Hmmm. Not only would this move the cost to the sender, but it would also prevent spoofing. The mail client could ignore mail from "dotted quad" addresses and only go to legitimate DNS entries. It would be nice to have dates in DNS entries, too, indicating that the domain has been around for a while.
You need to work out some kind of public key authentication so that nobody else can swipe your email from the server, but you don't have to exchange unique keys with every mail server on the internet.
Clean up the idea, buy some flameproof underwear, and post it on NANAE.
I noticed that I got joed after reporting spam. It would make sense for the spammers to try to kill off the reporters. Now I remove all identifying info when I report spam (including message IDs) and send the report from a gmail account. I also post that way on NANAS.
paulgraham.com is 66.163.161.45. When I check IP on either side, they are not listed. Only that one IP address is listed. Paul Graham is not a victim of collateral damage of a wideranging netblock, his single IP has been blocked because it was the previous address of textileshop.com. Its in the sbl evidence file SBL27945.
host paulgraham.com paulgraham.com has address 66.163.161.45
host 66.163.161.45 45.161.163.66.in-addr.arpa domain name pointer html3.store.vip.sc5.yahoo.com
44.161.163.66.sbl-xbl.spamhaus.org not found: 3(NXDOMAIN) 45.161.163.66.sbl-xbl.spamhaus.org has address 127.0.0.2 46.161.163.66.sbl-xbl.spamhaus.org not found: 3(NXDOMAIN)
This is supposed to be a review of high-end power supply specs and it never mentions ripple? How about overvoltage protection? Aging? Brownout or surge protection? So how stable is the output of this supply? Real stable. Oh, OK.
Then there are statements like this in the review: "Cheap power supply units... may blow up". Yeah, I think I saw this on Star Trek last week. If they would only stop making those cheap supplies out of TNT.
That's what this whole argument is about. To incorporate the platform-specific switches in the main gcc code, for example, would slow down every future gcc release to test on symbian/nokia. The alternative is to keep gcc simple(r) and let the minor platform (symbian in this case) use their own branch. There's something to be said for either approach.
If you slow down development to cover a broader set of platforms then you will be late with the latest new buzzword features (hyperthreaded keyboard support!). On the other hand, who gets to decide what counts as a minor platform? Do you decide based on age? (sorry, Amiga is too old) or based on number of units? (sorry, not enough IBM mainframes out there) or just interest level of the developers? Anything that you abandon is going to cause big problems to someone.
The truth everywhere I've worked is closer to the opposite. Professors supplying 75% of the ideas and the student who writes the paper taking all of the credit. Brash young pups bark but don't have much bite.
sendmail already tests that the sender domain of every email exists in a nameserver. How does adding user authentication add anything to that? Also, who is going to determine whether a domain is "legit" vs. "fly-by-night"? This single international database of which you speak, where will it be located? Will Cuban ISP's be listed in it? Palestinian? Taiwanese? Have you figured out why this is impossible yet?
Cringely never gets more than about 50% correct in his articles. In this case he calls it "raw tcp/ip sockets". Wrong. Raw sockets access IP, so you can forge tcp packets in a DOS attack. Every OS allows access to TCP/IP. How else would your browser work?
He then proposes a secure ID system. Gee. Maybe if every connection to the network had a unique 32-bit number that could be traced somehow? Maybe there could be a world-wide database connecting names and administrative information to these numbers? If only that were possible. Thanks, Bob.
Raw sockets isn't just for lying about your source address. It can be used if you want to write your own TCP/IP imlementation, write a new protocol on top of IP, or implement an ICMP packet that you can't do any other way, to give a few examples.
That said, its not clear that any of this HAS to work in user mode.
Most CPU cycles go unused on workstations and servers. iometer is probably a better benchmark for a non-compute-bound setting. It doesn't make much sense to have a 3GHz CPU and a single ATA hard drive.
Nope. C isn't a scripting language. Any language where the first 50 lines of your code are used to invoke obscure rituals cannot, by definition, be a scripting language. perl starts with the #! line, then can get right to: print "Hello, World\n" on the second line.
I used to do some work with video equipment. Shielded cable usually has a percentage of covering specified. If you cut apart a piece of cheap coax cable, you will find very thin braid and a lot of open space. Expensive coax looks like a solid layer. So I think your eyes are correct. There is a difference both in signal loss and in pickup of video noise.
At audio frequencies, the difference is going to be a lot less. I've always used untwisted pair (Radio Shack speaker cable) and never noticed any problem even on long runs.
The article is misquoting. Metcalfe said something like "usefulness" of a network. The squared term is because there are two endpoints to each connection, so it makes sense that the usefulness goes up as the number of possible connections.
Slashdot Mirror
With Verizon, I have to use a port redirector to run my own email, and their service sucks, but outages are rare.
Controllers? Is there a way to support 4 wheel-mice on a PC? The rest would be easy, I think.
mens sana in corpore sano
The best game ever for lunch breaks was 4-player Warlords on the 2600. Network Snipes is good for a variable number of players, too.
Don't leave out "Bill, the Galactic Hero"
Hmmm. Not only would this move the cost to the sender, but it would also prevent spoofing. The mail client could ignore mail from "dotted quad" addresses and only go to legitimate DNS entries. It would be nice to have dates in DNS entries, too, indicating that the domain has been around for a while.
You need to work out some kind of public key authentication so that nobody else can swipe your email from the server, but you don't have to exchange unique keys with every mail server on the internet.
Clean up the idea, buy some flameproof underwear, and post it on NANAE.
Except that when you get a virus from your friend, it really did come from your friend. So no, spf will not help here.
I noticed that I got joed after reporting spam. It would make sense for the spammers to try to kill off the reporters. Now I remove all identifying info when I report spam (including message IDs) and send the report from a gmail account. I also post that way on NANAS.
"Have I missed anything?"
I'm a frayed knot.
Cisco - Corporate Internet System Completely Obfuscated.
Have you ever been forced to program one of their routers? Gack. They must make all their money selling the courses.
I have Linksys and Netgear wireless links - Netgear wins. Less trouble to set up securely, and doesn't randomly forget what it was doing.
So I guess you wrote your own mail program? You wouldn't want to give Eric Allman control over your mail.
paulgraham.com is 66.163.161.45. When I check IP on either side, they are not listed. Only that one IP address is listed. Paul Graham is not a victim of collateral damage of a wideranging netblock, his single IP has been blocked because it was the previous address of textileshop.com. Its in the sbl evidence file SBL27945.
host paulgraham.com
paulgraham.com has address 66.163.161.45
host 66.163.161.45
45.161.163.66.in-addr.arpa domain name pointer html3.store.vip.sc5.yahoo.com
44.161.163.66.sbl-xbl.spamhaus.org not found: 3(NXDOMAIN)
45.161.163.66.sbl-xbl.spamhaus.org has address 127.0.0.2
46.161.163.66.sbl-xbl.spamhaus.org not found: 3(NXDOMAIN)
This is supposed to be a review of high-end power supply specs and it never mentions ripple? How about overvoltage protection? Aging? Brownout or surge protection? So how stable is the output of this supply? Real stable. Oh, OK.
... may blow up". Yeah, I think I saw this on Star Trek last week. If they would only stop making those cheap supplies out of TNT.
Then there are statements like this in the review: "Cheap power supply units
That's what this whole argument is about. To incorporate the platform-specific switches in the main gcc code, for example, would slow down every future gcc release to test on symbian/nokia. The alternative is to keep gcc simple(r) and let the minor platform (symbian in this case) use their own branch. There's something to be said for either approach.
If you slow down development to cover a broader set of platforms then you will be late with the latest new buzzword features (hyperthreaded keyboard support!). On the other hand, who gets to decide what counts as a minor platform? Do you decide based on age? (sorry, Amiga is too old) or based on number of units? (sorry, not enough IBM mainframes out there) or just interest level of the developers? Anything that you abandon is going to cause big problems to someone.
The truth everywhere I've worked is closer to the opposite. Professors supplying 75% of the ideas and the student who writes the paper taking all of the credit. Brash young pups bark but don't have much bite.
Also, the birth rate is falling. The population is increasingly composed of older monkeys at the typewriters.
"How many bills had Bush vetoed?"
You have to be able to read them to veto them.
You made the common error of backing up your claim with facts. Please re-read the posting FAQ. Well-supported arguments are not allowed here.
sendmail already tests that the sender domain of every email exists in a nameserver. How does adding user authentication add anything to that? Also, who is going to determine whether a domain is "legit" vs. "fly-by-night"? This single international database of which you speak, where will it be located? Will Cuban ISP's be listed in it? Palestinian? Taiwanese? Have you figured out why this is impossible yet?
Cringely never gets more than about 50% correct in his articles. In this case he calls it "raw tcp/ip sockets". Wrong. Raw sockets access IP, so you can forge tcp packets in a DOS attack. Every OS allows access to TCP/IP. How else would your browser work?
He then proposes a secure ID system. Gee. Maybe if every connection to the network had a unique 32-bit number that could be traced somehow? Maybe there could be a world-wide database connecting names and administrative information to these numbers? If only that were possible. Thanks, Bob.
Raw sockets isn't just for lying about your source address. It can be used if you want to write your own TCP/IP imlementation, write a new protocol on top of IP, or implement an ICMP packet that you can't do any other way, to give a few examples.
That said, its not clear that any of this HAS to work in user mode.
My favorite line:
"Such a thought is, of course, unthinkable"
Most CPU cycles go unused on workstations and servers. iometer is probably a better benchmark for a non-compute-bound setting. It doesn't make much sense to have a 3GHz CPU and a single ATA hard drive.
Nope. C isn't a scripting language. Any language where the first 50 lines of your code are used to invoke obscure rituals cannot, by definition, be a scripting language. perl starts with the #! line, then can get right to: print "Hello, World\n" on the second line.
I used to do some work with video equipment. Shielded cable usually has a percentage of covering specified. If you cut apart a piece of cheap coax cable, you will find very thin braid and a lot of open space. Expensive coax looks like a solid layer. So I think your eyes are correct. There is a difference both in signal loss and in pickup of video noise.
At audio frequencies, the difference is going to be a lot less. I've always used untwisted pair (Radio Shack speaker cable) and never noticed any problem even on long runs.
The article is misquoting. Metcalfe said something like "usefulness" of a network. The squared term is because there are two endpoints to each connection, so it makes sense that the usefulness goes up as the number of possible connections.