I agree that this is the reason ISPs block mail from dynamic IP addresses, but that doesn't mean they're right.
There are right ways and wrong ways to block spam. The wrong ways block a lot of mail that end users do wish to receive. Considerable violence is done to the end-to-end model that was responsible for the Internet's success. Arbitrary IP-level blocks on all dynamic IP addresses is a perfect example of a wrong way to block spam.
The right ways maintain the transparency of the internet, and leave end users in ultimate control of the email they receive. I have no objection to placing spam filters in the network for performance reasons (e.g., to reduce the traffic that would otherwise go over a user's slow link) but they must remain under end-user control. A Bayesian filter that diverts spam into a "Junk" folder on an IMAP server is an example of a right way to control spam. A user-controllable IP packet filter on the ISP end of a user's access link is another acceptable example.
All too many ISPs take heavy-handed approaches to stopping spam, and this has got to stop. The collateral damage from these methods is destroying the utility of email just as much (or more) than the spam itself.
So what's your take on the kind of VoIP that the FCC may be interested in regulating? Phone-to-phone VoIP, which looks to the parties just like ordinary POTS? Computer-to-phone VOIP, where the POTS network is used to deliver a call initiated on a computer? Computer-to-computer VOIP? Or all three?
I can understand the FCC being interested in the first category, and maybe even the second. So long as the FCC looks only at VoIP that involves one or more regular telephone calls, I can't get too overly excited. The harder they squeeze on such services, the faster everyone will move to true end-to-end VOIP that doesn't touch a telephone switch at all. And that will be a good thing.
But if they do try to control true, end-to-end VOIP, then things could turn ugly. End-to-end VOIP between computers is indistinguishable from any other Internet application, and rules that reach this mode of operation could do serious damage to the Internet model. As far as I'm concerned, the nature and the meaning of the data that consenting computers exchange over the Internet is none of the FCC's (or my carrier's) damn business. If they feel otherwise, they can try to crack my end-to-end encryption.
Still, considerable vigilance will be required to stop this threat. As we've seen with the Broadcast Flag debacle, technical realities have little sway at the Commission. They're lawyers and political hacks, not engineers, and are not about to let mere facts get in their way.
Amen! This is a perfect example of one of many serious threats to end-to-end transparency in the Internet. Between greedy service providers like Verisign that would break end-to-end for their own financial gain and overzealous and ill-conceived antispam mechanisms like dialup blacklisting, the end-to-end principle that made the Internet great is now in very serious jeopardy.
I don't know what can be done other than to find and promote better ways to fight spam at the endpoints, and to scream whenever an ISP does something really stupid.
Of course Microsoft implements lots of application functions in the operating system. This isn't just careless design; it's how they deliberately lock applications into Windows. That this results in hopelessly insecure
spaghetti code is an inevitable consequence of this policy.
This was the main point of the report on security published last month by Geer, Schneier et al.
I admire your deep insights into what "the unwashed masses" want.
Usage-sensitive billing need not result in higher bills. It has been well known for years that most residential telephone customers pay considerably more under flat-rate plans than they would under usage-sensitive plans.
Many creative service and pricing alternatives are available to the ISPs. Draconian laws like the one proposed in Tennessee are simply unnecessary. The one I've always liked is a continuous bandwidth auction. Each user has a "knob" that says how much they're currently willing to pay to send or receive a packet. Those bidding more get priority over those bidding less. If you prefer a flat-rated service, you simply bid zero. That lets you contend for whatever bandwidth is left over after the paying customers are served. If you're willing to wait, or if you don't want the complexity, great. You simply ignore the knob. But if you want your download to go faster, you have the option to turn up the knob and pay for the privilege.
It would take some work to implement this scheme, but it's now entirely doable and would, I think, satisfy pretty much everyone.
I suggest you go to college and take Economics 101. Pay special attention to the sections on demand-side elasticity and the "tragedy of the commons".
It may surprise you to learn that most commercial Internet users already pay by usage. They already have to deal with the problem of having to pay for DDos traffic. The answer here is not flat-rate billing, but new mechanisms to control such attacks. Routers should give the user of any IP address, without prior arrangement, the ability to create filters that block unwanted traffic to that address. We also need an "IP pushback" feature to automatically detect DoS and worm attacks and cut them off near their sources. This is a lot of work, but there's no alternative.
As for shady ISPs increasing revenue by dropping packets, perhaps you should take Networking 101 yourself and study how TCP behaves in the face of packet loss. TCP treats packet loss as an indicator of congestion and backs off its transmission rate to help the network. Deliberate packet dropping by the ISP in the absence of congestion would not only yield terrible performance that would drive the users to the competition, but also significantly reduce the number of packets entering the network and therefore the ISP's revenues.
Perhaps you misunderstood my original posting. I would not require any ISP to institute usage-sensitive charging. It's just that I strongly oppose laws to prop up draconian restrictions on the permitted uses of a network. To that end, I point out to the legislators that much less repressive alternatives are already available to the ISPs if abuse of flat-rate services really becomes a problem.
Seems to me the cable companies and telcos want this bill mainly to protect a fundamentally flawed business model -- the flat rate broadband plan.
Yes, residential customers really like flat rate plans because they know exactly how much they'll spend every month. But they have a Faustian downside: they give the carriers an excuse to severely limit and control how you use the service. Just as all-you-can-eat cafeterias have rules that regular restaurants do not (e.g., against sharing food or taking it home) most flat-rate broadband plans prohibit connection sharing, business use, running servers, etc.
If the carriers instead charged by usage for the shared part of their network, then they would have far less of an arguable case (i.e., none whatsoever) for claiming that a NAT box, even if you use it to provide service to your neighbor, constitutes "theft of service". If you pay for those bits, they're clearly yours to give away.
I know it's unpopular to argue for usage-based billing. But if I'm forced to choose (and I think I will be)
between flat rate plans with lots of heavy-handed restrictions and a pay-as-you-go plan with no restrictions at all, I know what I'd do.
Groups like those opposing the Tennessee bill should educate their lawmakers that it's simply not their job to protect unsustainable business models. Although broadband service is frequently provided over cable TV facilities, it is nothing like cable TV. With usage-based billing, even your average legislator
might see how analogies between NAT boxes, which support a two-way telecommunications service, and illegal cable descramblers, which gain access to a one-way broadcast service, simply don't apply.
Imagine also the public outrage that would finally be directed against Microsoft when end-users have to pay for all the traffic generated by their worm-infested machines. Not only might that create an incentive to get such machines quickly off the net, we just might see a lot of ordinary Joes defenestrating their copies of Windows. Clearly a good thing.
Even the MPAA and RIAA couldn't complain, since usage-sensitive billing would discourage file sharing. (We don't have to tell them that everyone would simply revert to the way music was widely pirated long before the Internet: by exchanging physical media.)
Oh, and the spammers would have to pay more, too. Wouldn't that alone make it worthwhile?
Let's compare apples with apples. If coverage weren't an issue, then nothing beats my home 802.11b network. It's faster than any of the commercial services, and it's free.
Coverage is everything in a wireless network. If it doesn't work reliably where you want to use it, then even a cheap rate is no bargain. So if we instead pick AT&T Wireless's GPRS network as having coverage more comparable to Verizon's 1x network, the price is exactly the same: $80/mo for unlimited traffic.
As for the cost of the PCMCIA card, I just checked Verizon's website. Their 1xRTT card is $200 with a $100 rebate. That makes it not much more expensive than their (admittedly overpriced) data cables, and probably not much more expensive than an external bluetooth adapter for your laptop. And these prices are likely to go down with volume, of course. In any event, the price of the card or cable is lost in the price of the service if you keep it for a year or two. Might as well get one and free up your phone for voice calls. No need for bluetooth or a data cable.
I have a Kyocera smartphone (CDMA phone + Palm pilot) and I back it up like any palm pilot: by dropping it into a desk stand and pushing a button. Works for me.
Who needs a data cable (at any price) to put a laptop on the net with cellular? Just get a Sierra Wireless CDMA 2000 1xRTT PCMCIA card; they're available for both Sprint and Verizon (the Verizon model I have is the 555). You stick it in your laptop, attach a little antenna (an optional extension cable is supplied), bring up PPP and away you go. Works on both Windows and Linux, and it doesn't interfere with the normal use of your cell phone. Verizon's 1xRTT coverage (which they call "ExpressNet") is remarkable; it really fills in the big gaps between public 802.11 base stations.
Verizon has just rolled out an even faster wireless IP service, 1xEV-DO, in San Diego and DC, with more cities to come. There's another Sierra Wireless PCMCIA card for it that also does 1xRTT.
Bluetooth was conceived at a time when cell phones were big and expensive, so maximizing their utility with short-range wireless links to small headsets, laptops, etc, made sense. But now that cell phones are tiny enough to be built directly into PCMCIA cards, who needs Bluetooth?
Actually, that's not true, at least not for the modems now in use.
Some early cable modem networks (e.g., those with Lan City gear) used "dumb bent pipe" repeaters at the head end. This required symmetric modulation and data rates on the upstream and downstream links. But the Motorola CyberSURFR and then the DOCSIS standard did away with symmetry. The typical upstream rate is only a few megabits/sec, and much of this is taken by polling overhead.
Uh, how about the right to due process and trial by jury if I am arrested and accused of wrongdoing? How about the right against secret government intrusions on my privacy? The right to talk privately to a lawyer representing me if I am accused of a crime? The right to choose my reading material without the government claiming the right to look secretly over my shoulder?
The list goes on. Just because I haven't yet had to excercise all these rights doesn't mean I don't consider them valuable.
I'm sure that as a fallback, you're next going to say that rights and freedoms don't mean anything if you're loved ones are dead and your buildings are in ruins. Well, guess what? Many Afghanis and Iraqis are now saying the same thing.
Yup, as far as I'm concerned my government is a bigger threat than al Qaeda.
In one of his post 9/11 interviews, Osama bin Laden openly gloated about the destruction of individual rights in the US. But he didn't do it. He didn't have the power to do it. But Bush and Ashcroft did.
Basically, since 9/11 the US has been suffering from the biggest case of anaphylactic shock in recorded history. If you don't know the term, look it up in a medical dictionary.
A new, open fiber instructure is indeed a possibility, but it will require a highly enlightened local government. And my few sample data points indicate that most local governments are anything but enlightened, especially when the subject is at all technical. They'll be easy marks for the money, FUD and bogus "grass roots" organizations of telcos and cable companies desperate to avoid the competition. The usual practice of your average local government is to give a utility franchise to whoever promises them the most money, not to the entity who will best serve the public with it.
When cable and DSL modems were designed, the assumption was that only a tiny fraction of residential customers would use the service. That drove the architectures in a certain direction. But when most of the homes on every block would like to get high speed Internet access if it's reasonably priced, then many more options open up. Some time ago I heard of some housing developments that were interconnected with Cat-5 cable. Whatever happened to them?
Yeah, that broadband stuff sounds pretty horrible from everything I've read. I can't believe that the FCC would even consider it.
I couldn't agree with you more about how the phone company breakup should have been done. The company owning and maintaining the wires in the streets should be regulated as a common carrier, all the stuff on both ends of those wires should be completely unregulated, and no company owning wires should be allowed to have any say or interest whatsoever in the stuff on the ends of the wires beyond some basic
electrical safety rules.
This is so obviously the right way to do things that one wonders how Congress, as corrupt and stupid as they are, could have gotten it so wrong.
That's a good question about remote vault access. I had assumed that the recent what-me-worry FCC decision that destroyed any vestige of competition in broadband Internet access had allowed the telcos to deny the CLECs access to those vaults, but I could be wrong. The telcos certainly spread their FUD effectively enough, whining about how "unfair" it would be to have to "give away" their broadband wires to their competitors. For some reason, the FCC ignored the many CLECs that went out of business because the telcos charged them more for the bare wires than they charged for their own DSL services.
Well, actually it's a duopoly, as I'm in one of the relatively few areas with both cable modem service (Road Runner) and DSL (SBC). Yes, Speakeasy DSL is technically available here, but it's substantially more expensive. It is also much slower; only 144 kb/s ISDL is available to my house, as I'm ~19,000 ft from the CO and I believe they do not have access to SBC's remote vaults.
You are quite right, Spamcop isn't perfect. I sometimes wonder why I bother with it at all, since I certainly haven't seen much of a reduction in my spam volume. It also takes time to manually verify all my spam before I submit it, because I don't want to falsely accuse somebody of spamming. (I once accidently reported myself as a spammer to Spamcop, but that's another story.)
But as Spamcop's fans point out, it's better than nothing. And the sustained DoS attacks that Spamcop has suffered in recent months tend to indicate that it's effective enough to really piss off some spammers.
You make a very good point. I like my backup MX because I can kick its queue right after I bring up my primary SMTP server and not have to wait for the remote senders to retry on their own. Also, I know I can lengthen my secondary's delivery timeouts if my primary server is ever down for an extended period (which has never happened).
Other than that, I'd be willing to let the original senders resend it as you do.
I just don't think it's an either/or thing. You've already implemented a scheme along the lines I've suggested that reduces much of the spam load on your links while leaving ultimate control in the hands of your users. That's all I ask.
Regarding reverse DNS, remember that I'm dealing with a large unregulated monopoly. They charge more for static IP addresses not because they're hard to provide, but because they can. They charge for reverse DNS because they can. They post my IP addresses in the residential ghetto because they can.
Guess what? Many MTAs also do not return 5xx messages during the SMTP transfer. Instead they swallow the message and generate a bounce to the usually-fake sender address. So this is nothing new.
Spammers may fake their return addresses, but I don't. So I can receive an error message from your client just as well as one from your MTA. Now if you want to send legitimate email with fake return addresses while still wanting to catch error returns, I can't help you.
What "other means" do you have in mind? Email has become so central to our way of life that we all have many people we only know how to reach by email. That's why heavy-handed spam filtering has become such a problem; real people can be really inconvenienced (or worse) by what the ISPs arbitrarily decide is "acceptable collateral damage".
Actually, you can perform the test during the RCPT command for the tuple of recipient and what IPs they want blocked. Then you can 550 the RCPT command if that user wants that IP blocked. There's no need to have all users agree.
That's basically what I meant. By "all users agree" I simply meant that the DATA transfer would not occur unless it is preceeded by at least one valid, nonblocking RCPT. Your idea accomplishes the same thing in a cleaner way; thanks.
Rejecting as much mail as possible during the SMTP session allows 5XX responses in place of queueing a bounce message (which in most spam cases can't be delivered anywhere anyway, so it sits in the queue for whatever the duration is set for).
I agree. It is important to send 5XX responses not only to avoid generating undeliverable bounce messages, but also to thwart spammers who use these bounce messages as a form of open third party relay. I put a quick stop to this on my own server when I first detected it, but I suspect a lot of servers will be harder to fix since they may not have a complete list of valid local recipients. This is true for many backup MX relays and for gateways outside company firewalls that are given as little information as possible in case they're compromised.
The only bandwidth that really matters is the bandwidth of your users' eyeballs. Everything else is hardware that's getting rapidly cheaper. Bayesian filters do a very good job of protecting my eyeballs.
It's been said many times before: spam is about conSent, not conTent.
Exactly! And that consent can only come from the ultimate recipient, not from you. Now your users may choose to delegate part or all of the task of spam filtering to you, but that's their choice and not yours. (You seem to be sensitive to this, so I'm not directing this at you personally but at the majority of ISPs that aggressively filter whether their users want it or not, and who can't be bothered with complaints about false positives.)
As for reverse DNS, this won't work in my case since both Road Runner and SBC (I have both cable modem and DSL) refuse to set up reverse DNS unless you pay an extortionate fee for the privilege of having what they call a "business" account. Since any spammer worth his salt can easily afford the "business" surcharge, this policy makes it clear that Road Runner and the other ISPs with similar policies aren't as interested in stopping spam as they are in simply getting a piece of the spammers' action.
"Filtering doesn't do a damn thing to stop spam"? Hello? It keeps it from reaching my eyeballs, which are the only resources that really matter. The rest is hardware that's getting cheaper all the time.
You have a good point about switching to an ISP that does just what I want. I already have. It's the "ISP" that I run at home for just my wife and me. Unfortunately, many ISPs make this increasingly difficult because they have decided that my IP address is a second-class citizen that cannot send them mail even if I have never spammed, or relayed spam, in my life.
It's fine for you to talk about how wonderful the free market is. But that's true only when the market actually is free. It isn't.
Slashdot Mirror
There are right ways and wrong ways to block spam. The wrong ways block a lot of mail that end users do wish to receive. Considerable violence is done to the end-to-end model that was responsible for the Internet's success. Arbitrary IP-level blocks on all dynamic IP addresses is a perfect example of a wrong way to block spam.
The right ways maintain the transparency of the internet, and leave end users in ultimate control of the email they receive. I have no objection to placing spam filters in the network for performance reasons (e.g., to reduce the traffic that would otherwise go over a user's slow link) but they must remain under end-user control. A Bayesian filter that diverts spam into a "Junk" folder on an IMAP server is an example of a right way to control spam. A user-controllable IP packet filter on the ISP end of a user's access link is another acceptable example.
All too many ISPs take heavy-handed approaches to stopping spam, and this has got to stop. The collateral damage from these methods is destroying the utility of email just as much (or more) than the spam itself.
I can understand the FCC being interested in the first category, and maybe even the second. So long as the FCC looks only at VoIP that involves one or more regular telephone calls, I can't get too overly excited. The harder they squeeze on such services, the faster everyone will move to true end-to-end VOIP that doesn't touch a telephone switch at all. And that will be a good thing.
But if they do try to control true, end-to-end VOIP, then things could turn ugly. End-to-end VOIP between computers is indistinguishable from any other Internet application, and rules that reach this mode of operation could do serious damage to the Internet model. As far as I'm concerned, the nature and the meaning of the data that consenting computers exchange over the Internet is none of the FCC's (or my carrier's) damn business. If they feel otherwise, they can try to crack my end-to-end encryption.
Still, considerable vigilance will be required to stop this threat. As we've seen with the Broadcast Flag debacle, technical realities have little sway at the Commission. They're lawyers and political hacks, not engineers, and are not about to let mere facts get in their way.
I don't know what can be done other than to find and promote better ways to fight spam at the endpoints, and to scream whenever an ISP does something really stupid.
This was the main point of the report on security published last month by Geer, Schneier et al.
I'm confused. This must be some strange new meaning of the term 'flat rate' with which I was previously unfamiliar. :-)
Usage-sensitive billing need not result in higher bills. It has been well known for years that most residential telephone customers pay considerably more under flat-rate plans than they would under usage-sensitive plans.
Many creative service and pricing alternatives are available to the ISPs. Draconian laws like the one proposed in Tennessee are simply unnecessary. The one I've always liked is a continuous bandwidth auction. Each user has a "knob" that says how much they're currently willing to pay to send or receive a packet. Those bidding more get priority over those bidding less. If you prefer a flat-rated service, you simply bid zero. That lets you contend for whatever bandwidth is left over after the paying customers are served. If you're willing to wait, or if you don't want the complexity, great. You simply ignore the knob. But if you want your download to go faster, you have the option to turn up the knob and pay for the privilege.
It would take some work to implement this scheme, but it's now entirely doable and would, I think, satisfy pretty much everyone.
It may surprise you to learn that most commercial Internet users already pay by usage. They already have to deal with the problem of having to pay for DDos traffic. The answer here is not flat-rate billing, but new mechanisms to control such attacks. Routers should give the user of any IP address, without prior arrangement, the ability to create filters that block unwanted traffic to that address. We also need an "IP pushback" feature to automatically detect DoS and worm attacks and cut them off near their sources. This is a lot of work, but there's no alternative.
As for shady ISPs increasing revenue by dropping packets, perhaps you should take Networking 101 yourself and study how TCP behaves in the face of packet loss. TCP treats packet loss as an indicator of congestion and backs off its transmission rate to help the network. Deliberate packet dropping by the ISP in the absence of congestion would not only yield terrible performance that would drive the users to the competition, but also significantly reduce the number of packets entering the network and therefore the ISP's revenues.
Perhaps you misunderstood my original posting. I would not require any ISP to institute usage-sensitive charging. It's just that I strongly oppose laws to prop up draconian restrictions on the permitted uses of a network. To that end, I point out to the legislators that much less repressive alternatives are already available to the ISPs if abuse of flat-rate services really becomes a problem.
Yes, residential customers really like flat rate plans because they know exactly how much they'll spend every month. But they have a Faustian downside: they give the carriers an excuse to severely limit and control how you use the service. Just as all-you-can-eat cafeterias have rules that regular restaurants do not (e.g., against sharing food or taking it home) most flat-rate broadband plans prohibit connection sharing, business use, running servers, etc.
If the carriers instead charged by usage for the shared part of their network, then they would have far less of an arguable case (i.e., none whatsoever) for claiming that a NAT box, even if you use it to provide service to your neighbor, constitutes "theft of service". If you pay for those bits, they're clearly yours to give away.
I know it's unpopular to argue for usage-based billing. But if I'm forced to choose (and I think I will be) between flat rate plans with lots of heavy-handed restrictions and a pay-as-you-go plan with no restrictions at all, I know what I'd do.
Groups like those opposing the Tennessee bill should educate their lawmakers that it's simply not their job to protect unsustainable business models. Although broadband service is frequently provided over cable TV facilities, it is nothing like cable TV. With usage-based billing, even your average legislator might see how analogies between NAT boxes, which support a two-way telecommunications service, and illegal cable descramblers, which gain access to a one-way broadcast service, simply don't apply.
Imagine also the public outrage that would finally be directed against Microsoft when end-users have to pay for all the traffic generated by their worm-infested machines. Not only might that create an incentive to get such machines quickly off the net, we just might see a lot of ordinary Joes defenestrating their copies of Windows. Clearly a good thing.
Even the MPAA and RIAA couldn't complain, since usage-sensitive billing would discourage file sharing. (We don't have to tell them that everyone would simply revert to the way music was widely pirated long before the Internet: by exchanging physical media.)
Oh, and the spammers would have to pay more, too. Wouldn't that alone make it worthwhile?
Coverage is everything in a wireless network. If it doesn't work reliably where you want to use it, then even a cheap rate is no bargain. So if we instead pick AT&T Wireless's GPRS network as having coverage more comparable to Verizon's 1x network, the price is exactly the same: $80/mo for unlimited traffic.
As for the cost of the PCMCIA card, I just checked Verizon's website. Their 1xRTT card is $200 with a $100 rebate. That makes it not much more expensive than their (admittedly overpriced) data cables, and probably not much more expensive than an external bluetooth adapter for your laptop. And these prices are likely to go down with volume, of course. In any event, the price of the card or cable is lost in the price of the service if you keep it for a year or two. Might as well get one and free up your phone for voice calls. No need for bluetooth or a data cable.
I have a Kyocera smartphone (CDMA phone + Palm pilot) and I back it up like any palm pilot: by dropping it into a desk stand and pushing a button. Works for me.
So does this mean we all get to kick Ballmer's ass?
Oh, I see. Microsoft disclaims all liability for their own products. So it's "do as we say, not as we do."
Verizon has just rolled out an even faster wireless IP service, 1xEV-DO, in San Diego and DC, with more cities to come. There's another Sierra Wireless PCMCIA card for it that also does 1xRTT.
Bluetooth was conceived at a time when cell phones were big and expensive, so maximizing their utility with short-range wireless links to small headsets, laptops, etc, made sense. But now that cell phones are tiny enough to be built directly into PCMCIA cards, who needs Bluetooth?
Some early cable modem networks (e.g., those with Lan City gear) used "dumb bent pipe" repeaters at the head end. This required symmetric modulation and data rates on the upstream and downstream links. But the Motorola CyberSURFR and then the DOCSIS standard did away with symmetry. The typical upstream rate is only a few megabits/sec, and much of this is taken by polling overhead.
The list goes on. Just because I haven't yet had to excercise all these rights doesn't mean I don't consider them valuable.
I'm sure that as a fallback, you're next going to say that rights and freedoms don't mean anything if you're loved ones are dead and your buildings are in ruins. Well, guess what? Many Afghanis and Iraqis are now saying the same thing.
In one of his post 9/11 interviews, Osama bin Laden openly gloated about the destruction of individual rights in the US. But he didn't do it. He didn't have the power to do it. But Bush and Ashcroft did.
Basically, since 9/11 the US has been suffering from the biggest case of anaphylactic shock in recorded history. If you don't know the term, look it up in a medical dictionary.
When cable and DSL modems were designed, the assumption was that only a tiny fraction of residential customers would use the service. That drove the architectures in a certain direction. But when most of the homes on every block would like to get high speed Internet access if it's reasonably priced, then many more options open up. Some time ago I heard of some housing developments that were interconnected with Cat-5 cable. Whatever happened to them?
Yeah, that broadband stuff sounds pretty horrible from everything I've read. I can't believe that the FCC would even consider it.
This is so obviously the right way to do things that one wonders how Congress, as corrupt and stupid as they are, could have gotten it so wrong.
That's a good question about remote vault access. I had assumed that the recent what-me-worry FCC decision that destroyed any vestige of competition in broadband Internet access had allowed the telcos to deny the CLECs access to those vaults, but I could be wrong. The telcos certainly spread their FUD effectively enough, whining about how "unfair" it would be to have to "give away" their broadband wires to their competitors. For some reason, the FCC ignored the many CLECs that went out of business because the telcos charged them more for the bare wires than they charged for their own DSL services.
Well, actually it's a duopoly, as I'm in one of the relatively few areas with both cable modem service (Road Runner) and DSL (SBC). Yes, Speakeasy DSL is technically available here, but it's substantially more expensive. It is also much slower; only 144 kb/s ISDL is available to my house, as I'm ~19,000 ft from the CO and I believe they do not have access to SBC's remote vaults.
But as Spamcop's fans point out, it's better than nothing. And the sustained DoS attacks that Spamcop has suffered in recent months tend to indicate that it's effective enough to really piss off some spammers.
Other than that, I'd be willing to let the original senders resend it as you do.
Regarding reverse DNS, remember that I'm dealing with a large unregulated monopoly. They charge more for static IP addresses not because they're hard to provide, but because they can. They charge for reverse DNS because they can. They post my IP addresses in the residential ghetto because they can.
The spammers aren't the ones blocking my email. I blame the ISPs because they're the ones doing it.
Spammers may fake their return addresses, but I don't. So I can receive an error message from your client just as well as one from your MTA. Now if you want to send legitimate email with fake return addresses while still wanting to catch error returns, I can't help you.
What "other means" do you have in mind? Email has become so central to our way of life that we all have many people we only know how to reach by email. That's why heavy-handed spam filtering has become such a problem; real people can be really inconvenienced (or worse) by what the ISPs arbitrarily decide is "acceptable collateral damage".
Exactly! And that consent can only come from the ultimate recipient, not from you. Now your users may choose to delegate part or all of the task of spam filtering to you, but that's their choice and not yours. (You seem to be sensitive to this, so I'm not directing this at you personally but at the majority of ISPs that aggressively filter whether their users want it or not, and who can't be bothered with complaints about false positives.)
As for reverse DNS, this won't work in my case since both Road Runner and SBC (I have both cable modem and DSL) refuse to set up reverse DNS unless you pay an extortionate fee for the privilege of having what they call a "business" account. Since any spammer worth his salt can easily afford the "business" surcharge, this policy makes it clear that Road Runner and the other ISPs with similar policies aren't as interested in stopping spam as they are in simply getting a piece of the spammers' action.
You have a good point about switching to an ISP that does just what I want. I already have. It's the "ISP" that I run at home for just my wife and me. Unfortunately, many ISPs make this increasingly difficult because they have decided that my IP address is a second-class citizen that cannot send them mail even if I have never spammed, or relayed spam, in my life.
It's fine for you to talk about how wonderful the free market is. But that's true only when the market actually is free. It isn't.