But which emails are lost? Tax-related emails or opt-in mailing lists? For most modern spam tools, false positives are usually grey anyway: marketing emails or mailing lists which the user does indeed want, but can't be easily distinguished from spam. I think that if you run SpamAssassin on a corpus of your important corporate tax-emails intermixed with real spam, the chance of the important emails being falsely tagged is considerably less than even the 1/25,000 mentioned.
As otherwise mentioned, if you absolutely have to have it, email isn't the tool anyway. Put something on your website which allows users to submit these tax forms. Or require that users provide their email addresses in advance (on the website) which automatically go into the whitelist.
Your requirement of 'perfectly filtered email from arbitrary sources' is silly. Have fun wading through spam. Can't wait until you accidentally delete one of those tax emails: human tagging of spam is much worse than these numbers you find so unacceptable.
Reiser4 implements a resier4() syscall with new semantics for their custom operations. That's the interesting thing about this; they're adding onto the Unix APIs. Therefore (from a syscall standpoint at least) they can do things like transactions which open() and write() can't support. Of course, this comes at the expense of compatibility with other filesystems, but if it works as advertised, I'd have no problem writing code which takes advantage of it.
How large (and long) can Reiser atomic transactions be?
Can I write an installation program which creates, replaces, moves, and deletes many files and directories, and have it all be under one transaction with a single commit at the end? Do other 'sessions' not see the transaction until it is complete? Are sessions based on processes or threads or something else?
That would be pretty amazing, to be able to roll back large sets of changes in case of an error. I know that database rollbacks can take large amounts of time (they optimize for the commit, which makes perfect sense) but nonetheless having rollback support in applications would be sensational.
The point isn't to guarantee that data was flushed to disk. But if it wasn't, you know it wasn't, and you have a previous copy of all the blocks in the transaction.
Atomicity isn't the same thing as synchronous writes, where the OS won't return from the write() call until the data has truly been flushed successfully to the disk controller (and with controllers which support it, to the platter itself). Atomicity says that a set of writes either happens completely, or not at all.
OT: Why is it that I have three different PDF viewers on my Fedora Core 2 system, and I've found at least one rendering bug in each of them? Honestly, I have to keep track of which viewer will render which PDF correctly.
"BZZT Wrong answer" has to be the most annoying and childish catch-phrase of Slashdotters. It implies enormous intellectual superiority over the original poster, and yet a majority of the time, someone else (as happened here) proves the opposite is true.
Security by definition must limit functionality. The best you can hope for is that the functionality limited is less valuable than the security gained.
Microsoft management has finally realized that in order to avoid the gigantic fiascos of the past year's worms, they have to limit some functionality. My guess is Microsoft engineers have been telling their management this for a long time, and finally, they were heard.
M: Is our product secure? E: The only way to improve security is at the expense of features. M: No way. Features sell the product.
M: We need to patch this security hole. E: The only way to improve security is at the expense of features. M: I still can't accept this.
M: Please, dear god, do ANYTHING to fix these security problems! E: The only way to improve security is at the expense of features. M: All right, all right! Do it!
I have the same setup (Fedora Core 2 does a nice job of automounting my camera) but a real problem is that FAM (which monitors drives for changes and alerts Nautilus) has to keep file handles on the mounted drive open in order to monitor directories, and this prevents unmounting the volume. The easiest workaround I know of is killing FAM (actually by restarting xinetd), which is ugly. Unfortunately, from what I understand, it's a kernel problem: the file/directory monitoring API requires open file handles, and last I heard the kernel hackers didn't have a solution they liked. Does anyone know if this is fixed (possibly by DBUS)? I imagine the solution is more in the unmounting code (possibly alerting FAM that the volume is trying to unmount) rather than the monitoring API.
The meaning of KB is currently being debated (in this very thread) but KiB is definitely 1024.
To continue the argument: you don't have to make a fool of yourself by saying 'kibibytes', which sounds like you have a speech impediment. But is (writing) KiB really that bad? It's being exact, and isn't precise language part of geek culture? And it's got mixed case, which geeks love almost as much as LotR.
Use 'kilobytes' when you want to be quick, and use 'KiB' when you want to be precise.
Dictionaries list kilobyte as 1000 bytes because that's correct. The SI prefixes existed before kilobytes did. Just because 2^10 is 'conveniently close' to 1000 doesn't make it kilo-.
Now, it often does help marketers to have more kilobytes in their product...
I'm looking for a good, small cell phone which also functions as a basic PDA. I'm tired of keeping my cell phone address book updated manually. I'd also like basic access to my calendar, but I don't want something a) separate from my phone or b) larger than my phone.
It doesn't have to allow for convenient text entry, as long as it will sync with a computer. (You can't really enter text into something as small as a cell phone anyway).
Let me know when you write an operating system which runs on everything Windows does, and handles upgrades better. In fact, let me know when ANYONE writes an operating system that does a better job with crazier hardware. Microsoft has a lot of problems, but they hire good engineers. It's a complex problem.
Have you ever developed software? Do you know anything about deploying software? Do you know how difficult it is to upgrade software on millions of machines that have had near-infinite permutations of software written by either malicious or ignorant third-party developers installed and uninstalled?
There's a lot Microsoft could do better. But I really can't stand you implying that they're a bunch of idiots when it's painfully obvious it's the other way around.
Do you really care whether Slashdot is implemented with CSS? The savings mentioned would be for OSDN, Slashdot's parent. Apparently they don't care enough about the potential savings to make the change, and while that may be a bad business decision, unless you're an investor in OSDN I really don't see why you should care.
It isn't like a puppy is killed every time a GB of OSDN bandwidth is wasted.
I send mail from my home server through my ISP as a smarthost. DNS is managed by another company (easyDNS). I assume that I would have to have my DNS provider enter the SPF information, since I don't manage it myself. Do most DNS providers allow the user to enter data like this in the TXT record?
Of course there is nothing illegal about implementing a standard from scratch. But copying another person's implementation is indeed infringement. It's possible that a kernel developer was careless and thought "why should I go to the effort of writing an ELF implementation from scratch, when the UNIX implementation is pretty much in the public domain?"
If they did that, it would be copyright infringement.
Personally, I love streaming music. My stereo at home is connected to my PC, which is always connected to the net. On Rhapsody I can play nearly every album I've ever owned or wanted to listen to, for a flat monthly fee.
The best thing about unlimited streaming is that I can listen to albums which I would probably never buy, or even take the time to borrow or copy. When someone says 'hey, listen to this band' I can check them out right away, for no extra money.
Speakeasy is great. They have a great range of services from 56k to 3.0/768, and they don't give a rat's ass what you do with your connection as long as it's legal.
Slackware is for people who want a classic UNIX system. Debian, Red Hat, etc. all have their places, but Slack is for people who grew up on, administer, use, and love UNIX.
The broadcast flag really cracks me up. TV and radio stations put up freakin' huge antennas so that they can broadcast their signal so strongly I can practically hear it through my orthodontics...and then don't want anyone to record it.
If I stood on top of a mountain and sang a song so loud nobody within twenty miles can avoid hearing it, can I complain if people record it?
Private performances, and things like cable and satellite, are different, because there is an expectation of some privacy: it's not being distributed in a completely public manner. But broadcasting? How can you possibly constrain what people do with what you broadcast?
Apple's new "Spotlight" search technology is by far Tiger's most dominant feature, and it can be accessed from almost every corner of the system, literally.
A blue-colored Spotlight search button appears in the upper-right-hand corner of the Mac OS menubar, and remains accessible at that point from any Mac OS X application. Selecting the Spotlight icon reveals a search field that will expand to display results in real-time.
The British Empire was most certainly as dominating a power as the US for at least a half-century. By definition, nothing is at its peak for more than a moment.
Slashdot Mirror
But which emails are lost? Tax-related emails or opt-in mailing lists? For most modern spam tools, false positives are usually grey anyway: marketing emails or mailing lists which the user does indeed want, but can't be easily distinguished from spam. I think that if you run SpamAssassin on a corpus of your important corporate tax-emails intermixed with real spam, the chance of the important emails being falsely tagged is considerably less than even the 1/25,000 mentioned.
As otherwise mentioned, if you absolutely have to have it, email isn't the tool anyway. Put something on your website which allows users to submit these tax forms. Or require that users provide their email addresses in advance (on the website) which automatically go into the whitelist.
Your requirement of 'perfectly filtered email from arbitrary sources' is silly. Have fun wading through spam. Can't wait until you accidentally delete one of those tax emails: human tagging of spam is much worse than these numbers you find so unacceptable.
Reiser4 implements a resier4() syscall with new semantics for their custom operations. That's the interesting thing about this; they're adding onto the Unix APIs. Therefore (from a syscall standpoint at least) they can do things like transactions which open() and write() can't support. Of course, this comes at the expense of compatibility with other filesystems, but if it works as advertised, I'd have no problem writing code which takes advantage of it.
How large (and long) can Reiser atomic transactions be?
Can I write an installation program which creates, replaces, moves, and deletes many files and directories, and have it all be under one transaction with a single commit at the end? Do other 'sessions' not see the transaction until it is complete? Are sessions based on processes or threads or something else?
That would be pretty amazing, to be able to roll back large sets of changes in case of an error. I know that database rollbacks can take large amounts of time (they optimize for the commit, which makes perfect sense) but nonetheless having rollback support in applications would be sensational.
The point isn't to guarantee that data was flushed to disk. But if it wasn't, you know it wasn't, and you have a previous copy of all the blocks in the transaction.
Atomicity isn't the same thing as synchronous writes, where the OS won't return from the write() call until the data has truly been flushed successfully to the disk controller (and with controllers which support it, to the platter itself). Atomicity says that a set of writes either happens completely, or not at all.
OT: Why is it that I have three different PDF viewers on my Fedora Core 2 system, and I've found at least one rendering bug in each of them? Honestly, I have to keep track of which viewer will render which PDF correctly.
"BZZT Wrong answer" has to be the most annoying and childish catch-phrase of Slashdotters. It implies enormous intellectual superiority over the original poster, and yet a majority of the time, someone else (as happened here) proves the opposite is true.
Legally they're right only if EULAs are enforceable, which is certainly debateable.
BTW, by reading this comment you have agreed to give me your first-born son.
Security by definition must limit functionality. The best you can hope for is that the functionality limited is less valuable than the security gained.
Microsoft management has finally realized that in order to avoid the gigantic fiascos of the past year's worms, they have to limit some functionality. My guess is Microsoft engineers have been telling their management this for a long time, and finally, they were heard.
M: Is our product secure?
E: The only way to improve security is at the expense of features.
M: No way. Features sell the product.
M: We need to patch this security hole.
E: The only way to improve security is at the expense of features.
M: I still can't accept this.
M: Please, dear god, do ANYTHING to fix these security problems!
E: The only way to improve security is at the expense of features.
M: All right, all right! Do it!
I have the same setup (Fedora Core 2 does a nice job of automounting my camera) but a real problem is that FAM (which monitors drives for changes and alerts Nautilus) has to keep file handles on the mounted drive open in order to monitor directories, and this prevents unmounting the volume. The easiest workaround I know of is killing FAM (actually by restarting xinetd), which is ugly. Unfortunately, from what I understand, it's a kernel problem: the file/directory monitoring API requires open file handles, and last I heard the kernel hackers didn't have a solution they liked. Does anyone know if this is fixed (possibly by DBUS)? I imagine the solution is more in the unmounting code (possibly alerting FAM that the volume is trying to unmount) rather than the monitoring API.
The meaning of KB is currently being debated (in this very thread) but KiB is definitely 1024.
To continue the argument: you don't have to make a fool of yourself by saying 'kibibytes', which sounds like you have a speech impediment. But is (writing) KiB really that bad? It's being exact, and isn't precise language part of geek culture? And it's got mixed case, which geeks love almost as much as LotR.
Use 'kilobytes' when you want to be quick, and use 'KiB' when you want to be precise.
Dictionaries list kilobyte as 1000 bytes because that's correct. The SI prefixes existed before kilobytes did. Just because 2^10 is 'conveniently close' to 1000 doesn't make it kilo-.
Now, it often does help marketers to have more kilobytes in their product...
I'm looking for a good, small cell phone which also functions as a basic PDA. I'm tired of keeping my cell phone address book updated manually. I'd also like basic access to my calendar, but I don't want something a) separate from my phone or b) larger than my phone.
It doesn't have to allow for convenient text entry, as long as it will sync with a computer. (You can't really enter text into something as small as a cell phone anyway).
Let me know when you write an operating system which runs on everything Windows does, and handles upgrades better. In fact, let me know when ANYONE writes an operating system that does a better job with crazier hardware. Microsoft has a lot of problems, but they hire good engineers. It's a complex problem.
Have you ever developed software? Do you know anything about deploying software? Do you know how difficult it is to upgrade software on millions of machines that have had near-infinite permutations of software written by either malicious or ignorant third-party developers installed and uninstalled?
There's a lot Microsoft could do better. But I really can't stand you implying that they're a bunch of idiots when it's painfully obvious it's the other way around.
Do you really care whether Slashdot is implemented with CSS? The savings mentioned would be for OSDN, Slashdot's parent. Apparently they don't care enough about the potential savings to make the change, and while that may be a bad business decision, unless you're an investor in OSDN I really don't see why you should care.
It isn't like a puppy is killed every time a GB of OSDN bandwidth is wasted.
I send mail from my home server through my ISP as a smarthost. DNS is managed by another company (easyDNS). I assume that I would have to have my DNS provider enter the SPF information, since I don't manage it myself. Do most DNS providers allow the user to enter data like this in the TXT record?
Of course there is nothing illegal about implementing a standard from scratch. But copying another person's implementation is indeed infringement. It's possible that a kernel developer was careless and thought "why should I go to the effort of writing an ELF implementation from scratch, when the UNIX implementation is pretty much in the public domain?"
If they did that, it would be copyright infringement.
Personally, I love streaming music. My stereo at home is connected to my PC, which is always connected to the net. On Rhapsody I can play nearly every album I've ever owned or wanted to listen to, for a flat monthly fee.
The best thing about unlimited streaming is that I can listen to albums which I would probably never buy, or even take the time to borrow or copy. When someone says 'hey, listen to this band' I can check them out right away, for no extra money.
Speakeasy is great. They have a great range of services from 56k to 3.0/768, and they don't give a rat's ass what you do with your connection as long as it's legal.
Fundamentalist Christians would be blowing people up if they were living in Palestine under Muslim rule in Jerusalem.
Slackware is for people who want a classic UNIX system. Debian, Red Hat, etc. all have their places, but Slack is for people who grew up on, administer, use, and love UNIX.
The broadcast flag really cracks me up. TV and radio stations put up freakin' huge antennas so that they can broadcast their signal so strongly I can practically hear it through my orthodontics...and then don't want anyone to record it.
If I stood on top of a mountain and sang a song so loud nobody within twenty miles can avoid hearing it, can I complain if people record it?
Private performances, and things like cable and satellite, are different, because there is an expectation of some privacy: it's not being distributed in a completely public manner. But broadcasting? How can you possibly constrain what people do with what you broadcast?
Anyone have recommendations on a video card that will run this, and still be (very) Linux-compatible?
Don't worry, even if someone erases the writing on the wall, VMS users will be able to see it, along with the 20 previous versions.
Finally, a correct use of the word 'literally':
Apple's new "Spotlight" search technology is by far Tiger's most dominant feature, and it can be accessed from almost every corner of the system, literally.
A blue-colored Spotlight search button appears in the upper-right-hand corner of the Mac OS menubar, and remains accessible at that point from any Mac OS X application. Selecting the Spotlight icon reveals a search field that will expand to display results in real-time.
The British Empire was most certainly as dominating a power as the US for at least a half-century. By definition, nothing is at its peak for more than a moment.