Back in the good ol' days of the 1800's, people would form Political Parties to deal with this overwhelming opposition, even in the sense that the Democrats and Republicans are two different parties.
Essentially both want the same thing, just implemented in two different ways.
Now, in the United States of Complacency, we're going to be stuck with a central government completely ruled by one oppressive political power. Liberals nor Conservatives will be able to save us from what's coming.
This nation needs a "proletariat" party; a party who represents the average American and not the corporate entity that America has become internally, and in the face of the rest of the world. I think in the wake of this natural disaster and the inability for the government to even get up off its ass and do anything about it for five days should be quite the impetuous to start such a party.
I'd like to end with a famous slashdot sig: "There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order." the ballot box isn't working anymore. Time to bring out the ammo.
Because the Supreme Court is more important than the President and Congress.
Um, what? Just because the Supreme Court stays in office longer than both the president and all of congress, doesn't make them any more powerful/important than the rest of the system.
The Supreme Court is a balance against the law passing powers of the Congress and the President's veto power. The whole system works because one unit in the system balances out the other.
One thing that's happening right now, though, is a lot of power is getting shifted to the President. And now that the President can replace two Supreme Court officials, I'm sure you will continue to see a power shift away from the court system and towards the legislator and President in the years to come. I'll hold my hopes against it, but my better judgement tells me that it will happen regardless.
So while the Supreme Court may be the last stop for a law, they are only as strong as the rest of the government. Hopefully the new justices will notice this, and the balance of power will restore itself.
PayPal's done this before to SA for other reasons as well.. it's fucking unacceptable.
And yes, the SA servers are in the same facility as the DirectNIC building, just a floor below. Yesterday they finally got the bandwidth to post a temporary homepage for those of us seeking refuge in Slashdot for the time being.
I hope the SA community stands with the rest of us who never use PayPal. This is fucking bullshit.
..that will never make it to human trials in America. Reason? It's another one of those taboo research topics; it's fine and dandy to clone a sheep or a mouse, it's fine to use crocodiles to fight HIV, it's acceptible to take a look at the human genes for eye-color and hair color, but the minute you even mention any of these actually going into clinical trials, or even attempting to get government funding, and you're shutdown for life.
The research climate in this country's starting to get ridiculous. We hear about all of these new advances almost daily in the news, but we're still waiting to see any practical use come from them. These are things that save lives, things that make terrible diseases easier to fight.
I know if I lost an arm or a leg or more importantly a heart or lung, I'd love the ability to grow one back..
Eh, it's all just signs of Microsoft cracking. Right now it's running around in so many directions, trying to do so many things that one side of Microsoft can't tell what the other's doing.
One section of Microsoft is trying to find a way to diversify into other fields (as it always has been). This means as soon as anything gets popular, instantly releasing that they will have a competitor to that product. See previous articles..
The next section of Microsoft is designing Vista. More or less, they're looking over at Apple and saying "hmm, now how do we do this for ourselves". Hey, if you're going to copy, make sure you copy from the best.
Next, Microsoft's patent team is doing everything they can to churn out as many patents for as many things as possible, no matter what relevance they have to anything. Patents are the new gold; having them makes you rich, no matter in what shape, color, or form.
Then you have the Microsoft gaming committee putting together the XBox 360.. Good luck with that xboxers.
And then you end up with the "future of technology" department; the one where they write all of these magnificent things, designing things like Palladium and giving them crazy names. The only problem is, while this section's doing the designing, all of the other sections of Microsoft are doing their own thing; it seems as if there isn't any communication in the entire process.
Microsoft is like a three hundred pound kid on a tricycle on a very big hill. They've got a lot of business henged on a small amount of products, and they've got to ensure that these products don't collaspe. And the best way of doing that is Advertising, the media, product placement, and the public (get the picture yet? good). The more of these documents coming out that don't mean anything at all, the more Microsoft looks like it's doing something.
Now THAT is disaster aid. Thank you Cingular, you're doing a great service to the community.
T-Mobile, take a look at Cingular; they're putting you to shame. All people want to do is to tell their loved ones that they are alright. The Internet may have that capacity (if they can find their loved ones), but is a terrible medium in which to have to search, or email and hope. Phone calls will always be better in that respect.
I'm glad to hear my phone company's doing some good for a change. Money well spent I hope.
What about phone cards? T-Mobile's a phone company, they can give away phone card numbers to people so they can get ahold of loved ones, even if it's just 20 minutes; long enough to say "Hi Mom, I'm still alive, I love you." Besides, phone service is much more ubiquitous than Internet service, and phone cards are cheap.
It really feels like a promotion instead of general charity. I mean come on, WiFi? How many people would even be able to use this service verses general telephony, or even handing out "disposable cellphones" (those prepaid things you can find at any gas station)?
Lastly, giving blood is something you should be doing anyways, if you can. Disasters shouldn't make you wake up and realize the need for blood, and I really wish services like the Red Cross would get that message out.
Give it all time. I remember it took about a day before Apple's page was updated with the Tsunami information, I would suggest it might take a bit longer in this case as things are just now starting to get organized in the disaster relief front.
But, I'm glad others are donating as well. Well worth the hundred twenty bucks left at the bottom of the bank account.
From CNN, I believe they are providing this service to try to help find loved ones; earlier I was listening in, and it sounds like tonnes of E-mail services, and website forums are being flooded by people looking for their families online.
If T-Mobile set up a centralized "find-me" server, drove around with some laptops and actually tried to lend a helping hand, I'd love to support them, but yeah, at this point, it does look like a PR move.
Hey! You forgot the camara. God I love that feature!</sarcasm>
I dunno if that's a camera or a camaro, but hey, I'll take my photo-taking, fast-moving, phone-calling machine any day. Now if only my phone had an electric razor..
It's probably a port of the iPod interface in software, with the iPod shuffle's hardware player, tied into Motorola's cellphone. Nothing too complicated really, and it fits with the need for Apple to get more flash ram, now (as moving parts plus cellphones equals dead device).
I'll still get one, even if it means my Treo 600 will have to take the backseat for a while.
For one, you're probably just arguing just to be arguing, as most people here on Slashdot do, but I thought since I gave SuperKendal a try at responding to all of this million holes in my argument, I would at least give you a fair chance. Forgive me if I butcher your arguments a bit; I simply don't care enough to leave everything in tact after a 20 hour shift.
For one, the database is implemented as a byte-heap; to the user, it really doesn't look all that different. You have a heirarchy, but instead of manually building it, it's a series of SQL queries. That way, instead of making a bunch of new folders, moving everything around, and then realizing that you want to change something, it's simply a change of a few characters in an SQL query, and your folders work for you. It's not that hard to understand.
As for your biggest question; why do you need to remove the current filesystem to do this: you don't, it just helps. Apple's version works fairly well, but when I go to save a file, I still have to choose a folder and hit save. I still have to deal with decompressed, flat-file configurations instead of storing all of this information in a database, transparently compressed. I still have to deal with file fragmentation, file location, dot files (UNIX hidden files), etc. Under my implementation, a lot of these unnessicary bits are removed, and a cleaner implementation takes its place. Oh, and if you know anything about filesystems, you know that most typical file systems block-allocate files, which means that a 2 byte text file takes up 4 kilobytes. At first you think "eh, this doesn't matter", but then you realize how fast it adds up; of the nearly million files currently on my RAID server taking up 263.32 gigabytes, roughly six gigabytes of that are files exactly 32kb in size (due to the hugeness of the FAT), while most of those files actually don't use but 4kb of that. Sure, I could compress them all, but then I can't search them as I need to be able to.
While even I admit that this is a small annoyance, it's one that is completely dissolved with a database file system. Media that can be compressed, is, and is still fastly searchable. Data that is mutable is in a mutable table and has fast access writes and reads. Data that are media files (especially pictures, mp3s, and movies) are stored in tables that are immutable and fixed-size, autocategorized by their respective header informations, and files that are compressed are now searchable compressed files. Files that are executable are protected from the rest of the files. Files that are system files are hidden from non-power users. ACLs are fast generated by database permissions. Authentication is fed at a database-login level. The file system is easily spanned across volumes, computers, even the Internet, without nasty implemenations involving FTP or WebDAV. The list of advantages goes on, and on and on. The list of disadvantages in my book at least are very short, and I can deal: the database needs a few more CPU cycles to index information. Crawlers need CPU cycles. It's not very portable; you need an implementation in every operating system, or a very good database browser. The data is more vulnerable to failures as it is more densely concentrated.
Well, that's about all I've got to say. I'm gonna move on now.
Dude, I wish I could get away with buying my girlfriend a dirty, black, carbon based rock for a ring. You must have a really accepting significant other:(
Um, no. Diamonds currently retain value as expensive the same way Oil does. It's controlled by a company who's got overwhelming control over the supply, and thus, can charge any price they want for the goods.
That being said, synthetic diamonds have been on the market for a while now. In fact, my sister just bought a ring with one in it.
No. I mean you shouldn't have to. Just as I wouldn't have to spell things incorrectly if all text boxes had a default spellchecker in Linux as they do in Mac OS X.
Instead, try "find *filename*" or "find file associated with *filename* by *date of creation*" or "find file associated with *username*", or "find file with name like *blah*, that is a *filetype*, and was created by user group *administrators*". Try THAT with your current Linux terminal. It'll bark a bunch of program errors and junk. Yet all of those find functions above are ways I feel you should be able to find files.
Co-gawd, you outdate me. I had to do some pretty creative googling just to find out what the heck you were talking about, and I'm quite impressed. Why this never made it as an idea, I'll never know, and be sad because of. But I can assure you, I never meant to invent anything, those days have come and gone. These days I just want something that makes my life easier, and that gets rid of all those headaches I get when I look at the number of harddisks, in the number of machines, the number of different pieces of media, and the number of different media devices I have. Now that we're talking about terabyte-size devices, now that I have a cluster with a 2 terabyte array, I want to make sure that when I put files on it, I'll never lose them due to accidental deletion ("Hmm, nothing I need in this folder, *delete* OH SHIT NOT MY ORIGINAL HOT GRITS PIX!!!") (yes, the array is a raid 10 array; two teras is the physical size).
Ok, fine... you have just heaps of data, with a myriad of references to them.
Sure, but explicit references are something we're trying not to deal with, as we've passed that age IMO.
What then is delete? How does a user distinguish between "remove an association from the blob of data" vs "remove this blob of data altogether". Should the blob automatically delete when you remove all metadata around it? If not, how will you find it again? If so, would you really want data vanishing just because you removed a keyword?
Delete is delete. Remove link is remove link. These kinds of things are relatively easy to understand, even for the layman. Nothing should automatically remove just because you remove something attached to it. If you're looking to remove it, you click delete and it's gone. If you're looking to remove an association to it you choose "disassociate" or "remove link" or whatever your GUI might implement it as. As this currently doesn't exist (as links are represented as files), we don't have a way of explaining this as of yet. How will you find a file if you break a namespace association? Easy; search for the name. You never want to remove all associations, and as the file is stored in a database, the fileid should still be around, and relatively easy to find. This is really a null argument.
What does partial backup look like on a system? How can you have a combination of partial backups and know you have a whole? I can do that with a set of five directories. Let's say you tag a set of files with "project fred". But one small file, that you almost never care about, gets tagged with "project ferd". What good is the ol' Fred backup now?
Wow, as a typical slashdotter is an expert in absolutely everything, you obviously aren't an expert at database designs. In databases we have transactions that allow us to roll forward and backward, and we have differential backup systems that work just as they do on filesystems. Lastly, partial backups are really not where you want to be if you're implementing a system like this. Partials really work best on text files, and as textfiles would hopefully be stored in a seperate database, maybe even with some kind of compression, all you have to do is dump it, and do a binary compare with your past dump. If you keep the dumps in order, then you'll always have a full copy. Backups are really a null point; we've learned how to do backups of databases since the 80's
At some core level these blobs of data that users place on a system need ONE meaningful location where they always "are". You need someplace where the file will always be, no matter what other associations you remove. You need somewhere you know it will be to assure yourself EVERYTHING you care about is backed up or moved between systems.
Well, as I recall correctly, there is one central, meaningful location to where all of the files are located. It's call the hard disk they're stored on. This is the root level; everything's right there laid out in bits, right in order. All my system would do is remove the whole idea of "folders" or locations within locations of files. It would then replace that idea with the idea that you can group files by how they are alike and disalike. This shifts the focus of how the files are organized away from the disk, and into the software, and as computers are fast enough to deal with this shift, and now that humans are vastly more accepting to this shift (as our media devices often do this, as our programs often do this kind of abstraction... it's really actually amazing how this kind of abstraction has been invented and reinvented every time someone comes up with a program), the time is now to implement it at an operating system level, instead of an application level.
The perfection you seek can just as easily be obtained with files in directories that allow metadata on top of them and things like smart folders that are essentially queries over the user-defined and automatically ext
Why does a human have to enter the metadata? Why not let the machine do what it can to derive what the file is about, and ask the human whether or not it's right? Such a system can be taught to "learn" when it's right or wrong, and it'd get better with time.
People can't update their ID3 tags. But they can download a program like MusicBrainz which is a database of ID3 tags that those of us with the time do things right.
To be truthful, the worse metadata that exists today in my opinion is that of Photos, and that of Movies. Everyone's got a solution out there that works for music, but that solution disentegrates when you point it at a collection of pictures, or a bunch of movies. It's also hard (but not impossible) to write a program which can look at a movie and tell you what movie it is, or look at a bunch of pictures and based on previous experience, tell you who's in that picture, or what that picture is of. That makes visual programs distinctly harder to categorize, verses a machine which can listen to music and instantly identify it.
Lastly, your example more or less proves my point. Look at all of those links. Why do you need a link that points at a file, ever? If you have the damned file, you don't need someone pointing at it and saying "there it is!", you need the software saying "here's your file sir".
Um.. why? Why do you need name your file with six hundred underscores, don't spaces still work? Why do you even need to use that long of a file name; if you use a unique id system, you can associate a file to a program, and thus, give the file a whole seperate namespace to work within. Then add a session namespace identifier, and you're done. Of course, this will be hidden from you within the database, but it will prevent namespace collision. it works off of this idea that humans have had for a long time called "relevance" and "situational awareness".
If you find your current situation ideal, than this system is not for you, and it's easy enough to disable. Hell, it'll even save you time because you won't have to run it, thus reducing the overhead on your computer. But for those of us with literally millions of files to index of different types dating back ten years, it's distinctly difficult.
Lastly, I'm a man of the GUI situation. I can't stand "locate filename | grep "s0m#thin%^9&%&3". It looks horrible it's hard to remember the different syntaxes, and it's terrible for new users. And I can't even mention how stupid the names are; "rm" for "remove"? "cp" for "copy"? These are short words people.. not that difficult to remember. Oh, I have a solution! Make a symlink from cp to copy and everything's fine, right? Wrong; only works on your system, new users get confused when they go to other machines, et cetera.
It's time we grow up. Some of us want to live in a GUI world, and for those of us who do, we shouldn't be held back by those of us who are too unwilling to learn something new and easy. But, in this environment, this kind of arguement is a troll, and falls on completely deaf ears.
Well, I didn't say WinFS could do it, but I did say a database file system could.
What a database filesystem can do is stick a file right against another file, without the otherwise nessicary buffer space, without fragmentation, and without needing to have the files laid out in blocks.
If a program needs to be able to dynamically change the size of the file, keep that file in a seperate index until the point in which the program terminates, or the file is deemed static once again. I would implement this as having a "Mutable" flag in the database.
As the typical person doesn't have a file that changes dynamically other than log files, the system works. For a good example, media files would rarely need a mutable flag, thus can be stored in a more tightly aligned table. As for text files, they could even be stored in a seperate database, one that I would personally have automatically and transparently compressed.
But these are all really implemetation-specific things. Perhaps I should put my database-filesystem report online somewhere..
Beagle is a carbon-copy of Apple's Spotlight, which I noted in my original post. As far as I can tell, inotify was added to the kernel for the explicit purpose of allowing something like this to be created.
That being said, I cannot solicit Beagle, as much as it is a part of GNOME. First of all, it's written in C#, which I am against, but even averting that point, Beagle is slow, it's very, very buggy, and for some insane reason, they decided to go with Lucene as an Index server, instead of a fully qualified SQL server which could be connected through ODBC or any other database abstraction method.
I've said these things before and been modified as troll, with people responding with "if you could do it better, do it yourself". Well, this isn't my capacity at this point in time; I'm simply observing and reporting on the product. I understand that it's deep in alpha right now, and I do have hope that it'll get better, but in the meantime, it's connection to C#, Lucene, and fundamental archetecture problems as to where the program is allowed to index makes me doubt it's future relevance.
My point is that we need a database file system, but that Linux as a whole will be in last place to get one. Beagle is a good attempt, but I can't see it as anything more than a graduate project. I offered to port it to C++, a database agnostic implementation, and to add Kerberos/PAM support to it as my Google Summer-of-Code entry, but as I was declined, and because I do need to stay alive and eat, I can't just code it for free.
Um, no offense, but GNOME Storage was CRAP. I've been following Database filesystems for a long, long time, and have worked out a number of different implementation-schemas on paper, and I have to tell you, the way that GNOME storage was going about things was entirely bogus.
The sytsem worked off the idea of installing a CORBA orb in the kernel to communicate back to userspace, where the query utility was located. This has advantages, but the enormous, gigantic disadvantage of having to have a CORBA orb inside of the kernel, and being dependent on this orb to keep up with the kernel's development. This of course, didn't happen, and development stagnated on this particular project.
After a while of working, I decided it wasn't worth my time to implement a database file system simply because Apple's Spotlight was almost exactly the system I figured would work best; a scanning indexer that would find all of the file information and put it inside of a database, leaving the files around the disk where they were located in the first place. This would require less hacking, less re-developing of the whoe UNIX virtual filesystem, etc. etc.
To be truthful, I really wouldn't mind developing the filesystem, but the Linux kernel makes it a pain really; it's a very fast moving target to aim at, so many other filesystems depend on the virtual filesystem staying the way that it is currently. But additions like inotify will definitely help in this area, but it'd still be a lot of work.
Slashdot Mirror
Back in the good ol' days of the 1800's, people would form Political Parties to deal with this overwhelming opposition, even in the sense that the Democrats and Republicans are two different parties. Essentially both want the same thing, just implemented in two different ways.
Now, in the United States of Complacency, we're going to be stuck with a central government completely ruled by one oppressive political power. Liberals nor Conservatives will be able to save us from what's coming.
This nation needs a "proletariat" party; a party who represents the average American and not the corporate entity that America has become internally, and in the face of the rest of the world. I think in the wake of this natural disaster and the inability for the government to even get up off its ass and do anything about it for five days should be quite the impetuous to start such a party.
I'd like to end with a famous slashdot sig: "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." the ballot box isn't working anymore. Time to bring out the ammo.
Because the Supreme Court is more important than the President and Congress.
Um, what? Just because the Supreme Court stays in office longer than both the president and all of congress, doesn't make them any more powerful/important than the rest of the system.
The Supreme Court is a balance against the law passing powers of the Congress and the President's veto power. The whole system works because one unit in the system balances out the other.
One thing that's happening right now, though, is a lot of power is getting shifted to the President. And now that the President can replace two Supreme Court officials, I'm sure you will continue to see a power shift away from the court system and towards the legislator and President in the years to come. I'll hold my hopes against it, but my better judgement tells me that it will happen regardless.
So while the Supreme Court may be the last stop for a law, they are only as strong as the rest of the government. Hopefully the new justices will notice this, and the balance of power will restore itself.
Oh the irony, my account's locked for contributing to SA's account. I think I'm going to give my lawyer a ring this afternoon...
PayPal's done this before to SA for other reasons as well.. it's fucking unacceptable.
And yes, the SA servers are in the same facility as the DirectNIC building, just a floor below. Yesterday they finally got the bandwidth to post a temporary homepage for those of us seeking refuge in Slashdot for the time being.
I hope the SA community stands with the rest of us who never use PayPal. This is fucking bullshit.
Screenshots? Zonk, you shouldn't have.. when did slashdot get the bandwidth for that!?
how does a lawsuit that hasn't even yet taken place have a length value, in pages?
..that will never make it to human trials in America. Reason? It's another one of those taboo research topics; it's fine and dandy to clone a sheep or a mouse, it's fine to use crocodiles to fight HIV, it's acceptible to take a look at the human genes for eye-color and hair color, but the minute you even mention any of these actually going into clinical trials, or even attempting to get government funding, and you're shutdown for life.
The research climate in this country's starting to get ridiculous. We hear about all of these new advances almost daily in the news, but we're still waiting to see any practical use come from them. These are things that save lives, things that make terrible diseases easier to fight.
I know if I lost an arm or a leg or more importantly a heart or lung, I'd love the ability to grow one back..
Eh, it's all just signs of Microsoft cracking. Right now it's running around in so many directions, trying to do so many things that one side of Microsoft can't tell what the other's doing.
One section of Microsoft is trying to find a way to diversify into other fields (as it always has been). This means as soon as anything gets popular, instantly releasing that they will have a competitor to that product. See previous articles..
The next section of Microsoft is designing Vista. More or less, they're looking over at Apple and saying "hmm, now how do we do this for ourselves". Hey, if you're going to copy, make sure you copy from the best.
Next, Microsoft's patent team is doing everything they can to churn out as many patents for as many things as possible, no matter what relevance they have to anything. Patents are the new gold; having them makes you rich, no matter in what shape, color, or form.
Then you have the Microsoft gaming committee putting together the XBox 360.. Good luck with that xboxers.
And then you end up with the "future of technology" department; the one where they write all of these magnificent things, designing things like Palladium and giving them crazy names. The only problem is, while this section's doing the designing, all of the other sections of Microsoft are doing their own thing; it seems as if there isn't any communication in the entire process.
Microsoft is like a three hundred pound kid on a tricycle on a very big hill. They've got a lot of business henged on a small amount of products, and they've got to ensure that these products don't collaspe. And the best way of doing that is Advertising, the media, product placement, and the public (get the picture yet? good). The more of these documents coming out that don't mean anything at all, the more Microsoft looks like it's doing something.
Now THAT is disaster aid. Thank you Cingular, you're doing a great service to the community.
T-Mobile, take a look at Cingular; they're putting you to shame. All people want to do is to tell their loved ones that they are alright. The Internet may have that capacity (if they can find their loved ones), but is a terrible medium in which to have to search, or email and hope. Phone calls will always be better in that respect.
I'm glad to hear my phone company's doing some good for a change. Money well spent I hope.
What about phone cards? T-Mobile's a phone company, they can give away phone card numbers to people so they can get ahold of loved ones, even if it's just 20 minutes; long enough to say "Hi Mom, I'm still alive, I love you." Besides, phone service is much more ubiquitous than Internet service, and phone cards are cheap.
It really feels like a promotion instead of general charity. I mean come on, WiFi? How many people would even be able to use this service verses general telephony, or even handing out "disposable cellphones" (those prepaid things you can find at any gas station)?
Lastly, giving blood is something you should be doing anyways, if you can. Disasters shouldn't make you wake up and realize the need for blood, and I really wish services like the Red Cross would get that message out.
Give it all time. I remember it took about a day before Apple's page was updated with the Tsunami information, I would suggest it might take a bit longer in this case as things are just now starting to get organized in the disaster relief front.
But, I'm glad others are donating as well. Well worth the hundred twenty bucks left at the bottom of the bank account.
From CNN, I believe they are providing this service to try to help find loved ones; earlier I was listening in, and it sounds like tonnes of E-mail services, and website forums are being flooded by people looking for their families online.
If T-Mobile set up a centralized "find-me" server, drove around with some laptops and actually tried to lend a helping hand, I'd love to support them, but yeah, at this point, it does look like a PR move.
Hey! You forgot the camara. God I love that feature!</sarcasm>
I dunno if that's a camera or a camaro, but hey, I'll take my photo-taking, fast-moving, phone-calling machine any day. Now if only my phone had an electric razor..
Hah, I honestly doubt it'll have any of that.
It's probably a port of the iPod interface in software, with the iPod shuffle's hardware player, tied into Motorola's cellphone. Nothing too complicated really, and it fits with the need for Apple to get more flash ram, now (as moving parts plus cellphones equals dead device).
I'll still get one, even if it means my Treo 600 will have to take the backseat for a while.
For one, you're probably just arguing just to be arguing, as most people here on Slashdot do, but I thought since I gave SuperKendal a try at responding to all of this million holes in my argument, I would at least give you a fair chance. Forgive me if I butcher your arguments a bit; I simply don't care enough to leave everything in tact after a 20 hour shift.
For one, the database is implemented as a byte-heap; to the user, it really doesn't look all that different. You have a heirarchy, but instead of manually building it, it's a series of SQL queries. That way, instead of making a bunch of new folders, moving everything around, and then realizing that you want to change something, it's simply a change of a few characters in an SQL query, and your folders work for you. It's not that hard to understand.
As for your biggest question; why do you need to remove the current filesystem to do this: you don't, it just helps. Apple's version works fairly well, but when I go to save a file, I still have to choose a folder and hit save. I still have to deal with decompressed, flat-file configurations instead of storing all of this information in a database, transparently compressed. I still have to deal with file fragmentation, file location, dot files (UNIX hidden files), etc. Under my implementation, a lot of these unnessicary bits are removed, and a cleaner implementation takes its place. Oh, and if you know anything about filesystems, you know that most typical file systems block-allocate files, which means that a 2 byte text file takes up 4 kilobytes. At first you think "eh, this doesn't matter", but then you realize how fast it adds up; of the nearly million files currently on my RAID server taking up 263.32 gigabytes, roughly six gigabytes of that are files exactly 32kb in size (due to the hugeness of the FAT), while most of those files actually don't use but 4kb of that. Sure, I could compress them all, but then I can't search them as I need to be able to.
While even I admit that this is a small annoyance, it's one that is completely dissolved with a database file system. Media that can be compressed, is, and is still fastly searchable. Data that is mutable is in a mutable table and has fast access writes and reads. Data that are media files (especially pictures, mp3s, and movies) are stored in tables that are immutable and fixed-size, autocategorized by their respective header informations, and files that are compressed are now searchable compressed files. Files that are executable are protected from the rest of the files. Files that are system files are hidden from non-power users. ACLs are fast generated by database permissions. Authentication is fed at a database-login level. The file system is easily spanned across volumes, computers, even the Internet, without nasty implemenations involving FTP or WebDAV. The list of advantages goes on, and on and on. The list of disadvantages in my book at least are very short, and I can deal: the database needs a few more CPU cycles to index information. Crawlers need CPU cycles. It's not very portable; you need an implementation in every operating system, or a very good database browser. The data is more vulnerable to failures as it is more densely concentrated.
Well, that's about all I've got to say. I'm gonna move on now.
Dude, I wish I could get away with buying my girlfriend a dirty, black, carbon based rock for a ring. You must have a really accepting significant other :(
Um, no. Diamonds currently retain value as expensive the same way Oil does. It's controlled by a company who's got overwhelming control over the supply, and thus, can charge any price they want for the goods.
That being said, synthetic diamonds have been on the market for a while now. In fact, my sister just bought a ring with one in it.
No. I mean you shouldn't have to. Just as I wouldn't have to spell things incorrectly if all text boxes had a default spellchecker in Linux as they do in Mac OS X.
Instead, try "find *filename*" or "find file associated with *filename* by *date of creation*" or "find file associated with *username*", or "find file with name like *blah*, that is a *filetype*, and was created by user group *administrators*". Try THAT with your current Linux terminal. It'll bark a bunch of program errors and junk. Yet all of those find functions above are ways I feel you should be able to find files.
Co-gawd, you outdate me. I had to do some pretty creative googling just to find out what the heck you were talking about, and I'm quite impressed. Why this never made it as an idea, I'll never know, and be sad because of. But I can assure you, I never meant to invent anything, those days have come and gone. These days I just want something that makes my life easier, and that gets rid of all those headaches I get when I look at the number of harddisks, in the number of machines, the number of different pieces of media, and the number of different media devices I have. Now that we're talking about terabyte-size devices, now that I have a cluster with a 2 terabyte array, I want to make sure that when I put files on it, I'll never lose them due to accidental deletion ("Hmm, nothing I need in this folder, *delete* OH SHIT NOT MY ORIGINAL HOT GRITS PIX!!!") (yes, the array is a raid 10 array; two teras is the physical size).
Ok, fine... you have just heaps of data, with a myriad of references to them.
Sure, but explicit references are something we're trying not to deal with, as we've passed that age IMO.
What then is delete? How does a user distinguish between "remove an association from the blob of data" vs "remove this blob of data altogether". Should the blob automatically delete when you remove all metadata around it? If not, how will you find it again? If so, would you really want data vanishing just because you removed a keyword? Delete is delete. Remove link is remove link. These kinds of things are relatively easy to understand, even for the layman. Nothing should automatically remove just because you remove something attached to it. If you're looking to remove it, you click delete and it's gone. If you're looking to remove an association to it you choose "disassociate" or "remove link" or whatever your GUI might implement it as. As this currently doesn't exist (as links are represented as files), we don't have a way of explaining this as of yet. How will you find a file if you break a namespace association? Easy; search for the name. You never want to remove all associations, and as the file is stored in a database, the fileid should still be around, and relatively easy to find. This is really a null argument.
What does partial backup look like on a system? How can you have a combination of partial backups and know you have a whole? I can do that with a set of five directories. Let's say you tag a set of files with "project fred". But one small file, that you almost never care about, gets tagged with "project ferd". What good is the ol' Fred backup now?
Wow, as a typical slashdotter is an expert in absolutely everything, you obviously aren't an expert at database designs. In databases we have transactions that allow us to roll forward and backward, and we have differential backup systems that work just as they do on filesystems. Lastly, partial backups are really not where you want to be if you're implementing a system like this. Partials really work best on text files, and as textfiles would hopefully be stored in a seperate database, maybe even with some kind of compression, all you have to do is dump it, and do a binary compare with your past dump. If you keep the dumps in order, then you'll always have a full copy. Backups are really a null point; we've learned how to do backups of databases since the 80's
At some core level these blobs of data that users place on a system need ONE meaningful location where they always "are". You need someplace where the file will always be, no matter what other associations you remove. You need somewhere you know it will be to assure yourself EVERYTHING you care about is backed up or moved between systems.
Well, as I recall correctly, there is one central, meaningful location to where all of the files are located. It's call the hard disk they're stored on. This is the root level; everything's right there laid out in bits, right in order. All my system would do is remove the whole idea of "folders" or locations within locations of files. It would then replace that idea with the idea that you can group files by how they are alike and disalike. This shifts the focus of how the files are organized away from the disk, and into the software, and as computers are fast enough to deal with this shift, and now that humans are vastly more accepting to this shift (as our media devices often do this, as our programs often do this kind of abstraction... it's really actually amazing how this kind of abstraction has been invented and reinvented every time someone comes up with a program), the time is now to implement it at an operating system level, instead of an application level.
The perfection you seek can just as easily be obtained with files in directories that allow metadata on top of them and things like smart folders that are essentially queries over the user-defined and automatically ext
Why does a human have to enter the metadata? Why not let the machine do what it can to derive what the file is about, and ask the human whether or not it's right? Such a system can be taught to "learn" when it's right or wrong, and it'd get better with time.
People can't update their ID3 tags. But they can download a program like MusicBrainz which is a database of ID3 tags that those of us with the time do things right.
To be truthful, the worse metadata that exists today in my opinion is that of Photos, and that of Movies. Everyone's got a solution out there that works for music, but that solution disentegrates when you point it at a collection of pictures, or a bunch of movies. It's also hard (but not impossible) to write a program which can look at a movie and tell you what movie it is, or look at a bunch of pictures and based on previous experience, tell you who's in that picture, or what that picture is of. That makes visual programs distinctly harder to categorize, verses a machine which can listen to music and instantly identify it.
Lastly, your example more or less proves my point. Look at all of those links. Why do you need a link that points at a file, ever? If you have the damned file, you don't need someone pointing at it and saying "there it is!", you need the software saying "here's your file sir".
Um.. why? Why do you need name your file with six hundred underscores, don't spaces still work? Why do you even need to use that long of a file name; if you use a unique id system, you can associate a file to a program, and thus, give the file a whole seperate namespace to work within. Then add a session namespace identifier, and you're done. Of course, this will be hidden from you within the database, but it will prevent namespace collision. it works off of this idea that humans have had for a long time called "relevance" and "situational awareness".
If you find your current situation ideal, than this system is not for you, and it's easy enough to disable. Hell, it'll even save you time because you won't have to run it, thus reducing the overhead on your computer. But for those of us with literally millions of files to index of different types dating back ten years, it's distinctly difficult.
Lastly, I'm a man of the GUI situation. I can't stand "locate filename | grep "s0m#thin%^9&%&3". It looks horrible it's hard to remember the different syntaxes, and it's terrible for new users. And I can't even mention how stupid the names are; "rm" for "remove"? "cp" for "copy"? These are short words people.. not that difficult to remember. Oh, I have a solution! Make a symlink from cp to copy and everything's fine, right? Wrong; only works on your system, new users get confused when they go to other machines, et cetera.
It's time we grow up. Some of us want to live in a GUI world, and for those of us who do, we shouldn't be held back by those of us who are too unwilling to learn something new and easy. But, in this environment, this kind of arguement is a troll, and falls on completely deaf ears.
*sigh* back to my Mac.
Well, I didn't say WinFS could do it, but I did say a database file system could.
What a database filesystem can do is stick a file right against another file, without the otherwise nessicary buffer space, without fragmentation, and without needing to have the files laid out in blocks.
If a program needs to be able to dynamically change the size of the file, keep that file in a seperate index until the point in which the program terminates, or the file is deemed static once again. I would implement this as having a "Mutable" flag in the database.
As the typical person doesn't have a file that changes dynamically other than log files, the system works. For a good example, media files would rarely need a mutable flag, thus can be stored in a more tightly aligned table. As for text files, they could even be stored in a seperate database, one that I would personally have automatically and transparently compressed.
But these are all really implemetation-specific things. Perhaps I should put my database-filesystem report online somewhere..
Beagle is a carbon-copy of Apple's Spotlight, which I noted in my original post. As far as I can tell, inotify was added to the kernel for the explicit purpose of allowing something like this to be created.
That being said, I cannot solicit Beagle, as much as it is a part of GNOME. First of all, it's written in C#, which I am against, but even averting that point, Beagle is slow, it's very, very buggy, and for some insane reason, they decided to go with Lucene as an Index server, instead of a fully qualified SQL server which could be connected through ODBC or any other database abstraction method.
I've said these things before and been modified as troll, with people responding with "if you could do it better, do it yourself". Well, this isn't my capacity at this point in time; I'm simply observing and reporting on the product. I understand that it's deep in alpha right now, and I do have hope that it'll get better, but in the meantime, it's connection to C#, Lucene, and fundamental archetecture problems as to where the program is allowed to index makes me doubt it's future relevance.
My point is that we need a database file system, but that Linux as a whole will be in last place to get one. Beagle is a good attempt, but I can't see it as anything more than a graduate project. I offered to port it to C++, a database agnostic implementation, and to add Kerberos/PAM support to it as my Google Summer-of-Code entry, but as I was declined, and because I do need to stay alive and eat, I can't just code it for free.
Um, no offense, but GNOME Storage was CRAP. I've been following Database filesystems for a long, long time, and have worked out a number of different implementation-schemas on paper, and I have to tell you, the way that GNOME storage was going about things was entirely bogus.
The sytsem worked off the idea of installing a CORBA orb in the kernel to communicate back to userspace, where the query utility was located. This has advantages, but the enormous, gigantic disadvantage of having to have a CORBA orb inside of the kernel, and being dependent on this orb to keep up with the kernel's development. This of course, didn't happen, and development stagnated on this particular project.
After a while of working, I decided it wasn't worth my time to implement a database file system simply because Apple's Spotlight was almost exactly the system I figured would work best; a scanning indexer that would find all of the file information and put it inside of a database, leaving the files around the disk where they were located in the first place. This would require less hacking, less re-developing of the whoe UNIX virtual filesystem, etc. etc.
To be truthful, I really wouldn't mind developing the filesystem, but the Linux kernel makes it a pain really; it's a very fast moving target to aim at, so many other filesystems depend on the virtual filesystem staying the way that it is currently. But additions like inotify will definitely help in this area, but it'd still be a lot of work.