Would you please quit linking to the NY Post as a "news source"? The last time this happened was when the NY Post was claiming that McDonalds would be giving away a billion iTunes songs... which turned out to be completely untrue, an unfounded rumour.
For the love of god, stop linking to tabloids as news! If it's reported somewhere respectable then fine, but it's not a story until you've got more than this pathetic 200 word paragraph from some grocery store checkout RAG.
Hey dude, would you PLEASE learn how to post to slashdot without inserting tags in your post? You're probably using "Plain old text" and hitting ENTER on each newline or something. Or perhaps you wrote your reply in a text editor and then pasted it into the box, which also copies the hard newlines, which are turned into by slashcode if you use "plain old text mode". In any case, your text has a hard line break after every 80 characters or so, and hence it looks like ASS when displayed. Your post taked up about 1/3 of the horizontal space on the screen, and thus is about 3 times taller than it should be.
Either remove the hard newlines, or use a mode (such as HTML) that doesn't convert newlines into breaks (but this requires that you explicitly add the BR tags to your post to make a new paragraph.)
First, this is a non-article. There are no actual findings reported, just that the author happened to notice that more of his labeled disks were failing compared to non-labeled. Well, that's hardly a scientific conclusion, it could be due to any number of things. Without a controlled experiment I don't lend any credance to this theory.
Second, I'd like to suggest that perhaps labeled disks are suffering from mechanical problems and not optical problems. What I mean by that is that if the label is not centered precisely it will cause the CDR to no longer balance. When spun at very high speeds, even a little bit of out-of-balance will be multiplied by the huge lateral accelerations, causing the disk to flop around.
If you know anything about CD optical pickups, you know that the closed-loop servo system used to keep them in focus does a remarkable job of compensating for defects in the disk. Somewhere on the net there was an analogy that it was the equivalent of flying 50 feet above a highway at 600 MPH and being able to maintain that height to several inches while counting small spots in the road a few inches wide. Whatever -- I don't remember the specifics, but basically the servo loop that maintains the precise positioning does an incredible job of keeping the pickup in the right place over the disk. But it can't do miracles...
So the idea I'm presenting here is that perhaps these labels cause enough mechanical flutter to bring the disk too far out of the bounds that the optical servo system can deal with. In this theory, all discs age to some degree and become harder to read, but in the case of a misapplied label it quickly gets to be too much for the drive, and thus the disk becomes unreadable.
I've always used one of those "CD Stomper" applicators for the disks that I did label, and I find that it puts the label on almost exactly on center. I have many CDR disks from early 1998 (approaching 6 years old) that I made in the manner, and all of them are still perfectly readable as far as I can tell -- even in the car stereo which is particularly unforgiving. Again this is not scientific at all, but I'm just proposing a theory here, that perhaps it has more to do with the mechanical properties of the label and nothing to do with it "eating the dye away" or anything like that.
Re:Bluetooth directional antennas: 2-way gain
on
Spammed by Bluetooth
·
· Score: 1
Yeah, I believe this is called the "reciprocity principle" or something. If you design something that has a particular RF radiation pattern (i.e. Tx) then it will have a corresponding absorbtion pattern (Rx).
NAT mitigates the problem for the most part on the client (or typical "end user") end, very true.
One thing that I have not seen many people mention in this discussion is the server-side analogue: Name-based virtual hosting. Circa 1996 or 1997, a majority of web sites operated on the 1:1 mapping of IP address to hostname, as most browsers were still old stupid HTTP/1.0 things. But with the advent of name-based virtual hosting made possible by the Host: field that was optional in HTTP/1.0 but required in HTTP/1.1, we now routinely have hundreds or thousands of web sites on a single IP address. This vastly cut down on the appetite of IP addresses on the server end of things. If you're ever bored or curious, browse over to http://whois.sc and try out their reverse-IP tool. It requires that you sign up for a free account, but once you've done that you can enter any IP address and see all the web sites that are hosted on that interface. It's often quite surprising how many sites are on some server. You can use this trick if you're paying for shared web space from a hosting company to see approximately how many other sites they have running on that same server.
The only thorn in the side of name-based vhosting is SSL. You cannot host more than a single https site per IP address. Or at least, you can't and still have each site present its own signed certificate. It's a chicken-and-egg problem. The server doesn't know which host it is supposed to be masquerading as until the HTTP headers arrive, but this happens after the secure channel has already been established. When the client initially asks for the server's certificate it has no choice but to respond with the default, meaning that you can really only realistically host one https site per IP address. But I'd say this is really a minority of sites out there, and there are workarounds such as hosting multiple secure sites from a single domain: www.secureserver.com/site1, www.secureserver.com/site2, etc.
Riiiiight. So, that's why if you visit the FAQ link you're currently greeted by an image of the goatse.cx guy spreading his anus? Because surely the trolls could never figure out how to just vote for that change enough times to get it accepted.
I think the point was that with a headline like "Not Your Father's Periodic Table" we (I, at least) expected to click on a link and see something earth shatteringly novel, unique, and different... Instead it's the same chart, with the same format, in the same shape, with the elements in the same order. Only instead of atomic number it's got abundancy and condensation temperature. Whohoo. Oh, and instead of being colored by solid/liquid/gas, it's something different. I'm sure this is interesting to cosmochemists (wtf?) but to me it's rather dull.
I contend that, indeed, my father WOULD recognise this as his periodic table.
boxens!! No offense, but reading people write "boxen" when they mean "boxes" is almost as nauseating as the incorrect "virii". But you took it to a new level with "boxens". I await "viriiuses". That would just about cap it off.
This has got to be one of the worst examples of gratuitous linking in a story write-up I've ever seen. Gee, thanks buddy, I would have never guessed that apple.com was Apple's web site, or that aol.com was where I could find AOL on "teh intarweb". Oh, and I was really wondering where Pepsico's web site was... good thing there's that link to pepsi.com.
If you can't find any relevent links for your little write-up then don't include those idiotic generic links, it's completely pointless.
I hope that you all realise the POP3 and IMAP send logins and passwords in cleartext. I know that most people view email as insecure anyway, but sending around your password in the clear repeatedly (your client checks every 10 or 20 minutes I'd imagine) is just asking for trouble. You wouldn't telnet into a server in this modern age, so why do the equivalent with your email?
You can teach any client to speak secure POP3 or IMAP by just using stunnel and a email provider that offers an SSL tunnel for these protocols. This, in my opinion, is one of the biggest reasons why free POP3 services suck.
The best solution to email, IMO, is to buy your own domain ($9 a year from godaddy.com) and then pay for very cheap hosting that includes email. You can get a lot for only $20 - $60/year. And since you control the domain you never have to worry about changing your email address again. If your provider goes out of business or just pisses you off, you get another one and just update your domain's DNS. You'll have no more of those "well, I was using someisp.net for email but now I'm with anotherisp.com, so please update your addressbook" emails. Another plus of actually paying for email hosting is that you often get a lot of options for spam filtering, such as RBLs or SpamAssassin. With yahoo/hotmail you are totally at their mercy as far as spam-detection. And I have found that SpamAssassin is orders of magnitude better than anything yahoo has ever done.
Uh, what? How does loading a 1x1 GIF reveal anything about your screen resolution? It will simply be another entry in a log file, which records the URL, the IP address, the time, the referer, and the user-agent. All of those fields are present in the log of the server that's serving the main html page.
In order to determine any further info about the user, you'd have to use Javascript to get this information from the DOM, and then somehow code that into a URL which gets submitted or posted to a server somewhere. From the blurb in the article there was no such code, just a simple IMG tag.
Please let this above post be a lesson to all you amateur spamfighters out there: You can never trust the From: line when dealing with spam.
If you have ever replied to spam, or "reported" spam or blacklisted based on who it purports to be from, then you are a part of the problem, not part of the solution. I use quotes around reported above because most people are clueless about what needs to be done to actually help the people that work the abuse desks.
joe jobs would not be nearly so annoying and devastating if there weren't so many people out there that just don't understand email. When you receive spam it makes you mad, and I guess the natural reaction is to reply to it or try to see who it's from and add that to a blacklist. But resist, as almost always the information you are basing this judgement on has absolutely nothing to do with who was responsible for that turdlet in your inbox.
Were those addresses that you put in the address book randomly created hotmail/yahoo (or any other popular free webmail) accounts? If so, I'm not surprised that they received spam within hours of being created. Spammers hammer the popular webmail servers with all sorts of permutations of letters and numbers, becuase with so many users they're likely to find enough addresses to make it worth their while. It doesn't mean that the infected machine had anything to do with it.
If you really wanted to prove something with your experiment, you need a control group. Pick to random strings and create those accounts on hotmail/yahoo/whatever. Then add one of them to the address book but keep the other completely secret. If the one you added to the address book truly gets significantly more spam then I'd be impressed by your experiment. Otherwise, it's unsubstantiated and unscientific.
Alternatively, you could create these new addresses on an obscure domain name that's not likely to be subject of a dictionary attack... If that was what you did then I'd also believe the findings.
Because Rob and the rest of the "editors" feel that by not giving a shit about typos, incomplete sentences, run-on sentences, duplicates, and spelling mistakes it gives the site the amateurish feel that it had when they first started it for the hell of it in 1997. Of course, a lot of people argue that since now they're PAID to run slashdot and since the site has become huge, some level of actual "editing" should be required of these people that claim to be "editors" yet seem to have no idea what the job of an "editor" actually entails.
Did you read the article or just rush to post a comment? The amount of money that your coupon is worth has nothing to do with how many times you listen to something. The use of tracking the number of times you play a song is so that the artist that you get the most out of gets a corresponding part of your coupon -- but you can donate your coupon to whomever you want in any proportion if you so desire. You can listen to as many or as few artists as many times as you want, the tax is a small and constant amount. This has nothing to do with "pay per listen".
Take all of those arguments and apply them to any public thing: roads, parks, libraries, etc.
"We should have private roads where you purchase the rights to use whatever freeways you frequent from whoever built them. I don't trust that fat government issuing a gas tax to pay for things we all use like roads. I only use a few different routes anyway, why should I help pay for all those other suckers to use the roads I don't?"
"I don't want public parks, I'd rather sign up for private memberships to closed, gated-off parks. I don't trust some filthy organization is really using my tax dollars to pay rangers, maintenance staff, etc."
Nobody said you had to pay artists. I'm sure there would be some way to direct your coupon towards some charity or some organization, or even back to yourself. The point is that people -want- to drop a penny in the tip jar, but they just can't at the moment. If you don't want to support an artist or artists you can send your coupon along to whoever you want (yourself included), but the point is that most people DO INDEED want to support artists, and this makes it very easy to do that directly.
So you pay yourself. Okay, fine, that's no better than the current situation. The point is that most people DO want to support the artist in some way. But in the current situation they fell that doing this by paying for CDs is either unreansonable or is not benefiting the artist at all. In the proposed scheme, the user has a guarantee that the money is making it directly to the artist. And since it's rolled in with other purchases, it's much easier to justify "doing the right thing."
In other words, this allows people to honestly express their desire to support an artist, and this is a desire that the majority of people have. Sure, some will find a way to pay themselves, but that's fine: it's no worse than the current situation, and as long as those people are in the minority it doesn't matter at all.
It's kind of like that checkbox at the top of your taxes that asks if you want some amount to be donated to whatever-it-is. I assume that most people check that, because either way they're still paying the same amount in taxes, and the cause that it goes to is probably reasonably worthy to most people.
I find the whole notion presented in this article deplorable. What ISP is going to want to self-inflict a barrel customer rage? Where does all this money go? Who's in charge of verifying claims? What's to stop malicious users from filing false reports, or clandistinely installing software to incriminate an enemy's PC? How are all these ignorant customers supposed to be educated that they're suddenly liable for tens or hundreds of dollars in fines? If you just start arbitralily fining people, I don't care how little it is, you will bring down a boatloat of wrath and ire. People -hate- fees they don't know of in advance.
But one good thing would come of such a plan: egress filtering by all ISPs. This means that source-spoofed packets would be dropped before they get very far. It would make it significantly harder to spoof anything. No more RFC1918 packets on the public internet. If you ever run a public server on the internet, sometime try adding firewall rules to log and then drop all Bogon packets: those from unrouteable IP space, reserved or unallocated space, etc. You will be surprised how much of that stuff is floating around on the public internet, just soaking up legitimate bandwidth. Egress filtering would cause a much higher level of net-hygeine, in my opinion.
Slashdot Mirror
Would you please quit linking to the NY Post as a "news source"? The last time this happened was when the NY Post was claiming that McDonalds would be giving away a billion iTunes songs... which turned out to be completely untrue, an unfounded rumour.
For the love of god, stop linking to tabloids as news! If it's reported somewhere respectable then fine, but it's not a story until you've got more than this pathetic 200 word paragraph from some grocery store checkout RAG.
Hey dude, would you PLEASE learn how to post to slashdot without inserting
tags in your post? You're probably using "Plain old text" and hitting ENTER on each newline or something. Or perhaps you wrote your reply in a text editor and then pasted it into the box, which also copies the hard newlines, which are turned into
by slashcode if you use "plain old text mode". In any case, your text has a hard line break after every 80 characters or so, and hence it looks like ASS when displayed. Your post taked up about 1/3 of the horizontal space on the screen, and thus is about 3 times taller than it should be.
Either remove the hard newlines, or use a mode (such as HTML) that doesn't convert newlines into breaks (but this requires that you explicitly add the BR tags to your post to make a new paragraph.)
First, this is a non-article. There are no actual findings reported, just that the author happened to notice that more of his labeled disks were failing compared to non-labeled. Well, that's hardly a scientific conclusion, it could be due to any number of things. Without a controlled experiment I don't lend any credance to this theory.
Second, I'd like to suggest that perhaps labeled disks are suffering from mechanical problems and not optical problems. What I mean by that is that if the label is not centered precisely it will cause the CDR to no longer balance. When spun at very high speeds, even a little bit of out-of-balance will be multiplied by the huge lateral accelerations, causing the disk to flop around.
If you know anything about CD optical pickups, you know that the closed-loop servo system used to keep them in focus does a remarkable job of compensating for defects in the disk. Somewhere on the net there was an analogy that it was the equivalent of flying 50 feet above a highway at 600 MPH and being able to maintain that height to several inches while counting small spots in the road a few inches wide. Whatever -- I don't remember the specifics, but basically the servo loop that maintains the precise positioning does an incredible job of keeping the pickup in the right place over the disk. But it can't do miracles...
So the idea I'm presenting here is that perhaps these labels cause enough mechanical flutter to bring the disk too far out of the bounds that the optical servo system can deal with. In this theory, all discs age to some degree and become harder to read, but in the case of a misapplied label it quickly gets to be too much for the drive, and thus the disk becomes unreadable.
I've always used one of those "CD Stomper" applicators for the disks that I did label, and I find that it puts the label on almost exactly on center. I have many CDR disks from early 1998 (approaching 6 years old) that I made in the manner, and all of them are still perfectly readable as far as I can tell -- even in the car stereo which is particularly unforgiving. Again this is not scientific at all, but I'm just proposing a theory here, that perhaps it has more to do with the mechanical properties of the label and nothing to do with it "eating the dye away" or anything like that.
Yeah, I believe this is called the "reciprocity principle" or something. If you design something that has a particular RF radiation pattern (i.e. Tx) then it will have a corresponding absorbtion pattern (Rx).
NAT mitigates the problem for the most part on the client (or typical "end user") end, very true.
One thing that I have not seen many people mention in this discussion is the server-side analogue: Name-based virtual hosting. Circa 1996 or 1997, a majority of web sites operated on the 1:1 mapping of IP address to hostname, as most browsers were still old stupid HTTP/1.0 things. But with the advent of name-based virtual hosting made possible by the Host: field that was optional in HTTP/1.0 but required in HTTP/1.1, we now routinely have hundreds or thousands of web sites on a single IP address. This vastly cut down on the appetite of IP addresses on the server end of things. If you're ever bored or curious, browse over to http://whois.sc and try out their reverse-IP tool. It requires that you sign up for a free account, but once you've done that you can enter any IP address and see all the web sites that are hosted on that interface. It's often quite surprising how many sites are on some server. You can use this trick if you're paying for shared web space from a hosting company to see approximately how many other sites they have running on that same server.
The only thorn in the side of name-based vhosting is SSL. You cannot host more than a single https site per IP address. Or at least, you can't and still have each site present its own signed certificate. It's a chicken-and-egg problem. The server doesn't know which host it is supposed to be masquerading as until the HTTP headers arrive, but this happens after the secure channel has already been established. When the client initially asks for the server's certificate it has no choice but to respond with the default, meaning that you can really only realistically host one https site per IP address. But I'd say this is really a minority of sites out there, and there are workarounds such as hosting multiple secure sites from a single domain: www.secureserver.com/site1, www.secureserver.com/site2, etc.
Riiiiight. So, that's why if you visit the FAQ link you're currently greeted by an image of the goatse.cx guy spreading his anus? Because surely the trolls could never figure out how to just vote for that change enough times to get it accepted.
I think the point was that with a headline like "Not Your Father's Periodic Table" we (I, at least) expected to click on a link and see something earth shatteringly novel, unique, and different... Instead it's the same chart, with the same format, in the same shape, with the elements in the same order. Only instead of atomic number it's got abundancy and condensation temperature. Whohoo. Oh, and instead of being colored by solid/liquid/gas, it's something different. I'm sure this is interesting to cosmochemists (wtf?) but to me it's rather dull.
I contend that, indeed, my father WOULD recognise this as his periodic table.
boxens!! No offense, but reading people write "boxen" when they mean "boxes" is almost as nauseating as the incorrect "virii". But you took it to a new level with "boxens". I await "viriiuses". That would just about cap it off.
Since she works for the usgs I'm sure she's used to socially inept nerds calling at odd hours for no apparent reason.
This has got to be one of the worst examples of gratuitous linking in a story write-up I've ever seen. Gee, thanks buddy, I would have never guessed that apple.com was Apple's web site, or that aol.com was where I could find AOL on "teh intarweb". Oh, and I was really wondering where Pepsico's web site was... good thing there's that link to pepsi.com.
If you can't find any relevent links for your little write-up then don't include those idiotic generic links, it's completely pointless.
But does it run linux?
...and pray tell what exactly does BitTorrent have to do with the volume of email moving through these ISPs' mail servers? (Answer: nothing)
I hope that you all realise the POP3 and IMAP send logins and passwords in cleartext. I know that most people view email as insecure anyway, but sending around your password in the clear repeatedly (your client checks every 10 or 20 minutes I'd imagine) is just asking for trouble. You wouldn't telnet into a server in this modern age, so why do the equivalent with your email?
You can teach any client to speak secure POP3 or IMAP by just using stunnel and a email provider that offers an SSL tunnel for these protocols. This, in my opinion, is one of the biggest reasons why free POP3 services suck.
The best solution to email, IMO, is to buy your own domain ($9 a year from godaddy.com) and then pay for very cheap hosting that includes email. You can get a lot for only $20 - $60/year. And since you control the domain you never have to worry about changing your email address again. If your provider goes out of business or just pisses you off, you get another one and just update your domain's DNS. You'll have no more of those "well, I was using someisp.net for email but now I'm with anotherisp.com, so please update your addressbook" emails. Another plus of actually paying for email hosting is that you often get a lot of options for spam filtering, such as RBLs or SpamAssassin. With yahoo/hotmail you are totally at their mercy as far as spam-detection. And I have found that SpamAssassin is orders of magnitude better than anything yahoo has ever done.
Agreed. Since when did Ask Slashdot become "random tech support for anything tangentially related to open source"?
Uh, what? How does loading a 1x1 GIF reveal anything about your screen resolution? It will simply be another entry in a log file, which records the URL, the IP address, the time, the referer, and the user-agent. All of those fields are present in the log of the server that's serving the main html page.
In order to determine any further info about the user, you'd have to use Javascript to get this information from the DOM, and then somehow code that into a URL which gets submitted or posted to a server somewhere. From the blurb in the article there was no such code, just a simple IMG tag.
Please let this above post be a lesson to all you amateur spamfighters out there: You can never trust the From: line when dealing with spam.
If you have ever replied to spam, or "reported" spam or blacklisted based on who it purports to be from, then you are a part of the problem, not part of the solution. I use quotes around reported above because most people are clueless about what needs to be done to actually help the people that work the abuse desks.
joe jobs would not be nearly so annoying and devastating if there weren't so many people out there that just don't understand email. When you receive spam it makes you mad, and I guess the natural reaction is to reply to it or try to see who it's from and add that to a blacklist. But resist, as almost always the information you are basing this judgement on has absolutely nothing to do with who was responsible for that turdlet in your inbox.
Were those addresses that you put in the address book randomly created hotmail/yahoo (or any other popular free webmail) accounts? If so, I'm not surprised that they received spam within hours of being created. Spammers hammer the popular webmail servers with all sorts of permutations of letters and numbers, becuase with so many users they're likely to find enough addresses to make it worth their while. It doesn't mean that the infected machine had anything to do with it.
If you really wanted to prove something with your experiment, you need a control group. Pick to random strings and create those accounts on hotmail/yahoo/whatever. Then add one of them to the address book but keep the other completely secret. If the one you added to the address book truly gets significantly more spam then I'd be impressed by your experiment. Otherwise, it's unsubstantiated and unscientific.
Alternatively, you could create these new addresses on an obscure domain name that's not likely to be subject of a dictionary attack... If that was what you did then I'd also believe the findings.
You forgot to add "you insensitive clod!"
Because Rob and the rest of the "editors" feel that by not giving a shit about typos, incomplete sentences, run-on sentences, duplicates, and spelling mistakes it gives the site the amateurish feel that it had when they first started it for the hell of it in 1997. Of course, a lot of people argue that since now they're PAID to run slashdot and since the site has become huge, some level of actual "editing" should be required of these people that claim to be "editors" yet seem to have no idea what the job of an "editor" actually entails.
What the heck are you talking about? I didn't get infected by no stinkin' RPC garbage, and I don't use Zone Alarm or anything like it. It's called:
- Being up to date with WindowsUpdate
- Not allowing those ports access to the internet through the router
Please, Zone Alarm the only way? Riiiiight.
Did you read the article or just rush to post a comment? The amount of money that your coupon is worth has nothing to do with how many times you listen to something. The use of tracking the number of times you play a song is so that the artist that you get the most out of gets a corresponding part of your coupon -- but you can donate your coupon to whomever you want in any proportion if you so desire. You can listen to as many or as few artists as many times as you want, the tax is a small and constant amount. This has nothing to do with "pay per listen".
Take all of those arguments and apply them to any public thing: roads, parks, libraries, etc.
"We should have private roads where you purchase the rights to use whatever freeways you frequent from whoever built them. I don't trust that fat government issuing a gas tax to pay for things we all use like roads. I only use a few different routes anyway, why should I help pay for all those other suckers to use the roads I don't?"
"I don't want public parks, I'd rather sign up for private memberships to closed, gated-off parks. I don't trust some filthy organization is really using my tax dollars to pay rangers, maintenance staff, etc."
Nobody said you had to pay artists. I'm sure there would be some way to direct your coupon towards some charity or some organization, or even back to yourself. The point is that people -want- to drop a penny in the tip jar, but they just can't at the moment. If you don't want to support an artist or artists you can send your coupon along to whoever you want (yourself included), but the point is that most people DO INDEED want to support artists, and this makes it very easy to do that directly.
So you pay yourself. Okay, fine, that's no better than the current situation. The point is that most people DO want to support the artist in some way. But in the current situation they fell that doing this by paying for CDs is either unreansonable or is not benefiting the artist at all. In the proposed scheme, the user has a guarantee that the money is making it directly to the artist. And since it's rolled in with other purchases, it's much easier to justify "doing the right thing."
In other words, this allows people to honestly express their desire to support an artist, and this is a desire that the majority of people have. Sure, some will find a way to pay themselves, but that's fine: it's no worse than the current situation, and as long as those people are in the minority it doesn't matter at all.
It's kind of like that checkbox at the top of your taxes that asks if you want some amount to be donated to whatever-it-is. I assume that most people check that, because either way they're still paying the same amount in taxes, and the cause that it goes to is probably reasonably worthy to most people.
I find the whole notion presented in this article deplorable. What ISP is going to want to self-inflict a barrel customer rage? Where does all this money go? Who's in charge of verifying claims? What's to stop malicious users from filing false reports, or clandistinely installing software to incriminate an enemy's PC? How are all these ignorant customers supposed to be educated that they're suddenly liable for tens or hundreds of dollars in fines? If you just start arbitralily fining people, I don't care how little it is, you will bring down a boatloat of wrath and ire. People -hate- fees they don't know of in advance.
But one good thing would come of such a plan: egress filtering by all ISPs. This means that source-spoofed packets would be dropped before they get very far. It would make it significantly harder to spoof anything. No more RFC1918 packets on the public internet. If you ever run a public server on the internet, sometime try adding firewall rules to log and then drop all Bogon packets: those from unrouteable IP space, reserved or unallocated space, etc. You will be surprised how much of that stuff is floating around on the public internet, just soaking up legitimate bandwidth. Egress filtering would cause a much higher level of net-hygeine, in my opinion.