The main reason for choosing Red Hat as a distribution is usually the "security and hardware certificatations".
Oracle should either find a way of provinding that or otherwise use some other distribution.
Debian would certainly profit very much if chosen for this;-)
I can't think of such an example. But I'm sure there are a few. And from these few even fewer are "honest/correct", mostly it's just a way of getting money from something you didn't do.
Imagine if Donnald Knut had pattented KMP pattern search algorithm. He did invent it didn't he? But he did NOT patent it, and because of that it has been soo helpful.
I do NOT think that patents should be estinguished, rather the system should be refactored.
The problem is not judging if pattents are good or bad. The issue IMO is that a bad pattent is easily issued, and that can be very bad for software developers.
Patents (quoting RMS) "are good only for those who have many of them and can cross license."
IBM for instance has over 3000 pattents. The 1st income they have with those pattents is, against what most people assume, CROSS LICENSING.
That is a very good example that shows us that pattents are no benefit for developers, only for big companies.
A week ago I attended a speech by RMS here in Brazil. The theme was "Software Patents". He gave us some nice examples of why software patents are so bad. I shall name some:
- SpreadSheet calculation using topological sorting is patented, eventhough it was invented 40 year ago.
- An MIT professor managed to pattent an Electricity Law (ridiculous).
- Software are very complex and pattents were not built for that complexity.
That seems like a good practice. I'm gonna adopt it. The only thing is that rejecting an e-mail that is not an existing user should (acording to RFC) generate a bounce.
On our e-mail ISP we are running a bayesian spam filter engine. Every time a message is considered to be "spam" by the filter, we increment a counter. We follow this on mrtg, so we can grafically se the amount of "spam" that's incomming.
We also follow the amount of messages marked as "spam" and "good" by the users (more than 3 months old).
The number we get, is the one mentioned on the topic. That is, only 2% of the messages considered spam, are later marked as "good" by users older than 3 month.
"You referred to "their DSL" - are you the service provider to the DSL-based spammer or not?"
Nope.
"how can a "triple-bounce" happen in a properly configured mailserver?"
MAIL FROM: inexistent-1@domain RCPT TO: inexistent-2@domain
will make qmail say
triple bounce: discarding...
it's properly configured, but still consumes CPU time..
Basically, the thing is, people blacklist domains when they should blacklist e-mails, or blacklist ips. Blocking an entire domain is very, very bad practice. What happens is that if you block an entire domain, the spammer will use another one, and that other domain will be blocked to, then you'll block it, and so on.
Our network that you mentioned as blocked, is not "our" network. They blocked.com.br which is ridiculous. Seriously, they are blocking any e-mail comming from brazil. Do you really think that:
- @oracle.com.br
- @sun.com.br
- @amcham.com.br (american chamber of commerce)
are sources of spam?
I think you are mistaken about how e-mail works. The bounce goes to the "MAIL FROM:" part of the SMTP connection.
We are a brazilian ISP (targeted on providing the best IMAP e-mail there is). We work on two fronts, corporate and non-corporate. On the corporate front, we host company's MX.
Our core business is all open source (qmail, courier, squirrel, etc). We provide a "dropin" solution for Exchange servers. The only problem is we don't connect with outlook:-(
We have been searching for comercial/non-comercial solution for this for quite a long time. There are some solutions, but none of them are IMO good.
I am sure our company would be willing to colaborate on efforts to build such a tool. I've seen some comments of people willing to "sponsor" something like that.
Why don't we try to organize something?
Jakarta has too many projects that do the same!
on
Struts 1.1 Released
·
· Score: 2, Insightful
A few months ago, Jakarta started to host several projects that do the same thing (torque ojb, ant maven, etc). The same happens with strutus turbine.
I have written some applications using turbine, and like the framework. The thing is, if "now finally all those IDE vendors are going to put real struts 1.1 support in their software" they will probably not do this for tubine. Should I give up on turbine?
These spammer are using the spam program either in their own DSL connections or using open relays. Let me give you an example:
1 - spammer sends e-mail from IP a.b.c.d with the spam program and signs the e-mail as @mandic.com.br ( is changed by the spam program for every email)
Note that the IP a.b.c.d (DSL) is blocked on my MX cluster. 2 - hotmail, for instance (who I cannot block) receives this e-mail and bounces it to? (me:-))
You go to a news store here in brazil, buy a spam program, install it, write a spam message and click on run. It sends e-mail to (as they announce) 16.000.000 e-mail signing them all as @mandic.com.br. Why is that so? Because we ARE A KNOWN ENTITY and therefore any RESPECTFUL e-mail provider will NEVER black list our domain. So the spammer uses the fact that we are a known entity in his benefit, and since the 16.000.000 are obviously not real, a great part of it will bounce and get to my postmaster.
I didn't get it, what's dodgy about my IP? What do you consider as "my IP"? Our cluster "lives" in a well known data-center, and there is nothing dodgy about it.
Now, about the "real postmasters". When there are 100.000 bounced messages on the queue and they are starting to triple bounce, what am I supposed to do? If I stop the mail daemons and run a program to remove the messages, it takes more than an our. How am I supposed to read them?
>> If you can't handle the Email, you need to close up shop and get a new domain name and IP address.
sorry, no can do! Our name is very strong and... we don't want to loose this.
But, anyways, tks for your "tip".
The only problem is most provider don't use GOOD blacklists. I had a problem with 2m (ie: mmm.com and 3m.com) for instance, I called their support and had to listen to the "technical contact" yelling at me because they were receiving spam with recipients from my domains. Same happend with lycos.com, footlocker.com and many other medium to big providers.
I work on an ISP that hosts a very old domain (mandic.com.br). Several Brazilian spam programs send spam with our domains as a recipient. They do not use my servers, they are not my users, but still we get black-listed.
What's the point on securing my users then?
I run an e-mail server with over 20.000 acounts.
This is what happens (and I am not RFC ignorant):
My domain (ie: mandic.com.br) is about 10 years old. So it was present on the first spam lists that ever existed. People use it to send spam. That is, they send spam and sign it as foo@mandic.com.br. That happens about every day. My postmaster@mandic.com.br receives about 40MB e-mail every day. I would need 2 persons reading this to get it read. What do I do?
Dude, BraSil is comming! In a few months these guys will know how to spell anything related to "free software" and "braSil" just like they know how to spell "microsoft"
Please note that the language spoken in Brazil is portuguese, the brazilian word for "free" is liVre. What you wrote on your article (ie: libre) is spanish.
Slashdot Mirror
How long did it take until engineering mgmt was not a black art?
The main reason for choosing Red Hat as a distribution is usually the "security and hardware certificatations". Oracle should either find a way of provinding that or otherwise use some other distribution. Debian would certainly profit very much if chosen for this ;-)
+1 and it works perfectly
Thawte offers free e-mail certificates.
- Can't those be used?
- Isn't that a good enough price?
I can't think of such an example. But I'm sure there are a few. And from these few even fewer are "honest/correct", mostly it's just a way of getting money from something you didn't do.
Imagine if Donnald Knut had pattented KMP pattern search algorithm. He did invent it didn't he? But he did NOT patent it, and because of that it has been soo helpful.
I do NOT think that patents should be estinguished, rather the system should be refactored.
The problem is not judging if pattents are good or bad. The issue IMO is that a bad pattent is easily issued, and that can be very bad for software developers.
Patents (quoting RMS) "are good only for those who have many of them and can cross license."
IBM for instance has over 3000 pattents. The 1st income they have with those pattents is, against what most people assume, CROSS LICENSING.
That is a very good example that shows us that pattents are no benefit for developers, only for big companies.
A week ago I attended a speech by RMS here in Brazil. The theme was "Software Patents". He gave us some nice examples of why software patents are so bad. I shall name some:
- SpreadSheet calculation using topological sorting is patented, eventhough it was invented 40 year ago.
- An MIT professor managed to pattent an Electricity Law (ridiculous).
- Software are very complex and pattents were not built for that complexity.
he he, forgot the most important thing.
it's bogofilter
That seems like a good practice. I'm gonna adopt it. The only thing is that rejecting an e-mail that is not an existing user should (acording to RFC) generate a bounce.
Anyways, I applied that "patch". TKS a LOT!
On our e-mail ISP we are running a bayesian spam filter engine. Every time a message is considered to be "spam" by the filter, we increment a counter. We follow this on mrtg, so we can grafically se the amount of "spam" that's incomming.
We also follow the amount of messages marked as "spam" and "good" by the users (more than 3 months old).
The number we get, is the one mentioned on the topic. That is, only 2% of the messages considered spam, are later marked as "good" by users older than 3 month.
"You referred to "their DSL" - are you the service provider to the DSL-based spammer or not?"
...
Nope.
"how can a "triple-bounce" happen in a properly configured mailserver?"
MAIL FROM: inexistent-1@domain
RCPT TO: inexistent-2@domain
will make qmail say
triple bounce: discarding
it's properly configured, but still consumes CPU time..
Basically, the thing is, people blacklist domains when they should blacklist e-mails, or blacklist ips. Blocking an entire domain is very, very bad practice. What happens is that if you block an entire domain, the spammer will use another one, and that other domain will be blocked to, then you'll block it, and so on.
Our network that you mentioned as blocked, is not "our" network. They blocked .com.br which is ridiculous. Seriously, they are blocking any e-mail comming from brazil. Do you really think that:
- @oracle.com.br
- @sun.com.br
- @amcham.com.br (american chamber of commerce)
are sources of spam?
I think you are mistaken about how e-mail works. The bounce goes to the "MAIL FROM:" part of the SMTP connection.
Anyways, thanks for the advice.
We are a brazilian ISP (targeted on providing the best IMAP e-mail there is). We work on two fronts, corporate and non-corporate. On the corporate front, we host company's MX.
:-(
Our core business is all open source (qmail, courier, squirrel, etc). We provide a "dropin" solution for Exchange servers. The only problem is we don't connect with outlook
We have been searching for comercial/non-comercial solution for this for quite a long time. There are some solutions, but none of them are IMO good.
I am sure our company would be willing to colaborate on efforts to build such a tool. I've seen some comments of people willing to "sponsor" something like that.
Why don't we try to organize something?
A few months ago, Jakarta started to host several projects that do the same thing
(torque ojb, ant maven, etc). The same happens with strutus turbine.
I have written some applications using turbine, and like the framework. The thing is, if "now finally all those IDE vendors are going to put real struts 1.1 support in their software" they will probably not do this for tubine. Should I give up on turbine?
These spammer are using the spam program either in their own DSL connections or using open relays. Let me give you an example:
:-))
1 - spammer sends e-mail from IP a.b.c.d with the spam program and signs the e-mail as @mandic.com.br ( is changed by the spam program for every email)
Note that the IP a.b.c.d (DSL) is blocked on my MX cluster.
2 - hotmail, for instance (who I cannot block) receives this e-mail and bounces it to? (me
Is this clearer now?
You go to a news store here in brazil, buy a spam program, install it, write a spam message and click on run. It sends e-mail to (as they announce) 16.000.000 e-mail signing them all as @mandic.com.br. Why is that so? Because we ARE A KNOWN ENTITY and therefore any RESPECTFUL e-mail provider will NEVER black list our domain. So the spammer uses the fact that we are a known entity in his benefit, and since the 16.000.000 are obviously not real, a great part of it will bounce and get to my postmaster. I didn't get it, what's dodgy about my IP? What do you consider as "my IP"? Our cluster "lives" in a well known data-center, and there is nothing dodgy about it. Now, about the "real postmasters". When there are 100.000 bounced messages on the queue and they are starting to triple bounce, what am I supposed to do? If I stop the mail daemons and run a program to remove the messages, it takes more than an our. How am I supposed to read them? >> If you can't handle the Email, you need to close up shop and get a new domain name and IP address. sorry, no can do! Our name is very strong and ... we don't want to loose this.
But, anyways, tks for your "tip".
The only problem is most provider don't use GOOD blacklists. I had a problem with 2m (ie: mmm.com and 3m.com) for instance, I called their support and had to listen to the "technical contact" yelling at me because they were receiving spam with recipients from my domains. Same happend with lycos.com, footlocker.com and many other medium to big providers.
I work on an ISP that hosts a very old domain (mandic.com.br). Several Brazilian spam programs send spam with our domains as a recipient. They do not use my servers, they are not my users, but still we get black-listed. What's the point on securing my users then?
I run an e-mail server with over 20.000 acounts. This is what happens (and I am not RFC ignorant): My domain (ie: mandic.com.br) is about 10 years old. So it was present on the first spam lists that ever existed. People use it to send spam. That is, they send spam and sign it as foo@mandic.com.br. That happens about every day. My postmaster@mandic.com.br receives about 40MB e-mail every day. I would need 2 persons reading this to get it read. What do I do?
Dude, BraSil is comming! In a few months these guys will know how to spell anything related to "free software" and "braSil" just like they know how to spell "microsoft"
Please note that the language spoken in Brazil is portuguese, the brazilian word for "free" is liVre. What you wrote on your article (ie: libre) is spanish.