Maintaining abuse desks. Not fun. I know, I work at one. Pulling 12 hour+ workdays, reading tons of email, and never catching up to the flood of spam and complaints.
Differentiating between spam and complaints is a non-trivial problem. Most clued administrators don't block by domain, but by IP address. This reduces the problem of blockages considerably.
And if you ever have to deal with Americans, you would know that they do the conference calling at 2 am IST, instead of a sane hour leaving the Indian to work 18 hours/day. And then they are rude on top of that.
Korea might have different usage scenarios from the US. And while new stuff is being built with Linux support, the others users can have a 3 year migration plan while the sysadmins get experience. Think longer term, and you might even see the logic.
Hopefully, the Korean hardware manufacturers will improve their Linux support.
And I have given up trying to keep Windows working on badly supported hardware. Linux works very well with supported hardware. "It just works" is the same argument that Mac OS users use, and I will use the same thing for my Linux box.
Sure, you can use delays in the SMTP dialog. Not a very difficult thing to do, and it works quite well.
However, you have to keep in mind that the delay also ties up a smtpd listener process/thread on your system, and spammers have more bandwidth, and more sending hosts than you do.
Essentially, this is the same problem as trying to deal with a ping flood from a few hundred thousand IP addresses, except that this is a tcp packet flood. When your pipes are choked, there isn't much you can do about it, except buy bigger pipes.
You proposal would require that we accept all those messages (or equivalent envelope) and keep a lookout for mail getting deleted by others. DJBs proposal for Internet mail 2000 works very well with qmail, which always sends one message per recipient.
Dealing with spam takes a little more effort in handling complexity.
Zombie hosts have the same IP for days. You assume responsible ISPs, which are _not_ the problem. The same ISPs which do nothing about zombies spewing stuff are the ones with the problem.
How do you propose to deal with mailing lists? I am on quite a few lists, where I read the list messages once a month (or even once in 6 months), and only glance over the list otherwise to ensure that I am not missing important stuff. Keep in mind that when I do read the list, I read a few thousand messages at a single shot.
These lists will simply become unusable, or unrunnable.
Oh, and _how_ does your solution stop spam finally?
Now you have to be online to retrieve mail and connect to lots of different servers, and then download messages. That is the equivalent of POP3, only from a few thousand servers instead of the one (zero for me, actually). More of my time _and_ bandwidth is being spent in trying to deal with mail, rather than less.
Whereas with my current server, the mail lands up directly in my inbox, all sorted and filtered for my reading pleasure.
You are assuming that the certifying authorities will revoke the certificate, and people will get to know about the revocation fast enough to prevent the mail from being accepted in the first place.
Oh, and you had better be willing to store messages for months (my priority queue has unread messages that old). And then you have the issue of users who are not always online.
And also consider that if a message is sent to five recipients at a single domain, only one copy of that message is sent. With your proposal, 5 downloads happen. This can be a significant issue for mailing lists.
To overcome greylisting, a spammer must be willing to implement a full mail-server on thier end.
No. The spammer just needs to resend the same message after $time when a majority of greylisting periods expire. Keep in mind that the spammer is using zombies, so it doesn't pay any real money.
idiotic sys admins who firewall out entire IP blocks for, seemingly, no reason.
Or they just decided that they didn't do any significant business with the owners of thatIP block which warranted lifting the block. Also keep in mind ISPs like UUNet/MCI (now Verizon) which used ordinary users as shields while allowing spammers to spam unhindered that network.
The only thing that worked was wholesale blocking of mail from all of UUnet until their users left (or complained enough that the cost of keeping the spammer was more than the cost of handling complaints). SPEWS worked, for a _very_ good reason. If ISPs don't clean up, we will be glad to move back to using SPEWS to block spam.
You support a spamming ISP, you get blocked. If you don't like it, vote with your money.
Why would a combination of the small changes not produce sexual incompatibility? One or two changes would lead to sterility (such as mules), and from there on, the differences just increase.
Nobody wants to hire someone who is going to challenge them.
Some of us would love to be able to hire someone better than us, or who would challenge us. If only because it makes the work more enjoyable. Which is why we have three admin type positions empty for 6 months.
And Google cares about Fortune 50 company users accessing them from work because? Unless they make for a significant proportion of ad revenue, Google really doesn't care.
Your value to Google is the number of eyeballs you can offer them, or the advertising revenue they make from you. Do Fortune 50 corporations offer enough eyeballs to be a globally significant number?
Spam is NOT about freedom of speech. Spam is about Unsolicited Bulk Messaging. No one is taking away your right to freedom of speech. You are free to stand on a soapbox and make a speech, setup a website, print posters and flyers and distribute them... On the other hand, my mailbox is _my_ property. Your freedom of speech does not exist there.
Well, all that you need is a fission device _just_ below the threshold of explosion. The additional neutrons from the fusion reactor would then trigger it off.
Maintaining abuse desks. Not fun. I know, I work at one. Pulling 12 hour+ workdays, reading tons of email, and never catching up to the flood of spam and complaints.
Differentiating between spam and complaints is a non-trivial problem. Most clued administrators don't block by domain, but by IP address. This reduces the problem of blockages considerably.
Uhm, typical salaries would be sitting at ~ 30K/year at the top of the technical chain. 65K/yr is pretty much senior management.
And if you ever have to deal with Americans, you would know that they do the conference calling at 2 am IST, instead of a sane hour leaving the Indian to work 18 hours/day. And then they are rude on top of that.
Korea might have different usage scenarios from the US. And while new stuff is being built with Linux support, the others users can have a 3 year migration plan while the sysadmins get experience. Think longer term, and you might even see the logic.
Hopefully, the Korean hardware manufacturers will improve their Linux support.
And I have given up trying to keep Windows working on badly supported hardware. Linux works very well with supported hardware. "It just works" is the same argument that Mac OS users use, and I will use the same thing for my Linux box.
Sure, you can use delays in the SMTP dialog. Not a very difficult thing to do, and it works quite well.
However, you have to keep in mind that the delay also ties up a smtpd listener process/thread on your system, and spammers have more bandwidth, and more sending hosts than you do.
Essentially, this is the same problem as trying to deal with a ping flood from a few hundred thousand IP addresses, except that this is a tcp packet flood. When your pipes are choked, there isn't much you can do about it, except buy bigger pipes.
The usenet reference was wrt downloading messages from a server (pull instead of push).
This is what we see at work: Graph
You proposal would require that we accept all those messages (or equivalent envelope) and keep a lookout for mail getting deleted by others. DJBs proposal for Internet mail 2000 works very well with qmail, which always sends one message per recipient.
Dealing with spam takes a little more effort in handling complexity.
Zombie hosts have the same IP for days. You assume responsible ISPs, which are _not_ the problem. The same ISPs which do nothing about zombies spewing stuff are the ones with the problem.
How do you propose to deal with mailing lists? I am on quite a few lists, where I read the list messages once a month (or even once in 6 months), and only glance over the list otherwise to ensure that I am not missing important stuff. Keep in mind that when I do read the list, I read a few thousand messages at a single shot.
These lists will simply become unusable, or unrunnable.
Oh, and _how_ does your solution stop spam finally?
Now you have to be online to retrieve mail and connect to lots of different servers, and then download messages. That is the equivalent of POP3, only from a few thousand servers instead of the one (zero for me, actually). More of my time _and_ bandwidth is being spent in trying to deal with mail, rather than less.
Whereas with my current server, the mail lands up directly in my inbox, all sorted and filtered for my reading pleasure.
No. You only need to fail to send the food to the people who need it to precipitate a hunger crisis.
Most modern famines are not food _production_ issues, they are food distribution issues. Even if the food is free, the shipping is too expensive.
You are assuming that the certifying authorities will revoke the certificate, and people will get to know about the revocation fast enough to prevent the mail from being accepted in the first place.
We call it usenet.
Oh, and you had better be willing to store messages for months (my priority queue has unread messages that old). And then you have the issue of users who are not always online.
And also consider that if a message is sent to five recipients at a single domain, only one copy of that message is sent. With your proposal, 5 downloads happen. This can be a significant issue for mailing lists.
To overcome greylisting, a spammer must be willing to implement a full mail-server on thier end.
No. The spammer just needs to resend the same message after $time when a majority of greylisting periods expire. Keep in mind that the spammer is using zombies, so it doesn't pay any real money.
idiotic sys admins who firewall out entire IP blocks for, seemingly, no reason.
Or they just decided that they didn't do any significant business with the owners of thatIP block which warranted lifting the block. Also keep in mind ISPs like UUNet/MCI (now Verizon) which used ordinary users as shields while allowing spammers to spam unhindered that network.
The only thing that worked was wholesale blocking of mail from all of UUnet until their users left (or complained enough that the cost of keeping the spammer was more than the cost of handling complaints). SPEWS worked, for a _very_ good reason. If ISPs don't clean up, we will be glad to move back to using SPEWS to block spam.
You support a spamming ISP, you get blocked. If you don't like it, vote with your money.
Why would a combination of the small changes not produce sexual incompatibility? One or two changes would lead to sterility (such as mules), and from there on, the differences just increase.
It currently just works. The whole replication set issue is pretty much moot, because you can build another one trivially.
And DDL _is_ rare for well designed database applications.
Actually, it takes all of 10 minutes to get replication working reliably.
Try this: link for more information, howtos and scripts to assist you in setting up replication.
Except that the closest competitor to Oracle is PostgreSQL, and _that_ one is slightly harder to buy out.
MySQL just had the hype.
Nobody wants to hire someone who is going to challenge them.
Some of us would love to be able to hire someone better than us, or who would challenge us. If only because it makes the work more enjoyable. Which is why we have three admin type positions empty for 6 months.
Uhm, over-rated? INR is Indian rupees. That was a mild joke on outsourcing.
And Google cares about Fortune 50 company users accessing them from work because? Unless they make for a significant proportion of ad revenue, Google really doesn't care.
Your value to Google is the number of eyeballs you can offer them, or the advertising revenue they make from you. Do Fortune 50 corporations offer enough eyeballs to be a globally significant number?
Chanserv sets mode +o dodobh
dodobh sets mode +m
dodobh sets mode -v pclminion
You were saying something about tones?
Spam is NOT about freedom of speech. Spam is about Unsolicited Bulk Messaging. No one is taking away your right to freedom of speech. You are free to stand on a soapbox and make a speech, setup a website, print posters and flyers and distribute them... On the other hand, my mailbox is _my_ property. Your freedom of speech does not exist there.
and you can call blogs a "LiveJournal",
:P
Or Slashdot
Well, all that you need is a fission device _just_ below the threshold of explosion. The additional neutrons from the fusion reactor would then trigger it off.