive unit tests in place? How many have an accurate and correct set of design documents? Heck, how many put down more than two sentences of design before they start coding?
InstallShield has one, and only one, advantage over your typical Linux/BSD/UNIX package. In every other way it sucks. The one place it has an advantage is that it allows the user to somewhat customize the installation. On the other hand, the "less is more" GNOME folks might argue that this is a Bad Thing(tm).
Here's how a Windows user installs a program with InstallShield: Download executable. Double click on icon. Read obnoxious and threatening license screen. Click some buttons. Click, click, click. Some more clicks. Clickety clickety click. Reboot. Run program.
Here's how a typical UNIX user installs a program with a typical package manager. Download executable. Double click on icon. Run Program.
I hate InstallShield programs. Put everything into one directory tree, and zip it up. If you need to manipulate the registry, or other system wide interference, do it at first execution. Uninstall is as easy as removing that single directory. Isn't that what Mac users do? And don't people generally consider the Mac to be easier to use than Windows?
Frankly, the UNIX world doesn't need an install shield.
You laugh, but I've seen it in action. Two guys standing in front of various shrink wrapped distros in the aisle at Fry's, arguing over which one to get, finally deciding that a Distro 7.2 was better than Kernel 2.4. Sigh.
How long are people here in the US going to put up with this monkey business?
What monkey business? Let's say there's one million subscribers with 100 different programs on at a particular instance. That's a few magnitudes lower bandwidth than one million subscribers each watching a distinct program at the same time. Sending out one movie at 9:00pm on Showtime to everyone is a heck of a lot cheaper than sending out a different movie to every subscriber at the same time.
The culture of the Pacific Northwest Indians revolved around ostenatious display of wealth. Just one example out of many. Just because they didn't have money and most didn't have the concept of land as property, doesn't mean they didn't desire to accumulate wealth.
Slashdot Mirror
Or does it require a million KDE libraries to compile?
The answer is no, since there obviously aren't a million KDE libraries. It only needs two library packages: qt and kdelibs.
It seems to suggest that the only reason BSD sites are increasing is because of mainly Yahoo and a few other hosting sites who use BSD.
Since when is Yahoo a hosting site? The reason FreeBSD sites are increasing is because webhosters love FreeBSD.
Bingo!
The purpose of a file system is for peristant storage of data. The purpose of a database is for efficient organized access to data. But all too often we see database groupies advocating the filesystem-as-database, or the filesystem groupies advocating database-as-filesystem.
I use a database when I want to quickly find or correlate some data. "What was the company's net revenue from the GruntMaster2K product for April 2001?" I use a filesystem when I already know where the data is. "find" is a useful command, but I only use it about twice a week, most other times I know where the file is or can trivially browse for it, so the hierarchical structure of the filesystem is more than adequate.
Err, that would be second free operating system...
I agree that GConf was a bad example. But more than that, the whole article was "gee whiz" mental masturbation.
The problem with GConf is that it uses very many very small files. The author's solution seems to be to take this to an extreme, and make each individual key a file. Huh? Opening a file is slow. Opening one hundred files is one hundred times as slow.
You know those database programmers who think everything should be done a database, even screensavers? And those XML programmers who think everything should be done with XML, even screensavers? I almost expected this filesystem programmer to come right out and say screensavers should be done in the filesystem...
Many detractors of the UNIX security model point to the lack of ACLs and fine-grained security.
You don't throw away the door just because it doesn't have a lock. You simply put a lock on it!
FreeBSD now has ACLs, and it did it without throwing away UFS. It didn't need to replace the "everything is a file" model to do it. It just expanded the available extended attributes. To get ACLs you will certainly need to extend the filesystem, but you don't need to replace it completely.
p.s. FreeBSD also got soft updates without throwing away UFS. In short, you don't need to throw away the old to get the new, something a lot of developers don't seem to understand.
Presumably, on a machine that handles file metadata at the filesystem level, you would abandon this barbaric idea
Why is this idea barbaric? Filesystem metadata should only be for file metadata, and not content metadata. There is a crucial difference. ID3 tags are part of the content of MP3 files.
What is this obsession with reinventing filesystems? Yes, we all want blazingly fast and efficient filesystems that are self tuning. But why this rush to throw out the basic concepts that have been proven to work?
The article is a grabbag of featurettes, but with no coherent model. The classic UNIX filesystem (ufs, ffs, ext) has a simple model: everything is a file. Directories are files. Devices are files. Files are files. 99.99% of the time this is sufficient. It could use some tuning, but there is no reason to replace it with a big accretion of warts. Environment variables as files is an excellent idea. But the example of "gconf keys as files" is a wart, since gconf files are already files. If there is a pressing need for inode-free files (metadata only) then extend the existing model with inode-free files. Simple.
When you do a brainstorming session, you don't take all of the ideas on the chalkboard and roll them up into one new system. Instead you take the best ideas that work together and leave the others behind. Sometimes you even have to leave a good idea on the table. It's great that people are coming up with great ideas. But those ideas need to be fit into a working model that is lucid and elegant.
Then you're selling service, not software. Put your /etc edits on CDROM, put it on the shelf at Fry's, and no one will buy it.
The FSF groupies keep saying it's free speech not free beer. They keep saying that the GPL doesn't prevent you from selling the software. They keep saying Free Software can be commercial software.
So just sell the software! You'll go broke, but at least you can hang out with the FSF groupies in the unemployment line.
p.s. The FSF itself does NOT say this, only those who think RMS is their savior, but never bothered to read any of his writings.
p.p.s. And of course, if you're the typical slashdot reader, you still won't get it. So let me spell it out: you can't make a living selling Free Software by itself.
Say it ain't so! I've been told now for three years that it was Linux doing all of the special effects for LOTR, and now you tell me that IRIX was part of it!
Does this mean... that... Linux can't... do everything? Noooo!
Yes, monocultures are bad, and since GNU/Linux by definition can't become a monoculture (too many sources of distributions), it's a better deal than any one vendor's products.
While it's certainly better than a single-vendor monoculture, it's still a monoculture. In all probability I still won't get a choice of OS, and will probably have to pay a "Redhat tax". (ever see a Linux preload where they did NOT charge for Linux?)
1. Better tech support from ISPs, hardware vendors, PC OEMs, etc, etc etc.
"Okay bring up the DrakX and..."
"But I'm using Slackware!"
I'm not being facetious here. I use FreeBSD and I once had to lie to an ISP and tell them I was using Redhat in order to get support, because they only supported Windows, Mac and Redhat.
This is Slashdot. You know the drill. Stick your head in the sand and chant "US is Great Satan." This is the land of the double standard.
All you need to know: [US|Microsoft|RIAA|MPAA|corporations] are [evil|great satan|fascist] and only [GPL|Linux|socialism] will save us.
Because Finland under the EU is not a sovereign nation. If the two senators from Wyoming (for example) don't vote for a law, but all the other US states do, Wyoming doesn't get to opt out.
Take over the desktop. then worry about a browser.
What is this obsession you Linux guys have with taking over the desktop? How do I know the AC was a Linux guy? Simply because he was talking about taking over the dekstop!
Monocultures are bad. I don't want one bad situation with another.
For those of you asleep under a rock, let me present this "excuse" in a nutshell:
1) Someone remarks that FreeBSD and BSD/OS dominate the Netcraft uptime rankings.
2) Linux advocate offers the excuse "but the uptime counter thingy rolls over after 497 days, so the Netcraft rankings are not fair."
If you look at this thread from the beginning, you'll see that eht wasn't bitching about the Netcraft rankings, but bitching about the people bitching about Netcraft rankings. And if you had bothered reading his damn post, you would have seen that eht actually read the article, which wasn't about uptime at all!
The reason just windows is because that as much as we hate it, we are in the minority of computer uses
That doesn't stop the registrar of voters in my state from printing ballots in Tagalog, Vietnamese and Farsi.
You just have to take his extreme web-centricity into account. All applications are about data. What can you do with data? Create it, manipulate it, and retrieve it. By using the terms he did, he is arguing that the browser should be the platform (even though he says he's not).
When you're playing Quake you're most certainly not interacting with a database, but you are manipulating data.
But he's not the one using this "bug" as an excuse.
Liberals at every free-thinking college are too busy copyrighting their stream-of-consciousness rantings-as-textbook to care.
How many Open Source project do *real* SQA work? How many do code reviews *before* commit? How many have comprehensive unit tests in place? How many have an accurate and correct set of design documents? Heck, how many put down more than two sentences of design before they start coding?
InstallShield has one, and only one, advantage over your typical Linux/BSD/UNIX package. In every other way it sucks. The one place it has an advantage is that it allows the user to somewhat customize the installation. On the other hand, the "less is more" GNOME folks might argue that this is a Bad Thing(tm).
Here's how a Windows user installs a program with InstallShield: Download executable. Double click on icon. Read obnoxious and threatening license screen. Click some buttons. Click, click, click. Some more clicks. Clickety clickety click. Reboot. Run program.
Here's how a typical UNIX user installs a program with a typical package manager. Download executable. Double click on icon. Run Program.
I hate InstallShield programs. Put everything into one directory tree, and zip it up. If you need to manipulate the registry, or other system wide interference, do it at first execution. Uninstall is as easy as removing that single directory. Isn't that what Mac users do? And don't people generally consider the Mac to be easier to use than Windows?
Frankly, the UNIX world doesn't need an install shield.
You laugh, but I've seen it in action. Two guys standing in front of various shrink wrapped distros in the aisle at Fry's, arguing over which one to get, finally deciding that a Distro 7.2 was better than Kernel 2.4. Sigh.
How long are people here in the US going to put up with this monkey business?
What monkey business? Let's say there's one million subscribers with 100 different programs on at a particular instance. That's a few magnitudes lower bandwidth than one million subscribers each watching a distinct program at the same time. Sending out one movie at 9:00pm on Showtime to everyone is a heck of a lot cheaper than sending out a different movie to every subscriber at the same time.
Native americans weren't doing it either.
The culture of the Pacific Northwest Indians revolved around ostenatious display of wealth. Just one example out of many. Just because they didn't have money and most didn't have the concept of land as property, doesn't mean they didn't desire to accumulate wealth.