spam filtering strikes again
I met mexgeek and we bonded over our interest in cheap/free entertainment. He asked me to email him and keep him in the loop about invites if I were to go out. So I sent him a few e-mails about events, with hyperlinks to information describing them, not getting any response.
Relating this story to a group of friends, they said: "When you don't get a response, that's supposed to be the way of knowing that they got your message and don't want to reply." Of course, I argued that is a stupid way to think, and we shouldn't encourage people to express their preferences through silence.
As luck would have it, I found him in person. I told him the story, and he said he'd never gotten the messages. Unlike other people who aren't tech-savvy enough to be able to go in and figure out why a spam filter would reject your message, he could. Here were the analyses:
(Message 1)
BodySum: dbc7793cb26a422182ac1987a992d671
SpamPts: 95
Verified:no
ScoreLog: NEW(5) FREE(15) DOLLAR(15) PERC(15) HTTP(20) FROM-YAHOO(25)
Length: 2600
(Message 2)
BodySum: e94d1cde7a4e7e335bd197088c0aa5e6
SpamPts: 105
Verified:no
ScoreLog: FREE(15) DOLLAR(15) HTTP(20) FROM-YAHOO(25) EXCLAMATION(30)
Length: 2518
(Message 3)
BodySum: 24ad4c03f7be3556ce7160ec49b2b8f3
SpamPts: 85
Verified:no
ScoreLog: DEAL(10) FREE(15) DOLLAR(15) HTTP(20) FROM-YAHOO(25)
Length: 3904
This points out a fundamental flaw in filtering. Since spam capitalizes on things that intrigue us, we will be very disappointed when we throw out mail for offers that turn out to be sincere ("free money and sex!") Most distressing is when you don't get a "bounce" (or a SMTP reject) notification to tell you that your message was not relayed. Very, very annoying. I suppose if you want someone to get a message you should not send them weblinks and avoid mentioning anything that sounds good.
I wrote this treatise on The Importance of Reliable Communication
I've noticed a growing tendency of people to use silence to indicate a rejection of a message, and a cultural assumption that this is the intention when a response is not received. This sets a dangerous precedent-especially in an era where internet services are automatically throwing away content that has been mechanically analyzed as spam. The practice facilitates interpersonal misunderstanding, as well as the abuse of power by governments or corporations that wish to suppress information about competitors (this has happened on major sites).
You would think that senders would get some kind of indication when an automatic mechanism sweeps a message under the rug. It is claimed that this is not done because it would multiply the impact of spam-since the networks would now have to carry the traffic for the mail AND the receipts. The second argument is that giving the feedback would allow spammers to quickly evolve their messages until they beat the tests. This is faulty reasoning, and I think we should be very wary of relying on an infrastructure which is not able to provide a bounce message to blocked mail.
Ideally each person who received a question would respond. This response doesn't have to be as lengthy as what was received, but should at least depend somehow on what was written. Any formulated and automated response-no matter how instant and "informative"-leads to the same systematic weaknesses as giving no response. Being "overwhelmed" with messages is not an excuse for ignoring these policies, because if someone is that popular (perhaps due to celebrity status) then they shouldn't have trouble delegating the duty of responding. A policy of not being upset by those who repeat attempts to communicate is also a very important key to the process.
Relating this story to a group of friends, they said: "When you don't get a response, that's supposed to be the way of knowing that they got your message and don't want to reply." Of course, I argued that is a stupid way to think, and we shouldn't encourage people to express their preferences through silence.
As luck would have it, I found him in person. I told him the story, and he said he'd never gotten the messages. Unlike other people who aren't tech-savvy enough to be able to go in and figure out why a spam filter would reject your message, he could. Here were the analyses:
(Message 1)
BodySum: dbc7793cb26a422182ac1987a992d671
SpamPts: 95
Verified:no
ScoreLog: NEW(5) FREE(15) DOLLAR(15) PERC(15) HTTP(20) FROM-YAHOO(25)
Length: 2600
(Message 2)
BodySum: e94d1cde7a4e7e335bd197088c0aa5e6
SpamPts: 105
Verified:no
ScoreLog: FREE(15) DOLLAR(15) HTTP(20) FROM-YAHOO(25) EXCLAMATION(30)
Length: 2518
(Message 3)
BodySum: 24ad4c03f7be3556ce7160ec49b2b8f3
SpamPts: 85
Verified:no
ScoreLog: DEAL(10) FREE(15) DOLLAR(15) HTTP(20) FROM-YAHOO(25)
Length: 3904
This points out a fundamental flaw in filtering. Since spam capitalizes on things that intrigue us, we will be very disappointed when we throw out mail for offers that turn out to be sincere ("free money and sex!") Most distressing is when you don't get a "bounce" (or a SMTP reject) notification to tell you that your message was not relayed. Very, very annoying. I suppose if you want someone to get a message you should not send them weblinks and avoid mentioning anything that sounds good.
I wrote this treatise on The Importance of Reliable Communication
I've noticed a growing tendency of people to use silence to indicate a rejection of a message, and a cultural assumption that this is the intention when a response is not received. This sets a dangerous precedent-especially in an era where internet services are automatically throwing away content that has been mechanically analyzed as spam. The practice facilitates interpersonal misunderstanding, as well as the abuse of power by governments or corporations that wish to suppress information about competitors (this has happened on major sites).
You would think that senders would get some kind of indication when an automatic mechanism sweeps a message under the rug. It is claimed that this is not done because it would multiply the impact of spam-since the networks would now have to carry the traffic for the mail AND the receipts. The second argument is that giving the feedback would allow spammers to quickly evolve their messages until they beat the tests. This is faulty reasoning, and I think we should be very wary of relying on an infrastructure which is not able to provide a bounce message to blocked mail.
Ideally each person who received a question would respond. This response doesn't have to be as lengthy as what was received, but should at least depend somehow on what was written. Any formulated and automated response-no matter how instant and "informative"-leads to the same systematic weaknesses as giving no response. Being "overwhelmed" with messages is not an excuse for ignoring these policies, because if someone is that popular (perhaps due to celebrity status) then they shouldn't have trouble delegating the duty of responding. A policy of not being upset by those who repeat attempts to communicate is also a very important key to the process.