The goal of the next version (as I understand it) is to allow virtual folders, so that you can search everything with a common set of tools.
Reminds me of Reiser4. The principle, at least, sounds the same; In Reiser4, files can act as directories, their behaviour controlled by plugins in the filesystem. This means you can, just as an example, do things like cat/etc/passwd/root and if a plugin knows how to parse the passwd file, you get:
root:x:0:0:root:/root:/bin/bash (or something like this)
That said, I prefer your approach myself; Make the kernel and filesystem driver a simple system that the user very rarely needs to interface with directly, and put virtualisation into applications and libraries. It's more flexible and more portable this way.
Slashdot doesn't use well standardised XHTML/CSS, either. Nor does Slash, the code back-end, use many good programming principles that any Slashdot editor and most users would advocate - or at least so I've been told. Nor do a great many Slashdotters use an open-source OS (remember that poll about operating systems a while ago, some people still have the more shocking parts of the results in their sig)
Ergo, saying "Slashdot is not using it" is not saying much at all...
Gives a whole new meaning to karma whoring, though. Perhaps this should be called karma masturbation;) ?
Re:Definition of Developer
on
Halloween VII
·
· Score: 1
I use Debian GNU/Linux (and I like the GNU/ part there) exclusively at home. While I know a little C, PERL and PHP (none of them well), I have never written any software longer than a thousand lines (except years ago in BASIC, and that was mostly lists of stuff) or submitted a patch to any project. Sure, in time I might become a developer, but at least right now I just classify myself as an IT dabbler. So yes, such people exist who can use software and know what they are doing but can't very realistically contribute.
Re:Variety of standards
on
Gnutella2?
·
· Score: 2, Insightful
Remember one fact about real P2P networks: They have a lot of overhead for searching and, in the case of Gnutella, pinging/ponging. IM networks are centralised enough to not have that overhead, so being logged in to several of them doesn't use that much bandwidth. But unless you have insane amounts of bandwidth or are going to be connected only to very few computers in each network at once, you do NOT want to be in more than one P2P network at the same time. Even if you do have a lot of bandwidth, it's a better service to be a supernode to one network than a normal node to several networks.
Re:Here's the article without registration
on
Microsoft's New Hurdles
·
· Score: 4, Insightful
I didn't need to register to read the article, either. That's because I had already registered (:) ). Trust me, it's in everybody's good interests, yours included, that you don't misuse the convenience methods that NY Times has granted to you only under certain (implied) conditions. They aren't forced to give you anything for free, nor without registration. No, I don't think giving a link with the partner=google link will make them stop putting free content on the 'net altogether. This is a matter of morality. It seems pretty much impossible to me that it would be more work for you to register and read the two (count it, two, and never more) e-mails they send you than it is inconvenient for them to have bogus statistics. So why should you think of only yourself? Because they're not you? Because they're a company and you're a person? Because you think you're doing others a service?
This completely inane (I mean it, what I wrote IS inane!) moral sermon brought to you by someone who's not an AC. (now just try to figure out what I mean by that...)
At first I was going to comment that it doesn't make sense to think of Gates in the same job as Linus, but then I remembered that supervillains always tell superheroes that they're quite similar people. And on an off-topic note, they are. Not real, that is. I guess Bill and Torvalds always say the same thing when they enter people's homes;).
Damn you, you stole my post idea. In fact, if all light goes in one direction, all you will see is one spot at a time. I smell something very wrong here...
Perhaps it's just meant to be cool. After all, if all light from it goes in only one direction, people will see a reflection of the stuff on the screen on your face, just like in movies:-).
$apt-cache show freedom-us Package: freedom-us Priority: important Section: base Maintainer: The People of the United States Architecture: Human Version: 2002 Depends: libconstitution (>= 0.1971), libbillofrights (= 0-10) Filename: pool/main/f/freedom/freedom_2002_human.deb Descri ption: Freedom of the people of the United States. Included in base install of all US human systems. Use the localized packages freedom-eu and freedom-world for humans in other nations.
P.S. It's fun to write posts like this when you're a Finn and have never even visited the US;)
What about just redrawing everything, then? Not many people will be moving their apps around so much that the overhead from that would be anything but negligible...
The X Window System:
The standard UNIX graphical environment. With Linux, this is usually
XFree86 (http://www.xfree86.org). You may call it X, XFree, the X
Window System, XF86, or a host of other things. Call it 'XWindows' and
someone will smack you and you will have deserved it. -- fortune
Someone on that thread asked about an X equivalent for screen, one that would allow you to detach an X program from one screen and run it on another. I don't have an account at Sourceforge and don't care to get one right now, so here goes:
I've tried xmove once but never got it to run without segfaulting. Which isn't really a big surprise, since the last release is from 1995. However, if it worked correctly, it would sit between your X server and clients (guzzling some performance and probably making DRI, DGA, XSHM and the like pretty hard to configure). Maybe it would be possible to upgrade xmove to modern X11 revisions, but I'm not up to the task...
OTOH, would it be possible to implement this in X servers and/or Xlib itself? As far as I can see, an X client could just store its state, close its connection to the X server and initiate a connection with another X server to move from one server to another. Doing changes only to servers, it could be implemented with a little stupid redirection even without doing any changes to clients at all, but that would cause a lot of connection overhead...
Actually, at least ext3 can do full data journaling (there's a mount option for that, but I've forgotten what it is). It's dog slow since all data needs to be written on disk twice, once to the file and once to the journal, but at least it keeps your data intact well.
At first I thought that Elvis also used full data journaling, since this quote implied it:
The current version of NTFS, the file system within Windows 2000 and XP, does not handle full- fledged journaling, sources said; change-journal logs note alterations to files but don't provide enough information to reverse them.
However, AFAIK full data journaling should cause a 50% performance drop (unless the journal copies aren't written when the data proper is written - I don't actually know how this is handled, I wish free karma to the one who explains what those kjournald processes on my system actually are doing every five seconds) so Elvis probably only does metadata... or then I just don't know enough about journaling filesystems.
Except, of course, they're all virgins themselves... But since most computer hackers are very self-righteous, I wouldn't be surprised if they tried seducing themselves;).
You forgot the third choice: Centralize. As long as you aren't slapped with a lawsuit, you can just make one Phynd or GnucleusLAN take care of all intra-campus filesharing. University gets much, much lower bandwidth costs. Students get screamin' fast and reliable downloads.
You know, modding this post as "Funny" (and it's Insightful +1, Informative +1 and Funny +2 at the moment) is like pointing and laughing at LowAmmoWarning...
That wouldn't work, except if they asked the webmasters of porn sites to apply colour filters (and if they did that, they could just as well ask to have the images tarred, feathered and disguised as a random file type the netadmin has never heard of - preferably one that doesn't exist, that is). The pictures that a work-place pornsurfer gets would already have gone through the firewall when they're on his computer, and colour filters don't help at that part.
Actually it is an open-source problem, although not a very bad one if we just take up some good signing practices. Since everybody's free to copy software around at will, there will be a lot of servers where any particular program can be found - and some of those servers are inevitably insecure. MS (as just as an example) doesn't have this problem because of two things: Few people have the source of their programs so only few can put trojan code in the source, and MS probably has a pretty tight FreeBSD configuration that it uses on all internal servers;-).
I wonder if news.com has observed the Artistic License when making this link. I don't know what DeCSS.exe contains, but if all it contains is a program for DeCSSing video streams and no help files (more specifically, help files that tell where to get the standard version of DeCSS from), I think they have. See point 4: If the package is bare (and thus contains no help files or source code) and standard they definitely do not comply to any of a, b, or c, and they probably have not done any special arrangements, either. It's quite clear to me that DeCSS itself is protected speech under the first amendment, but I wouldn't be so sure about the linking.
(and if you're thinking about modding me down, see the sig. This post contains so many contradictions that it should be a +1 Funny, anyway)
Don't get your hopes up. Have you ever had a conversation with a robot on a chat channel? Takes a while to recognize. I don't think a "are you legit? yes I am" message can be distinguishing enough.
Automating this kind of negotiation will be easy as hell.
However, AFAIU knowingly using a bot to tell that you a are a real file-trader doesn't really legitimise working around the limitations of the safe harbour. If someone asks you if a file on your system is legit it isn't much different from asking a bot you've installed if a file on your system is legit.
Extrapolating a little (I haven't read the article nor the bill, so take this with three grains of salt and please only point out where I was in error without flames if I am wrong) we could just put something to the effect "When someone advertises a file on the network he or she implies that this file is what its name says and not bogus" into the official Gnutella protocol specification (if such a thing exists). Ta da. Bill worked around, at least partly.
Slashdot Mirror
Slashdot doesn't use well standardised XHTML/CSS, either. Nor does Slash, the code back-end, use many good programming principles that any Slashdot editor and most users would advocate - or at least so I've been told. Nor do a great many Slashdotters use an open-source OS (remember that poll about operating systems a while ago, some people still have the more shocking parts of the results in their sig)
Ergo, saying "Slashdot is not using it" is not saying much at all...
Gives a whole new meaning to karma whoring, though. Perhaps this should be called karma masturbation ;) ?
I use Debian GNU/Linux (and I like the GNU/ part there) exclusively at home. While I know a little C, PERL and PHP (none of them well), I have never written any software longer than a thousand lines (except years ago in BASIC, and that was mostly lists of stuff) or submitted a patch to any project. Sure, in time I might become a developer, but at least right now I just classify myself as an IT dabbler. So yes, such people exist who can use software and know what they are doing but can't very realistically contribute.
Remember one fact about real P2P networks: They have a lot of overhead for searching and, in the case of Gnutella, pinging/ponging. IM networks are centralised enough to not have that overhead, so being logged in to several of them doesn't use that much bandwidth. But unless you have insane amounts of bandwidth or are going to be connected only to very few computers in each network at once, you do NOT want to be in more than one P2P network at the same time. Even if you do have a lot of bandwidth, it's a better service to be a supernode to one network than a normal node to several networks.
This completely inane (I mean it, what I wrote IS inane!) moral sermon brought to you by someone who's not an AC. (now just try to figure out what I mean by that...)
... that the amount of words, characters and paragraphs in this beast of a document might well have been counted by flushing it down (through wc)?
At first I was going to comment that it doesn't make sense to think of Gates in the same job as Linus, but then I remembered that supervillains always tell superheroes that they're quite similar people. And on an off-topic note, they are. Not real, that is. I guess Bill and Torvalds always say the same thing when they enter people's homes ;).
Ever heard of floppy disks? Schools tend to love them...
Damn you, you stole my post idea. In fact, if all light goes in one direction, all you will see is one spot at a time. I smell something very wrong here...
:-).
Perhaps it's just meant to be cool. After all, if all light from it goes in only one direction, people will see a reflection of the stuff on the screen on your face, just like in movies
ABR? Aberrant Bit Rate?
$apt-cache show freedom-usi ption: Freedom of the people of the United States. Included in base install of all US human systems. Use the localized packages freedom-eu and freedom-world for humans in other nations.
;)
Package: freedom-us
Priority: important
Section: base
Maintainer: The People of the United States
Architecture: Human
Version: 2002
Depends: libconstitution (>= 0.1971), libbillofrights (= 0-10)
Filename: pool/main/f/freedom/freedom_2002_human.deb
Descr
P.S. It's fun to write posts like this when you're a Finn and have never even visited the US
What about just redrawing everything, then? Not many people will be moving their apps around so much that the overhead from that would be anything but negligible...
No, it isn't.
The X Window System:
The standard UNIX graphical environment. With Linux, this is usually
XFree86 (http://www.xfree86.org). You may call it X, XFree, the X
Window System, XF86, or a host of other things. Call it 'XWindows' and
someone will smack you and you will have deserved it.
-- fortune
I've tried xmove once but never got it to run without segfaulting. Which isn't really a big surprise, since the last release is from 1995. However, if it worked correctly, it would sit between your X server and clients (guzzling some performance and probably making DRI, DGA, XSHM and the like pretty hard to configure). Maybe it would be possible to upgrade xmove to modern X11 revisions, but I'm not up to the task...
OTOH, would it be possible to implement this in X servers and/or Xlib itself? As far as I can see, an X client could just store its state, close its connection to the X server and initiate a connection with another X server to move from one server to another. Doing changes only to servers, it could be implemented with a little stupid redirection even without doing any changes to clients at all, but that would cause a lot of connection overhead...
Except, of course, they're all virgins themselves... But since most computer hackers are very self-righteous, I wouldn't be surprised if they tried seducing themselves ;).
You forgot the third choice: Centralize. As long as you aren't slapped with a lawsuit, you can just make one Phynd or GnucleusLAN take care of all intra-campus filesharing. University gets much, much lower bandwidth costs. Students get screamin' fast and reliable downloads.
* points and laughs at LowAmmoWarning *
That wouldn't work, except if they asked the webmasters of porn sites to apply colour filters (and if they did that, they could just as well ask to have the images tarred, feathered and disguised as a random file type the netadmin has never heard of - preferably one that doesn't exist, that is). The pictures that a work-place pornsurfer gets would already have gone through the firewall when they're on his computer, and colour filters don't help at that part.
Actually it is an open-source problem, although not a very bad one if we just take up some good signing practices. Since everybody's free to copy software around at will, there will be a lot of servers where any particular program can be found - and some of those servers are inevitably insecure. MS (as just as an example) doesn't have this problem because of two things: Few people have the source of their programs so only few can put trojan code in the source, and MS probably has a pretty tight FreeBSD configuration that it uses on all internal servers ;-).
Although a real geek would, of course, say that her number is 2002:564b:309:: :-).
(and if you're thinking about modding me down, see the sig. This post contains so many contradictions that it should be a +1 Funny, anyway)
Extrapolating a little (I haven't read the article nor the bill, so take this with three grains of salt and please only point out where I was in error without flames if I am wrong) we could just put something to the effect "When someone advertises a file on the network he or she implies that this file is what its name says and not bogus" into the official Gnutella protocol specification (if such a thing exists). Ta da. Bill worked around, at least partly.
Especially since this is pretty useful as the article is slashdotted already... * makes an eye-rolling movement *
;-). Or then my theory of Mass Idiocy (see journal) is correct...
Ah well, I guess some mods just don't have a clue