Really, all you need to do is manage your address properly from the beginning, don't do obvious spam-lure tactics with it, use sneakemail/other aliasing and you're set.
Ah... I remember the days when I tried to avoid spam.... Now I collect it!
It's relatively simple economics. Microsoft is a monopoly.
If it weren't, to maximise its profit, Microsoft should keep selling to the point where the marginal cost (cost to produce 1 more copy of Windows) is equal to the marginal revenue (what that 152,435,126th buyer is willing to pay). This would likely mean that the price of Windows would be much closer to $40 rather than $200. (The Classic Supply and Demand model)
However, since Microsoft holds essentially a monopoly (because of barriers to entry in the OS market--namely, compatibility and familiarity), Microsoft will produce to the point again where marginal revenue meets marginal cost. However, since to sell an additional unit, the price for all units would have to be reduced (unless Microsoft used price discrimination...) MR < P. Therefore, this point would result in less quantity than expected, and Microsoft would sell at the price demanded, much higher than under competition.
This is all a little easier to explain with diagrams... but anyways I hope I've explained it. To lower the price of windows, all that needs to be done is to start selling more and more of linux (or giving it away, either is good). This will break up Microsoft's monopoly, and they will have to sell at the most efficient price, rather than the price that the market will bear. As long as Microsoft can sell at such a high price, it will.
I am somewhat shocked as a Debian Developer, and a new SPI member that I hadn't heard of this earlier. I really think Joey ought to have made this a little more public than he did. A posting to debian-devel-announce@lists.debian.org or even debian-devel@lists.debian.org would have been nice. Saves people like me from getting the news from/. -- not that that's a bad thing, but it's strange that/. and NewsForge managed to scoop a story before it was posted to a Debian mailing list. (Unless I missed something obvious...)
Joey did send a mail, forwarded by Bdale Garbee to debian-devel-announce@lists.debian.org that essentially stated:
Some members of the current Board of Directors of Software in the Public Interest would like to expand its Board.
In that e-mail he solicited nominations and suggested that candidates must have time to devote to the board.
It is true, without a doubt that SPI, and its Board of Directors, have not done anything in the last year that makes me consider SPI to be anything other than a formality designed to collect money for Debian and other worthy causes. I think the SPI website explains it best:
News
No items for this year.
So, I applaud Joey's decision. Maybe someone will wake up and realise, SPI has a role to fulfill and it's not doing it right now.
And maybe Joey will have even more time to write the Debian Weekly News, since nobody else is willing to do it:-)
Net Nanny might be good software for uptight parents
No blocking proxy server is good software for uptight parents. If parents really don't want their kids to find out what the internet is about, they should simply watch what the kid is doing from time to time or even surf the net with their kids.
And really, if little Billy wants to get porn, he'll get porn in some way by going around the proxy server. And, more likely, Billy knows more about computers than his parents, and will be able to disable it.
Does anyone else find this just a little strange? I doubt anyone achieved anywhere near 2.5MBps, and even if they had, I don't think $11,000 is the price to pay for it! And really, 16 hours times 2.5 MBps, thats... 144 GB. What's he transferring anyways? No home user can use that much bandwidth.
This guy got screwed by a litigation-happy company. I hope he wins.
I am one of the developers of SpamAssassin and I'm going to agree; non-spam is far harder to collect and it is needed in just as high a quantity as spam.
The biggest problem with non-spam is that it's private and often sensitive. It would be impossible to collect a giant corpus of non-spam representative of the business world that could be used to tweak spam filters.
For SpamAssassin, we get users/developers to submit the results (tests hit) for each message when run through spamassassin, and plug spam and non-spam results into some sort of a Genetic Algorithm. This way, users only need to submit results of rules, not full messages to us for scoring. For the most recent score set, we had 169k non-spam messages and 29k spam messages. (The scores are very good!)
For testing individual rules, we have a similar mechanism in place, with a smaller volume of results.
I'd say the best testing you can do involves the user with the mail running the test, and sending you the results, rather than sending you the mail.
One problem with public corpuses is that they tend to get dated, and generally aren't representative of the messages you want to filter. Filters based on a Bayesian type mechanism will find this sort of an archive entirely useless, and there are clearly better methods for rules-based filters.
PayPal is unprofessional, and definitely unacceptable for such a large payment (One-Time Equipment Fee: $700) to a professional company.
$700?!? for what? It's not even theft protection (a thief can break the antenna). Really, a cell phone (or even a sattelite phone) provides essentially all the same functionality (and almost certainly for less money).
I'd say they go bankrupt within 3 years, making that lovely equipment you bought useless to you.
Slashdot Mirror
No
Really, all you need to do is manage your address properly from the beginning, don't do obvious spam-lure tactics with it, use sneakemail/other aliasing and you're set.
Ah... I remember the days when I tried to avoid spam.... Now I collect it!
SpamAssassin
It's relatively simple economics. Microsoft is a monopoly.
If it weren't, to maximise its profit, Microsoft should keep selling to the point where the marginal cost (cost to produce 1 more copy of Windows) is equal to the marginal revenue (what that 152,435,126th buyer is willing to pay). This would likely mean that the price of Windows would be much closer to $40 rather than $200. (The Classic Supply and Demand model)
However, since Microsoft holds essentially a monopoly (because of barriers to entry in the OS market--namely, compatibility and familiarity), Microsoft will produce to the point again where marginal revenue meets marginal cost. However, since to sell an additional unit, the price for all units would have to be reduced (unless Microsoft used price discrimination...) MR < P. Therefore, this point would result in less quantity than expected, and Microsoft would sell at the price demanded, much higher than under competition.
This is all a little easier to explain with diagrams... but anyways I hope I've explained it. To lower the price of windows, all that needs to be done is to start selling more and more of linux (or giving it away, either is good). This will break up Microsoft's monopoly, and they will have to sell at the most efficient price, rather than the price that the market will bear. As long as Microsoft can sell at such a high price, it will.
Joey did send a mail, forwarded by Bdale Garbee to debian-devel-announce@lists.debian.org that essentially stated:
Some members of the current Board of Directors of Software in the Public Interest would like to expand its Board.
In that e-mail he solicited nominations and suggested that candidates must have time to devote to the board.
It is true, without a doubt that SPI, and its Board of Directors, have not done anything in the last year that makes me consider SPI to be anything other than a formality designed to collect money for Debian and other worthy causes. I think the SPI website explains it best:
News
No items for this year.
So, I applaud Joey's decision. Maybe someone will wake up and realise, SPI has a role to fulfill and it's not doing it right now.
And maybe Joey will have even more time to write the Debian Weekly News, since nobody else is willing to do it
Try going shirtless for a while!
No blocking proxy server is good software for uptight parents. If parents really don't want their kids to find out what the internet is about, they should simply watch what the kid is doing from time to time or even surf the net with their kids.
And really, if little Billy wants to get porn, he'll get porn in some way by going around the proxy server. And, more likely, Billy knows more about computers than his parents, and will be able to disable it.
I don't know if I could last a full 16 hours looking at porn... :-)
Do you have a 150 GB hard disk at home?
Perhaps a public corpus can compare diffent methodologies, however, there are better ways of harvesting one.
Regardless, these corpora will never be accurate representations of normal mail.
And mutt is also more free (i.e. free in the DFSG sense), allowing you to distribute modified versions, etc.
Does anyone else find this just a little strange? I doubt anyone achieved anywhere near 2.5MBps, and even if they had, I don't think $11,000 is the price to pay for it! And really, 16 hours times 2.5 MBps, thats... 144 GB. What's he transferring anyways? No home user can use that much bandwidth.
This guy got screwed by a litigation-happy company. I hope he wins.
I am one of the developers of SpamAssassin and I'm going to agree; non-spam is far harder to collect and it is needed in just as high a quantity as spam.
The biggest problem with non-spam is that it's private and often sensitive. It would be impossible to collect a giant corpus of non-spam representative of the business world that could be used to tweak spam filters.
For SpamAssassin, we get users/developers to submit the results (tests hit) for each message when run through spamassassin, and plug spam and non-spam results into some sort of a Genetic Algorithm. This way, users only need to submit results of rules, not full messages to us for scoring. For the most recent score set, we had 169k non-spam messages and 29k spam messages. (The scores are very good!)
For testing individual rules, we have a similar mechanism in place, with a smaller volume of results.
I'd say the best testing you can do involves the user with the mail running the test, and sending you the results, rather than sending you the mail.
One problem with public corpuses is that they tend to get dated, and generally aren't representative of the messages you want to filter. Filters based on a Bayesian type mechanism will find this sort of an archive entirely useless, and there are clearly better methods for rules-based filters.
No kidding!
PayPal is unprofessional, and definitely unacceptable for such a large payment (One-Time Equipment Fee: $700) to a professional company.
$700?!? for what? It's not even theft protection (a thief can break the antenna). Really, a cell phone (or even a sattelite phone) provides essentially all the same functionality (and almost certainly for less money).
I'd say they go bankrupt within 3 years, making that lovely equipment you bought useless to you.
If I hear another bad joke about "flame wars" or "firewalls" in reference to this, I think I might puke. :-)