I realize I should clarify the 'load balance outgoing mail' comment.
You can certainly load balance the mail submission process for users. Set up ten servers with round-robin DNS, see if I care.
What I was taking issue with the idea that you have a bunch of mail on one server, and you hand it to other servers to make it 'go faster', which is incredibly silly. They're not going to send it any faster than you, and you're wasting time by passing mail around instead of sending it!
You don't load balance on outgoing stuff, you load balance on tasks, which means you try to give each server the same amount of work to start with.
What interface a packet goes out on is completely unimportant. What's important is that it goes out with an known IP. There are systems that have multiple IPs on different interfaces, but the mail server should know this. It's not quantum physics or anything. If your mail server knows enough to try outgoing mail via different interface, it damn well should know to give out different names in the HELO.
And NAT and load balancing are a red herrings.
You don't need to load balance outgoing servers, that's just crazy talk. I don't even know how you'd load balance outgoing mail, unless you're one of those loons who likes set up an 'incoming' mail submission server, and hand the mail, via NFS, to outgoing servers to mail out, in which case, STOP DOING THAT, as it's possibly the stupidest mail configuration in history.
If a mail server has a piece of mail to send, SEND IT. Passing it around via 'load balancing' isn't helping anything. You can do that to incoming mail, but it's completely insane for outgoing mail.
As for NAT...well, outgoing mail servers shouldn't be behind a NAT. However, if they are, they should be claiming to be the NAT, because, to the rest of the internet, they are.
I don't know in what universe it's a useful point to mention that you're removing invalid email address before you send mail to them. That mail wouldn't go through anyway! It's the valid addresses that are a problem.
But, hey, you gave me the last number. So...5%. That's about 200 pieces of mail you sent. And you got 8 valid responses, and 2 invalid.
So you sent out, basically, 192 spam messages, barring the occasional legit C/R you sent out that was ignored. (Which is also a failure of the system, it's just a failure that isn't spamming.)
To get 8.
To get 8 fucking messages, you sent 192. For every legitmate message you receive, 24 other people had to look at a spam you sent them.
Well, you're the moral paradigm I've come to expect from C/R people.
You know, that's the second time I've heard people complaining about blacklisted domains, and I have no idea what the fuck you're talking about. No one blacklists 'from' addresses as spammer domains except stupid users. (No spam fighter would ever claim email came fromyour website unless you were running an open formmail script, in which case, damn right they block you.)
Some blacklists list known spammer domains, but these are fairly well confirmed via the ownership of the domain. Many lists skip the actual 'sending spam' part and just list all domains owned by certain spammers.
What you are describing sounds like what people who have IPs near spammers go through. Can you point to one of these hundreds of domain blocklists that has listed you incorrect at some point?
The only thing I can think of is that you're running an affiliate system and can't keep your affiliates under control.
Don't use 127.0.0.1. Use 127.29.13.4, or some equally random address in the 127/8 loopback.
Spamming software is almost always poorly written. They'll filter out 127.0.0.1, but aren't smart enough to do anything else. Those bastards will probably try to deliver mail to themselves for a week.
No, the list was devised because fucktards think it's reasonably to run in propose a 'new' solution that's already been determined not to work. For several months in any new spam-fighting forum, 25% of the posts are people leaping in without doing any research at all, proposing solutions with quite obvious objections, like the fact they require spammers to cooperate or magic pixies or everyone to turn all their email over to Microsoft.
There are spam fighting tools that do not fall anywhere on the list, and they're already used.
They aren't blacklists, in the normal meaning of the word. There are domain based blacklists, but they almost always are 'domains known to be owned by spammers', or companies known to be spamming. They generally are run even tighter than IP blacklists...a domain appearing in a spam message is never enough to get on the lists.
However, there are a lot of stupid companies and mail providers who have a shared domain blacklist...if X people (sometimes as low as 2) block a domain, it's blocked for everyone.
C/R doesn't stop spam in general. It merely stops spam for you by redirecting it basically at random. I'll admit, a large fraction of the redirects vanish, but a lot of them end up hitting innocent people.
220 mailin.rfc2821compliant.com ESMTP CRLF
ehlo mailout.example.com CRLF
550 Policy violation, incorrect domain in HELO violated this site's policy of not accepting mail from stupid computers.
Or did you just happen to forget you can reject any command for 'policy violation'? All you have to do is make a policy that states you do not accept systems that don't know who they are. You can, within the RFC, reject all mail from people whose name starts with a 'q'. You're not required to accept anything at all, you're just required to accept it in a certain way if you do accept it, and reject it in a certain way if you do reject it.
That said, the RFC is seriously out of date WRT to how email functions. It implies, for example, that it's okay to run a relay, but you might want to limit it to known sites, when, currently, it's completely unacceptable to run a relay that's not limited in that way. It makes no mention at all of DNS-based blacklists, etc.
When a server HELOs, it's from that server. It's not part of the email message. Wherever an email message ends up, that server will make up its own hello to pass it on.
We don't want the MAIL FROM to match the server name, we want what the HELO command says the server name is to match the server name. It's not like the server can't know what its own goddamn name is.
Even servers on dynamic IPs can figure out their own rDNS name, it's a trivial lookup. So it wouldn't really work to block spam.
The only time it makes sense for that not to happen is when a server is behind a NAT and doesn't know it. Which can be trivially fixed by the server giving the NAT's name, instead, or just moving moving the network in a sane configuration.
This isn't really important, however. It's just a sign of the utter fucking incompetance of half the server admins out there.
In other words, you sent out 3992 pieces of spam to forged or invalid addresses, pissing off 2 people who knew what was going on bad enough that they confirmed your C/R.
Erm...all 2D sliding puzzles move the free space. The only other way to do is a puzzle is to have it rotate, and you can only only do that in one direction in 2D, which isn't much of a puzzle. (Rotate the clock hands until they read 3:47.)
(Note it would, in theory, be possible to build a 2D sliding puzzle with a back that 'tiles' could wrap around to, with no free space. I don't know of any puzzle like that, or if it would technically be 2D.)
Re:lets do it the hard way 'cause we got computers
on
Programming Puzzles
·
· Score: 1
A better way to state it might be:
Lay down 14 pieces. These pieces do not matter. They can go anyway.
If you put the two remaining pieces one way, it is solvable. If you put them the other way, it is unsolvable.
Any combination of 14 pieces already laid down results in two remaining possibilities, one right, and one not. Hence the probability is always 50/50.
You can even say 'We're going to make the first line 5, 2, 7, and 11.', and the odds are still the same, there's no way to alter them. (Well, unless you specificed 15 pieces, and then it would either be solvable or unsolvable with 100% odds, because there's only one possible position. But that's silly.)
I don't know why they have to star this Indiana. That would be pretty stupid, when all they have to do is give him a kid.
Assuming time's been running at the same pace, it should be about 1960 or so in Indiana's universe. Perfect to introduce a pre-teen son, or, to be non-traditional, a daughter. (Hopefully they do not bring the kid along, that's what screwed up Temple of Doom. Just two minutes of family life, please.)
Flashbacks explain that, ten years ago, at the end of WWII, Indiana fathered a kid and settled down, and this movie calls him back into action. (It would give Ford an excuse to be less physical, also, because he hasn't been running through temples for ten years.)
Then, for the next movie, they just skip ahead 10 years, cast a young stud or studette as the kid, and set it at the end of the 60s or start of the 70s. Have a 70 year old Indiana make a cameo. They can even have the kid go by 'Indiana Jr.', because they got saddled with a name they couldn't stand either.
I hereby public domain this idea, if any lawyers care.
They worked this way, let's see if we can make a table:
- 0 1 2 3 H - - X X L - X - X
So if the low burns out, or if you're using a normal light bulb, you get off-off-on-on as you click around. I can't count the number of 3-way lamps I've had to turn off and on twice because we always put normal bulbs in them. It gets so it's a reflex to flip the switch twice.
And you can use the 3-way bulbs, with the low burned out, as normal lightbulbs. (And, of course, you could do that anyway if the low wasn't burned out.)
However, the weirdness comes about if the high burns out. In that case, the pattern is off-on-off-on, as you can see above. However, you'd never notice that, because that looks identical to the off-on of normal lights! It's really funny if you forget it's a 3-way lamp because of that, and try to use the light bulb elsewhere, or stick a normal bulb in there.
3-way lamps are often confused with dimmers, but they aren't. Dimmers alter the flow of electricity to a normal bulb, I forget how exactly.
3-way lights, on the other hand, have two seperate filiments inside a single bulb. One 15 walt and one 40, or something like that. They have two seperate connections on the base to run in power. (And one to let the power back out, duh.)
Of these two incoming connections, the last one, the bright one, corrospond to the one incoming on a normal bulb, and on any fluorescent designed to plug into a normal socket. So if you put a normal bulb in, you get two 'offs' and two 'ons'.
This is why, when 3-way bulbs burn out, they sometimes burn out weird. You either lose the high power or the low power. If you lose high, you get off-on-off-on, and if you lose low, you get the off-off-on-on you also get from a regular bulb in a 3 way. (At that point, if you're clever, you can move said low-less bulb into a normal light instead of throwing it out.)
And, yes, you can put 3 ways on real dimmers, and they work just fine.
One of these chemicals, not in the base but coating the actual bulb, is mercury. (So you throw it away even in a normal fluorescent lamp, not just those that plug into a normal light socket.) It's what make fluorescent lights 'whitish' instead of ultraviolet, IIRC.
Actually, incandescent lights tend to last as long or even longer if you leave them on. It's the stress of startup that tends to kill the filament, and cooling after they shut off.
Instead of hours of operation, they really should have 'number of times off and on'. The average functioning time of a lightbulb without being turned off is unknowable, mainly because no company can test them that long.
Fluorescent lights aren't effected as much by startup, because, really, they turn off and on 60 times a second. (Although you can kill your ballast, which is a part of fluorescent lights that people tend to ignore. My mother's kitchen has had fluorescent lights for eight years, and one of the ballasts are already screwed up.)
I have no idea about the LEDs. LEDs don't have a problem with startup, as far as I know. Otherwise they wouldn't always be blinking at us.
Electricy heaters have 100% conversion of electricity into heat. This only matters to people who think the point of internal climate control is to generate heat the most effictively.
In the real word, we want to heat our house with the cheapest, best for enviroment, longest lasting, whatever method, not 'We can turn every watt of this into heat' method.
Why not just have warm air shipped to your house? Every bit of warm air that made it would be used to heat your house, wouldn't it? If you didn't want to ship it, you could just run your car engine, and put a pipe from the inside the car to your house...every erg of heat that made it to your house would be used to heat it!
The only thing electric heaters are good for is heating areas very quickly, like a fire, but without the disadvantages of a fire, and under circumstances where a heat pump doesn't work, like when it's -40 outside.
Of course, at that point, it'd be cheaper to have a gas heater instead of electric, but almost no one has both a heat pump and a gas heater, whereas all heat pumps have electric heaters built in. Gas is what you're really paying for when you purchase electricty, and if you burn it yourself you cut out that annoying conversion to electricity, which does generate heat...for the power company.
Slashdot Mirror
Yeah, that's the ticket. In addition to having to filter spam, I now should now have to keep up with the format of C/R messages to filter those too.
You can certainly load balance the mail submission process for users. Set up ten servers with round-robin DNS, see if I care.
What I was taking issue with the idea that you have a bunch of mail on one server, and you hand it to other servers to make it 'go faster', which is incredibly silly. They're not going to send it any faster than you, and you're wasting time by passing mail around instead of sending it!
You don't load balance on outgoing stuff, you load balance on tasks, which means you try to give each server the same amount of work to start with.
And NAT and load balancing are a red herrings.
You don't need to load balance outgoing servers, that's just crazy talk. I don't even know how you'd load balance outgoing mail, unless you're one of those loons who likes set up an 'incoming' mail submission server, and hand the mail, via NFS, to outgoing servers to mail out, in which case, STOP DOING THAT, as it's possibly the stupidest mail configuration in history.
If a mail server has a piece of mail to send, SEND IT. Passing it around via 'load balancing' isn't helping anything. You can do that to incoming mail, but it's completely insane for outgoing mail.
As for NAT...well, outgoing mail servers shouldn't be behind a NAT. However, if they are, they should be claiming to be the NAT, because, to the rest of the internet, they are.
But, hey, you gave me the last number. So...5%. That's about 200 pieces of mail you sent. And you got 8 valid responses, and 2 invalid.
So you sent out, basically, 192 spam messages, barring the occasional legit C/R you sent out that was ignored. (Which is also a failure of the system, it's just a failure that isn't spamming.)
To get 8.
To get 8 fucking messages, you sent 192. For every legitmate message you receive, 24 other people had to look at a spam you sent them.
Well, you're the moral paradigm I've come to expect from C/R people.
Fucker.
Some blacklists list known spammer domains, but these are fairly well confirmed via the ownership of the domain. Many lists skip the actual 'sending spam' part and just list all domains owned by certain spammers.
What you are describing sounds like what people who have IPs near spammers go through. Can you point to one of these hundreds of domain blocklists that has listed you incorrect at some point?
The only thing I can think of is that you're running an affiliate system and can't keep your affiliates under control.
Spamming software is almost always poorly written. They'll filter out 127.0.0.1, but aren't smart enough to do anything else. Those bastards will probably try to deliver mail to themselves for a week.
First confirm, then block.
If they're making you you see their spam, then you should make them see their spam.
If you want to be nice, you could write a form letter, confirm their spam, and then send the form letter explaining why you did that.
There are spam fighting tools that do not fall anywhere on the list, and they're already used.
Are you legally insane, by any chance?
However, there are a lot of stupid companies and mail providers who have a shared domain blacklist...if X people (sometimes as low as 2) block a domain, it's blocked for everyone.
C/R doesn't stop spam in general. It merely stops spam for you by redirecting it basically at random. I'll admit, a large fraction of the redirects vanish, but a lot of them end up hitting innocent people.
ehlo mailout.example.com CRLF
550 Policy violation, incorrect domain in HELO violated this site's policy of not accepting mail from stupid computers.
Or did you just happen to forget you can reject any command for 'policy violation'? All you have to do is make a policy that states you do not accept systems that don't know who they are. You can, within the RFC, reject all mail from people whose name starts with a 'q'. You're not required to accept anything at all, you're just required to accept it in a certain way if you do accept it, and reject it in a certain way if you do reject it.
That said, the RFC is seriously out of date WRT to how email functions. It implies, for example, that it's okay to run a relay, but you might want to limit it to known sites, when, currently, it's completely unacceptable to run a relay that's not limited in that way. It makes no mention at all of DNS-based blacklists, etc.
When a server HELOs, it's from that server. It's not part of the email message. Wherever an email message ends up, that server will make up its own hello to pass it on.
We don't want the MAIL FROM to match the server name, we want what the HELO command says the server name is to match the server name. It's not like the server can't know what its own goddamn name is.
Even servers on dynamic IPs can figure out their own rDNS name, it's a trivial lookup. So it wouldn't really work to block spam.
The only time it makes sense for that not to happen is when a server is behind a NAT and doesn't know it. Which can be trivially fixed by the server giving the NAT's name, instead, or just moving moving the network in a sane configuration.
This isn't really important, however. It's just a sign of the utter fucking incompetance of half the server admins out there.
In other words, you sent out 3992 pieces of spam to forged or invalid addresses, pissing off 2 people who knew what was going on bad enough that they confirmed your C/R.
I'm glad I'm not the only person who noticed that.
(Note it would, in theory, be possible to build a 2D sliding puzzle with a back that 'tiles' could wrap around to, with no free space. I don't know of any puzzle like that, or if it would technically be 2D.)
Lay down 14 pieces. These pieces do not matter. They can go anyway.
If you put the two remaining pieces one way, it is solvable. If you put them the other way, it is unsolvable.
Any combination of 14 pieces already laid down results in two remaining possibilities, one right, and one not. Hence the probability is always 50/50.
You can even say 'We're going to make the first line 5, 2, 7, and 11.', and the odds are still the same, there's no way to alter them. (Well, unless you specificed 15 pieces, and then it would either be solvable or unsolvable with 100% odds, because there's only one possible position. But that's silly.)
That's not fraud. It's not like the puzzle design was a secret.
It's perfectly allowable by the directions to take off all the labels, and thus solve the puzzle by making all the sides the same color.
Assuming time's been running at the same pace, it should be about 1960 or so in Indiana's universe. Perfect to introduce a pre-teen son, or, to be non-traditional, a daughter. (Hopefully they do not bring the kid along, that's what screwed up Temple of Doom. Just two minutes of family life, please.)
Flashbacks explain that, ten years ago, at the end of WWII, Indiana fathered a kid and settled down, and this movie calls him back into action. (It would give Ford an excuse to be less physical, also, because he hasn't been running through temples for ten years.)
Then, for the next movie, they just skip ahead 10 years, cast a young stud or studette as the kid, and set it at the end of the 60s or start of the 70s. Have a 70 year old Indiana make a cameo. They can even have the kid go by 'Indiana Jr.', because they got saddled with a name they couldn't stand either.
I hereby public domain this idea, if any lawyers care.
So if the low burns out, or if you're using a normal light bulb, you get off-off-on-on as you click around. I can't count the number of 3-way lamps I've had to turn off and on twice because we always put normal bulbs in them. It gets so it's a reflex to flip the switch twice.
And you can use the 3-way bulbs, with the low burned out, as normal lightbulbs. (And, of course, you could do that anyway if the low wasn't burned out.)
However, the weirdness comes about if the high burns out. In that case, the pattern is off-on-off-on, as you can see above. However, you'd never notice that, because that looks identical to the off-on of normal lights! It's really funny if you forget it's a 3-way lamp because of that, and try to use the light bulb elsewhere, or stick a normal bulb in there.
3-way lights, on the other hand, have two seperate filiments inside a single bulb. One 15 walt and one 40, or something like that. They have two seperate connections on the base to run in power. (And one to let the power back out, duh.)
Of these two incoming connections, the last one, the bright one, corrospond to the one incoming on a normal bulb, and on any fluorescent designed to plug into a normal socket. So if you put a normal bulb in, you get two 'offs' and two 'ons'.
This is why, when 3-way bulbs burn out, they sometimes burn out weird. You either lose the high power or the low power. If you lose high, you get off-on-off-on, and if you lose low, you get the off-off-on-on you also get from a regular bulb in a 3 way. (At that point, if you're clever, you can move said low-less bulb into a normal light instead of throwing it out.)
And, yes, you can put 3 ways on real dimmers, and they work just fine.
One of these chemicals, not in the base but coating the actual bulb, is mercury. (So you throw it away even in a normal fluorescent lamp, not just those that plug into a normal light socket.) It's what make fluorescent lights 'whitish' instead of ultraviolet, IIRC.
Instead of hours of operation, they really should have 'number of times off and on'. The average functioning time of a lightbulb without being turned off is unknowable, mainly because no company can test them that long.
Fluorescent lights aren't effected as much by startup, because, really, they turn off and on 60 times a second. (Although you can kill your ballast, which is a part of fluorescent lights that people tend to ignore. My mother's kitchen has had fluorescent lights for eight years, and one of the ballasts are already screwed up.)
I have no idea about the LEDs. LEDs don't have a problem with startup, as far as I know. Otherwise they wouldn't always be blinking at us.
Electricy heaters have 100% conversion of electricity into heat. This only matters to people who think the point of internal climate control is to generate heat the most effictively.
In the real word, we want to heat our house with the cheapest, best for enviroment, longest lasting, whatever method, not 'We can turn every watt of this into heat' method.
Why not just have warm air shipped to your house? Every bit of warm air that made it would be used to heat your house, wouldn't it? If you didn't want to ship it, you could just run your car engine, and put a pipe from the inside the car to your house...every erg of heat that made it to your house would be used to heat it!
The only thing electric heaters are good for is heating areas very quickly, like a fire, but without the disadvantages of a fire, and under circumstances where a heat pump doesn't work, like when it's -40 outside.
Of course, at that point, it'd be cheaper to have a gas heater instead of electric, but almost no one has both a heat pump and a gas heater, whereas all heat pumps have electric heaters built in. Gas is what you're really paying for when you purchase electricty, and if you burn it yourself you cut out that annoying conversion to electricity, which does generate heat...for the power company.