Bottom line: stay tuned to Ximian.com for the release of the Ximian Desktop 2.0, and when it comes out of beta, grab yourself a copy of Red Hat 9, and then install Ximian over it. I think you'll be pleased with the results....
I moved to a linux desktop with openoffice.org about a year ago, and I used Evolution in place of Outlook
About a year ago, things were in pretty good shape, but still not perfect. I find your conerns to be most heartening though, and while I'm sorry it was tough for you, I'm glad that you were having problems of the sort you were rather than the ones that you would have had 2 years ago.
I thought it was great - or least that's what I kept telling myself. What I really found was that it was great to feel like I was cool for using those things, but in all honesty, it sucked. Nothing worked well together
My mailer, Web browser, calendar, contact manager, task manager and command-line terminal all work well together with minor problems caused almost exclusively by the fact that I run a poorly integrated development-version desktop rather than a more stable, if older, version from a vendor who has the done the integration work (e.g. Ximian or Red Hat).
When it comes to office applications, I mostly don't bother. Folks who send me resumes in a proprietary format can suffer with what a conversion program does to them when trying to extract text. If they don't like that, they should offer a text version of their resume.
I couldn't copy text in one app and paste it into another
Pasted from terminal: 1526 Oct 30 12:39 x.html Pasted from mailer: To: ajs@ajs.com Pasted from calendar: Boston.pm
That's about all I really need....
Evolution didn't allow me to change my email font - it required changing the default HTML view of Gnome
Tools->Settings->Font Preferences
What's more, sadly, much of the world is still using MS Office. Sure, openoffice.org can open and even save in documents in office format, but they're not perfect.
Perfect is a relative term when you're talking about handling unspecified formats. Converting to text is about your best bet. Another approach would be to convert to a neutral XML format that lets anyone read it.
If it really bothers you, you should send people a reply and ask them to send you HTML or plain text, both of which are quite possible from Word.
The copyright extension case was very hard to call. It centers around a flaw in the constitution, and when the constitution has flaws the court is nominally supposed to side with the constitution over intent. Still, it was good to see the discent, and I would not be surprised if there were some "ok, someone needs to write the discent on this one, who's it going to be?" type of discussion.
It's one of those cases where Congress is really to blame for pushing the bounds of the constitution in a way that is clearly beyond the scope of the INTENT.
The DMCA on the other hand flagrently violates both established constitutional precident and the letter of the document.
It significantly weakens the public domain and the benefit of the arts to the common good. These are the foundation of copyright law as laid out in the constitution (not even the Bill of Rights, but the core constitution itself).
I cannot imagine this law holding up to USSC scrutiny, but we shall see....
I don't expect anyone short of the US Supreme Court to overturn this descision. There's no way to argue this case without turning to the constitutionality of the DMCA and its far-reaching privacy implications, and that's going to have to go to the USSC.
I'm actually looking forward to the DMCA's day in court, though I'm not sure this is the ideal case for it, since it combines too many gray areas of the DMCA with business-vs-privacy with the unconstitutional elements of the DMCA. Still, I expect the only possible outcomes would be a) it's not reviewed by the USSC or b) there's some setback for the DMCA here, large or small....
If it does go that far, the RIAA would be smart to simply drop the case.
If you haven't noticed DNS is centralized and is a major pain in the ass. Sure there may be millions of DNS servers, but they all depend on the 13 root servers. The RFC's don't prevent me from setting up my own root server and asking everyone to point to that instead, but efforts to do that have failed. Even when they mirror the root DNS in addition to their own domains, and hence depend on the 13 root
That's simply not DNS. You're talking about Internic's domain name infrastructure that your local implementation has some defaults for and that everyone relies on, but there are alternate roots, and a system such as this should probably introduce a few extra roots for just that purpose (e.g. roots that would defer any requests for anything other than ".idtrust" or the like TLD to the Internic roots).
Part of the installation would have to be installation of the additional root servers in your configs until name resolution software starts shipping with such add-ons.
DNS the *protocol* is already quite reasonably suited to this task.
The receiver of the e-mail sets the charge. That means if her e-mail is only known to her friends, she can set the fee low. If she needs to post it on her web-page and announce it on CNN, she can set the CPU cost so high that it takes two days to send her an e-mail.
You get exactly the same effect by requiring TLS with a particular key-size. You can also get this effect (though with some noise being generated) by giving a temporary failure code for all mail unless its message-ID (or perhaps crytographic checksum for security and to avoid simple-case replay attacks) matches in a database of previously-seen IDs, e.g. you bounce mail on the first try. 99.99% of spammers will never try to redeliver that message.
Both scenarios will result in refusing a large amount of valid mail while not establishing any sort of trust relationship, which I see as a flaw, but you can get what you want without code changes or with very minor changes. Your solution seems a bit too cumbersome.
Never said anything about revenue. What I said was that their LARGEST CUSTOMERS (e.g. the ones they do the most business with) are the US Federal Govt and bulk mailers. This has been true for at least 20 years or so since I heard the statistic quoted.
Every merchant knows that the bulk of their business may not be their biggest profit, but they still cater to whatever market segment that is in order to preserve the STABILITY of their business.
The point that I was making was that as the Internet becomes more a phenomenon of big business and less a phenomenon of startups, governments and accedemic institutions, you will see spam shift from a fly-by-night business to a "legitimate" enterprise based on strong partnerships with the infrastructure providers. I guarantee you that at some point you'll here some exec say, "using the Internet and not reading your spam is effectively stealing."
The problem is that the current means of combatting spam play into that scenario perfectly. As ISPs acquire more and more of the roll of "trusted intermediary", they get to dictate more and more of what the definition of spam is. The inevitable compromise is to get legislation passed that makes it illegal to contact their SMTP ports without a prior business arangement, then setting up an arrangement between a consortium of the largest ISPs. Governements, schools and businesses will hail this as the end of spam, and roll over. After all they get to send and receive mail to/from anyone who wants it (e.g. anyone willing to sign up with a provider who's a member of the consortium).
Anyone who sees AOL blocking residential users and sees a strong, healthy Internet coming out of it down the road is deluding themselves massively.
I look forward to the day that spammers and I don't have the same reputation on the Net because reputation is hard to build and easy to lose... I'm trying to build such a system, but first I'm finishing some rather large changes to SpamAssassin so that it's much faster. When that's done, I'll focus on MTAs.
Think about this the next time you advocate centralizing the Net (in terms of SMTP) on large ISPs in order to "solve" the problem of residental users spewing spam (directly or by relay).
The residential users are annoying because there are so many of them, but if, 10 years from now, the only way you can send mail is to relay through a large ISP's mail servers... who do you think said ISP's best business partners will be?
For an answer to that question look to the US Postal Service's largest customers: The US Federal Government and bulk mailers.
THAT is exactly the business niche that spammers are evolving into. All they need is for users to have slightly less choice and ISPs to have slightly more power to tell their users how the Internet works rather than the other way around.
Push to keep the Internet a network of peers while establishing a system of identity, trust and responsibility (which should in turn also by non-centralized, but rooted on an arbitrary number of certificate authorities and trust databases), and you will do yourself and the rest of the world a large favor!
Having a central authority for tying identity to e-mail not only concentrates power and points of failure, but also adds unneeded hasle and real dollar cost.
Ew! Why on Earth would you ever want to have a central authority?!
First, establish a series of certificate authorities (as many as you like). Now establish a protocol for verifying keys against those certifying authorities that is resliant and fast.
Luckily those two have been done.
Next, create a system of trust-association so that anyone who wants to voice an opinion about a particular identity can. This is relatively easy to do stand-alone, but DNS is also suited to it (look for the RFCs on key recovery via DNS).
Ultimately people will begin to standardize on a few trust-sources, but there's no reason that joe average can't fire up his own trust authority for his friends and familly to use.
Identity tracking != centralization!
As far as the idea for "charging someone" CPU time to deliver mail... spammers would love you for such an idea. Burning CPU on the relays doesn't hurt them at all, and even if they had to build out compute farms capable of delivering such mail, it's all worth it if your spam is profitable. At best, you would be culling the herd of spammers down to those who actually make money....
Identity is the way to go, not on the user level, but the server. I want my server to get some credit for having been a good Internet citizen for years (regardless of IP or other useless trivia) and the spammer's machine to have a bad rep that gets it filtered and/or ignored. That encourages me to work hard at maintaining my good citizen status and hurts the spammer at the same time.
That, as opposed to the "maximum collateral damage" approach that blocks access from huge chunks of the Internet containing valid servers along with spam sources.
Your other idea of delivering client mail through IMAP and then shutting off SMTP for all non-servers... um, what's a server?
Should my machines at home be able to connect to a target random system for SMTP transmission? If not, how do I send mail direct-to-MX for security and privacy reasons? How do I engage TLS if my provider doesn't support it? How, in short, do I partake of the network of peers that is the Internet? Or should we toss that out the window at the first opportunity because users don't like spam?
Odd that you advocate a decentralized Internet except when it comes to sending mail....
The obfuscated text image thing is long since broken. It's actually very easy with modern image manipulation and OCR techniques (take any random OCR software, add in a touch of image cleanup specialized to the output of your particular obfuscated images like the removal of the grid lines followed by some very liberal edge-detection, and go). However, it does do what it was intended to: prevent casual scripting of web UIs.
This is a consumer document meant to tell folks how to stop getting as much spam.
Useful insofar as it goes, but what would be much more helpful is an objective take on how spam gets to the end-system. It's very hard to generate this information. You can come up with the list of final-hop relays, but that's not as useful as you might think, since most of the really crappy spam software out there finds open relays dynamically and routes through them.
Slightly smarter software is now making it out there that performs some simple testing to determine how / if a given relay of choice can reach other sites. So for example, AOL's recent blocking of Commcast customers will help them in the short term, but over time they'll find that spammers simply stop using those relays and start using the ones that can get through. As new relays pop up, they will be used... eventually you would have to simply stop accepting mail in order to correctly prevent spam.
Like I say, it would have been useful to have the data on where spam is actually originating, but even without it, you can block spam with a very high degree of certainty based on the sender and relays with a much lower false positive (failure) rate than any of the bogus blacklist schemes out there. I'm about to add a module to SA to do just this, so stay tuned....
Reading the RFCs is a very good start to understanding how to solve this sort of problem. Giving everyone on the Internet (or at least all of the SMTP-sources) an Identity and then actually attaching a record of trust to those identities would be a wonderful idea, and does NOT require replacing SMTP. In fact, if you do it very, very carefully, it probably doesn't even require writing any (or at least very little) new code.
In that case, it would depend on who they are. If they're young and male and not prone to being shocked by sexuality (and you're not worried about the implications of recommending such a book), I'd say start them off with XXXenophile. The sex usally gets them interested and the humor gets them interested in more ("that glow tells you it's working!") I use humor to indtoduce folks to SF (Hitchhikers, Men Who Killed Mohamed, etc.) and Fantasy (Discworld, Myth, etc.) all the time.
If a sexually explicit book would not be appropriate (often the case, of course), then Bone is a very funny and very good series.
Perhaps they enjoy mythological deconstruction? Sandman is a great intro, but it can be hard to choose the right place in the stories. I recommend Fables and Reflections as a very good intro to the series without the clumsy tie-ins of Preludes Nocturnes or the heavy dose of mid-game plot in Season of Mists or World's End.
If they're not into that sort of thing, then perhaps Love and Rockets which tends to be more popular with women who would otherwise never read comics.
After that, you've narrowed your audience to the sort that I think *would* enjoy superheroes, but might not realize it. I very highly recommend the original Mage series, "Mage: The Hero Discovered". That's more of the updated classic fantasy sort of thing. There's also the best city-of-heroes book I've ever read: Astro City. Astro manages to actually be about everything in the city but the gobs of superheroes that litter its streets. Same goes for Top 10, though I only recommend that book to comics fans looking to do a little explorations WITHIN the genre.
I've been reading your columns and I'm pleased to see that there are those working in the comics industry who actually think about what they're creating and refuse to lend a helping hand to reversing the story-telling progress that was made in the 80s and 90s.
However, I have to disagree with you on The Authority. I don't think the sexuality of Superman and Batman... er, Apollo and Midnighter;) had anything to do with the decline of the book. Yes, the book then side-tracked too often toward that topic and yes, the original setup gave depth to the story without having to be a "gay superhero" thing.
But, do you think they could have been overtly gay and still managed to be "just another couple of members of the team"? To put that another way, if there had been a wealth of plot and character development available (as there was when you were writing it), can't you imagine simply dropping the answer to the question and moving on to other stories? Is it an inescapable trap or just an obvious one?
Ok, three question marks is too many in a Slashdot interview, even when they're really all the same question. So thank you again for great story telling, and good luck!
Re:how to get the norms
on
Ask Warren Ellis
·
· Score: 4, Insightful
And don't say Watchmen, cause that's (fantastic) genre crud.
"genre crud"? I can agree with "genre", since of course everything either *has* a pigeon-hole or *makes* one. "crud"? Hmmm.... nope, I don't think so.
So, what's "genre crud" to you? Did you dislike Watchmen because it had superheroes? Because it had... well, what it had at the end that I shouldn't spoil if folks haven't read it? When I hear "genre crud" it makes me think of something that sits comfortably within the lines defined by its genre, and Watchmen certainly did anything but! Granted, today it would be somewhat difficult to explain WHY that was the case, but that's because we have different expectations now.
My feeling is that Watchmen, The Dark Knight ("DK2".... *shudder*), Teenage Mutant Ninja Turtles, Sandman, Astro City, and many other titles between the mid 80s and the mid 90s helped comics readers to explore what it was that they wanted to get out of their superheroes, and each contributed to the genre significantly. Later works such as Top 10, The Authority, Planetary, Rising Stars and many others would never be mainstream (is Top Cow mainstream? Not sure) without the contributions of those books.
That's not to say there isn't "genre crud". I look at the recent Green Arrow series, and I see a few brilliant ideas up-front that Kevin Smith always brings to the table (though honestly the first few pages felt a bit like Dogma with superheroes) and then a few issues later... it starts to get bogged down in the need to introduce a villain and a "someone could die" moment right before the end of the issue.
Lucifer also started out with some interesting ideas and stalled. Granted, it made good reading for the first 15 or so issues, which is more than I can say for most sequels.
Then there's the flagship books. Every now and then I pick up a Superman or an X-Men, and I'm reminded that superhero story telling isn't always about telling a coherent story... Sometimes it's just about setting up a big fight, angsting over some "relationship issues" and beating the bad guy to a pulp while reciting a "truth and justice shall prevail" littany.
How such sorry, tired cruft could be compared to Watchmen, I'm seriously confused on.
"Remember that this is corporate America and the U.S. government with which we are dealing. The chance of their gathering data correctly, let alone devising a way to use it to their advantage, is remote."
Yes, that's quite true. Woefully, the chance of their gathering it incorrectly, taking no useful advantage of it and incidentally screwing over thousands of people's lives is pretty huge based on prior track records....
What people always forget is that most of the damage caused by large beauracracies is not caused by the focused, well-managed efforts of sinister authority figures. It's usually the broken bungling of incompetent peons who have been given a pointless role to serve and are terrified that someone will realize that fact.
However, the kernel headers do disagree with the rest of your statement. __PAGE_OFFSET determines the TASK_SIZE and that's 3GB by default on an x86. You'll find that even with CONFIG_HIMEM64GB, you are limited to a 3GB address space per process due to the fact that the x86 only addresses 4GB of the 64 at any given time.
According to page.h the kernel address space (which is always present, you don't page out the kernel address space) is the remaining 1GB of physmem.
If you search for "3GB linux address space x86" I think you'll find lots of discussion around this. Ours is not the only company that's been bitten by this particularly ugly problem that's symptomatic of the fact that the x86 simply wasn't designed for serious applications.
That's not really the concern. The real problem is that the x86's hardware page tables can only address 4GB of RAM at a time. We get smacked by this where I work quite a bit. This is why Linux for x86 has a 3GB address space per process (3 for the process, 1 for kernel-space). Go ahead, sit down at a box with 4GB of RAM and try to
malloc((2<<31)+(2<<30))
I have no clue if the Itanium addresses this. I would assume it does, since backward compatibility was only loosely maintained (mostly through emulation).
So, I would expect that an Itanium could address more than 4GB of RAM, but I have no data to suggest that. Someone want to chime in who has worked on one?
Have you reported them through Spamcop and/or sent mail to abuse@rackspace.com? You can't blame RS for failing to eliminate a spammer if you don't tell them. In most ISP's cases you can't even rely on telling them once, since that's when they send the warning.
The right thing to do is block the spammer, not Rackspace. Rackspace probably deserves an entry in an ISP-RBL somewhere, but most systems will weight that fairly low.
The really funny thing is that I got Slashdot's email-notification of your post, and there was a Rackspace ad in it;-)
I do this too, but the problem is that I only want my home system to become so much of a server. At some point, I would end up having to buy more hardware and get a decent pipe, which would be a major pain.
If I ever start thinking that way, I'll probably grab a server out at Rackspace and blow the $300/month or whatever it is. But, when/if I do that, I'll have to start charging my friends and that will require that I provide them with support...
There *are* good ISPs out there, but it takes time and effort to find them. Perhaps I should start doing that research. Sigh.
They are in a position to trump any "standards" because of their saturation.
No they're not. But if you think that way they are.
Imagine if you will that AOL had tens of thousands of support calls asking why friends and loved ones were getting bounces. I have a friend at work whose wife was asking why she couldn't send mail to their priest on AOL. Another friend can't get my mail, and I told him to just call AOL until they fix it. I've suggested to other friends and familiy that they switch.
My hunch is that most of the people that AOL is blocking this way are the technically savvy folks who their friends and familly go to for help. If we all start telling ten or so of our friends to call up AOL and ask why they can't get mail from us.... AOL's spam problems will begin to seem less important.
Understand this: I'm not suggesting spamming their phones. I'm not suggesting that anyone "get revenge". It's simply a matter that the service, as advertized, is broken. They don't actually accept mail from large chunks of the net, and that needs to get to all of their customers.
A status of 550 should only be sent in response to a command, not to connection.
Correct, and what's more they issue that 550 ending with "550 Goodbye" and then a connection reset (TCP-"R") packet, which is also in violation of the RFC.
If you run SpamAssassin, I highly recommend adding:
score RCVD_IN_RFCI 0 3 0 3
to your/etc/mail/spamassassin/local.cf. If everyone on the net does this, it won't block AOL's mail (or any other RFC-ignorant site), but it will mean that you have a much lower level of tollerance for spam-like mail from them.
It's not punative so much as showing them the right way to have solved this problem. Yes, AOL gets a lot of mail; yes, filtering spam out of it is hard; but if they simply weighted blacklists based on how accurate they are (as SA does) and then combined the results of several lists from dynips to rfci to relays with those weights, then they could make an accurate assessment, inform the sites that are blacklisted appropriately (in conformance with the RFC).
Ultimately, even after issuing that 554, if someone pushes on with a "RCPT To: postmaster@aol.com", they should accept it so that the site has a usable route for delivering mail to assert that the problem has been solved, but that would be a rare occurance if the lists were public and used/maintained correctly.
If you actually *break* the protocol on the otherhand, then things will probably get a little more ugly
Then it's time for it to get ugly. AOL breaks the protocol by issuing at 550 (not a 554) and not leaving the session open until timeout or client issues "QUIT" (you are allowd to say "553 Get bent" to every command issued, but you're not allowed to disconnect).
Let the blacklisting of AOL begin!
RFCs aside, though, they're blacklisting folks for getting an address assigned by a protocol. This is arbitrary and foolish. It also eliminates a lot of good mail.
I'll keep running my mail server, and AOL can keep ignoring me, but I'm going to start sending my friends and familly to AOL's competition, must as I hate to because that's mostly folks like MSN and the regional phone companies.
Slashdot Mirror
I moved to a linux desktop with openoffice.org about a year ago, and I used Evolution in place of Outlook
About a year ago, things were in pretty good shape, but still not perfect. I find your conerns to be most heartening though, and while I'm sorry it was tough for you, I'm glad that you were having problems of the sort you were rather than the ones that you would have had 2 years ago.
I thought it was great - or least that's what I kept telling myself. What I really found was that it was great to feel like I was cool for using those things, but in all honesty, it sucked. Nothing worked well together
My mailer, Web browser, calendar, contact manager, task manager and command-line terminal all work well together with minor problems caused almost exclusively by the fact that I run a poorly integrated development-version desktop rather than a more stable, if older, version from a vendor who has the done the integration work (e.g. Ximian or Red Hat).
When it comes to office applications, I mostly don't bother. Folks who send me resumes in a proprietary format can suffer with what a conversion program does to them when trying to extract text. If they don't like that, they should offer a text version of their resume.
I couldn't copy text in one app and paste it into another
Pasted from terminal: 1526 Oct 30 12:39 x.html
Pasted from mailer: To: ajs@ajs.com
Pasted from calendar: Boston.pm
That's about all I really need....
Evolution didn't allow me to change my email font - it required changing the default HTML view of Gnome What's more, sadly, much of the world is still using MS Office. Sure, openoffice.org can open and even save in documents in office format, but they're not perfect.
Perfect is a relative term when you're talking about handling unspecified formats. Converting to text is about your best bet. Another approach would be to convert to a neutral XML format that lets anyone read it.
If it really bothers you, you should send people a reply and ask them to send you HTML or plain text, both of which are quite possible from Word.
The copyright extension case was very hard to call. It centers around a flaw in the constitution, and when the constitution has flaws the court is nominally supposed to side with the constitution over intent. Still, it was good to see the discent, and I would not be surprised if there were some "ok, someone needs to write the discent on this one, who's it going to be?" type of discussion.
It's one of those cases where Congress is really to blame for pushing the bounds of the constitution in a way that is clearly beyond the scope of the INTENT.
The DMCA on the other hand flagrently violates both established constitutional precident and the letter of the document.
It significantly weakens the public domain and the benefit of the arts to the common good. These are the foundation of copyright law as laid out in the constitution (not even the Bill of Rights, but the core constitution itself).
I cannot imagine this law holding up to USSC scrutiny, but we shall see....
I don't expect anyone short of the US Supreme Court to overturn this descision. There's no way to argue this case without turning to the constitutionality of the DMCA and its far-reaching privacy implications, and that's going to have to go to the USSC.
I'm actually looking forward to the DMCA's day in court, though I'm not sure this is the ideal case for it, since it combines too many gray areas of the DMCA with business-vs-privacy with the unconstitutional elements of the DMCA. Still, I expect the only possible outcomes would be a) it's not reviewed by the USSC or b) there's some setback for the DMCA here, large or small....
If it does go that far, the RIAA would be smart to simply drop the case.
If you haven't noticed DNS is centralized and is a major pain in the ass. Sure there may be millions of DNS servers, but they all depend on the 13 root servers. The RFC's don't prevent me from setting up my own root server and asking everyone to point to that instead, but efforts to do that have failed. Even when they mirror the root DNS in addition to their own domains, and hence depend on the 13 root
That's simply not DNS. You're talking about Internic's domain name infrastructure that your local implementation has some defaults for and that everyone relies on, but there are alternate roots, and a system such as this should probably introduce a few extra roots for just that purpose (e.g. roots that would defer any requests for anything other than ".idtrust" or the like TLD to the Internic roots).
Part of the installation would have to be installation of the additional root servers in your configs until name resolution software starts shipping with such add-ons.
DNS the *protocol* is already quite reasonably suited to this task.
The receiver of the e-mail sets the charge. That means if her e-mail is only known to her friends, she can set the fee low. If she needs to post it on her web-page and announce it on CNN, she can set the CPU cost so high that it takes two days to send her an e-mail.
You get exactly the same effect by requiring TLS with a particular key-size. You can also get this effect (though with some noise being generated) by giving a temporary failure code for all mail unless its message-ID (or perhaps crytographic checksum for security and to avoid simple-case replay attacks) matches in a database of previously-seen IDs, e.g. you bounce mail on the first try. 99.99% of spammers will never try to redeliver that message.
Both scenarios will result in refusing a large amount of valid mail while not establishing any sort of trust relationship, which I see as a flaw, but you can get what you want without code changes or with very minor changes. Your solution seems a bit too cumbersome.
Never said anything about revenue. What I said was that their LARGEST CUSTOMERS (e.g. the ones they do the most business with) are the US Federal Govt and bulk mailers. This has been true for at least 20 years or so since I heard the statistic quoted.
Every merchant knows that the bulk of their business may not be their biggest profit, but they still cater to whatever market segment that is in order to preserve the STABILITY of their business.
The point that I was making was that as the Internet becomes more a phenomenon of big business and less a phenomenon of startups, governments and accedemic institutions, you will see spam shift from a fly-by-night business to a "legitimate" enterprise based on strong partnerships with the infrastructure providers. I guarantee you that at some point you'll here some exec say, "using the Internet and not reading your spam is effectively stealing."
The problem is that the current means of combatting spam play into that scenario perfectly. As ISPs acquire more and more of the roll of "trusted intermediary", they get to dictate more and more of what the definition of spam is. The inevitable compromise is to get legislation passed that makes it illegal to contact their SMTP ports without a prior business arangement, then setting up an arrangement between a consortium of the largest ISPs. Governements, schools and businesses will hail this as the end of spam, and roll over. After all they get to send and receive mail to/from anyone who wants it (e.g. anyone willing to sign up with a provider who's a member of the consortium).
Anyone who sees AOL blocking residential users and sees a strong, healthy Internet coming out of it down the road is deluding themselves massively.
All good points, and I agree.
I look forward to the day that spammers and I don't have the same reputation on the Net because reputation is hard to build and easy to lose... I'm trying to build such a system, but first I'm finishing some rather large changes to SpamAssassin so that it's much faster. When that's done, I'll focus on MTAs.
Think about this the next time you advocate centralizing the Net (in terms of SMTP) on large ISPs in order to "solve" the problem of residental users spewing spam (directly or by relay).
The residential users are annoying because there are so many of them, but if, 10 years from now, the only way you can send mail is to relay through a large ISP's mail servers... who do you think said ISP's best business partners will be?
For an answer to that question look to the US Postal Service's largest customers: The US Federal Government and bulk mailers.
THAT is exactly the business niche that spammers are evolving into. All they need is for users to have slightly less choice and ISPs to have slightly more power to tell their users how the Internet works rather than the other way around.
Push to keep the Internet a network of peers while establishing a system of identity, trust and responsibility (which should in turn also by non-centralized, but rooted on an arbitrary number of certificate authorities and trust databases), and you will do yourself and the rest of the world a large favor!
Having a central authority for tying identity to e-mail not only concentrates power and points of failure, but also adds unneeded hasle and real dollar cost.
Ew! Why on Earth would you ever want to have a central authority?!
First, establish a series of certificate authorities (as many as you like). Now establish a protocol for verifying keys against those certifying authorities that is resliant and fast.
Luckily those two have been done.
Next, create a system of trust-association so that anyone who wants to voice an opinion about a particular identity can. This is relatively easy to do stand-alone, but DNS is also suited to it (look for the RFCs on key recovery via DNS).
Ultimately people will begin to standardize on a few trust-sources, but there's no reason that joe average can't fire up his own trust authority for his friends and familly to use.
Identity tracking != centralization!
As far as the idea for "charging someone" CPU time to deliver mail... spammers would love you for such an idea. Burning CPU on the relays doesn't hurt them at all, and even if they had to build out compute farms capable of delivering such mail, it's all worth it if your spam is profitable. At best, you would be culling the herd of spammers down to those who actually make money....
Identity is the way to go, not on the user level, but the server. I want my server to get some credit for having been a good Internet citizen for years (regardless of IP or other useless trivia) and the spammer's machine to have a bad rep that gets it filtered and/or ignored. That encourages me to work hard at maintaining my good citizen status and hurts the spammer at the same time.
That, as opposed to the "maximum collateral damage" approach that blocks access from huge chunks of the Internet containing valid servers along with spam sources.
Your other idea of delivering client mail through IMAP and then shutting off SMTP for all non-servers... um, what's a server?
Should my machines at home be able to connect to a target random system for SMTP transmission? If not, how do I send mail direct-to-MX for security and privacy reasons? How do I engage TLS if my provider doesn't support it? How, in short, do I partake of the network of peers that is the Internet? Or should we toss that out the window at the first opportunity because users don't like spam?
Odd that you advocate a decentralized Internet except when it comes to sending mail....
The obfuscated text image thing is long since broken. It's actually very easy with modern image manipulation and OCR techniques (take any random OCR software, add in a touch of image cleanup specialized to the output of your particular obfuscated images like the removal of the grid lines followed by some very liberal edge-detection, and go). However, it does do what it was intended to: prevent casual scripting of web UIs.
This is a consumer document meant to tell folks how to stop getting as much spam.
Useful insofar as it goes, but what would be much more helpful is an objective take on how spam gets to the end-system. It's very hard to generate this information. You can come up with the list of final-hop relays, but that's not as useful as you might think, since most of the really crappy spam software out there finds open relays dynamically and routes through them.
Slightly smarter software is now making it out there that performs some simple testing to determine how / if a given relay of choice can reach other sites. So for example, AOL's recent blocking of Commcast customers will help them in the short term, but over time they'll find that spammers simply stop using those relays and start using the ones that can get through. As new relays pop up, they will be used... eventually you would have to simply stop accepting mail in order to correctly prevent spam.
Like I say, it would have been useful to have the data on where spam is actually originating, but even without it, you can block spam with a very high degree of certainty based on the sender and relays with a much lower false positive (failure) rate than any of the bogus blacklist schemes out there. I'm about to add a module to SA to do just this, so stay tuned....
You're missing something that just about everyone who talks about "the limitations of SMTP" misses: SMTP isn't limited. SMTP has a standard mechanism for introducing extensions such as cryptographically certifying mail servers, and mechanisms already exist to allow for fast, distributed key recovery and verification.
Reading the RFCs is a very good start to understanding how to solve this sort of problem. Giving everyone on the Internet (or at least all of the SMTP-sources) an Identity and then actually attaching a record of trust to those identities would be a wonderful idea, and does NOT require replacing SMTP. In fact, if you do it very, very carefully, it probably doesn't even require writing any (or at least very little) new code.
Ok, that makes sense. Thanks.
In that case, it would depend on who they are. If they're young and male and not prone to being shocked by sexuality (and you're not worried about the implications of recommending such a book), I'd say start them off with XXXenophile. The sex usally gets them interested and the humor gets them interested in more ("that glow tells you it's working!") I use humor to indtoduce folks to SF (Hitchhikers, Men Who Killed Mohamed, etc.) and Fantasy (Discworld, Myth, etc.) all the time.
If a sexually explicit book would not be appropriate (often the case, of course), then Bone is a very funny and very good series.
Perhaps they enjoy mythological deconstruction? Sandman is a great intro, but it can be hard to choose the right place in the stories. I recommend Fables and Reflections as a very good intro to the series without the clumsy tie-ins of Preludes Nocturnes or the heavy dose of mid-game plot in Season of Mists or World's End.
If they're not into that sort of thing, then perhaps Love and Rockets which tends to be more popular with women who would otherwise never read comics.
After that, you've narrowed your audience to the sort that I think *would* enjoy superheroes, but might not realize it. I very highly recommend the original Mage series, "Mage: The Hero Discovered". That's more of the updated classic fantasy sort of thing. There's also the best city-of-heroes book I've ever read: Astro City. Astro manages to actually be about everything in the city but the gobs of superheroes that litter its streets. Same goes for Top 10, though I only recommend that book to comics fans looking to do a little explorations WITHIN the genre.
Heehee! Yeah, it's just us guys. You can talk...... I'm sorry, do go on!
I've been reading your columns and I'm pleased to see that there are those working in the comics industry who actually think about what they're creating and refuse to lend a helping hand to reversing the story-telling progress that was made in the 80s and 90s.
;) had anything to do with the decline of the book. Yes, the book then side-tracked too often toward that topic and yes, the original setup gave depth to the story without having to be a "gay superhero" thing.
However, I have to disagree with you on The Authority. I don't think the sexuality of Superman and Batman... er, Apollo and Midnighter
But, do you think they could have been overtly gay and still managed to be "just another couple of members of the team"? To put that another way, if there had been a wealth of plot and character development available (as there was when you were writing it), can't you imagine simply dropping the answer to the question and moving on to other stories? Is it an inescapable trap or just an obvious one?
Ok, three question marks is too many in a Slashdot interview, even when they're really all the same question. So thank you again for great story telling, and good luck!
And don't say Watchmen, cause that's (fantastic) genre crud.
"genre crud"? I can agree with "genre", since of course everything either *has* a pigeon-hole or *makes* one. "crud"? Hmmm.... nope, I don't think so.
So, what's "genre crud" to you? Did you dislike Watchmen because it had superheroes? Because it had... well, what it had at the end that I shouldn't spoil if folks haven't read it? When I hear "genre crud" it makes me think of something that sits comfortably within the lines defined by its genre, and Watchmen certainly did anything but! Granted, today it would be somewhat difficult to explain WHY that was the case, but that's because we have different expectations now.
My feeling is that Watchmen, The Dark Knight ("DK2".... *shudder*), Teenage Mutant Ninja Turtles, Sandman, Astro City, and many other titles between the mid 80s and the mid 90s helped comics readers to explore what it was that they wanted to get out of their superheroes, and each contributed to the genre significantly. Later works such as Top 10, The Authority, Planetary, Rising Stars and many others would never be mainstream (is Top Cow mainstream? Not sure) without the contributions of those books.
That's not to say there isn't "genre crud". I look at the recent Green Arrow series, and I see a few brilliant ideas up-front that Kevin Smith always brings to the table (though honestly the first few pages felt a bit like Dogma with superheroes) and then a few issues later... it starts to get bogged down in the need to introduce a villain and a "someone could die" moment right before the end of the issue.
Lucifer also started out with some interesting ideas and stalled. Granted, it made good reading for the first 15 or so issues, which is more than I can say for most sequels.
Then there's the flagship books. Every now and then I pick up a Superman or an X-Men, and I'm reminded that superhero story telling isn't always about telling a coherent story... Sometimes it's just about setting up a big fight, angsting over some "relationship issues" and beating the bad guy to a pulp while reciting a "truth and justice shall prevail" littany.
How such sorry, tired cruft could be compared to Watchmen, I'm seriously confused on.
"Remember that this is corporate America and the U.S. government with which we are dealing. The chance of their gathering data correctly, let alone devising a way to use it to their advantage, is remote."
Yes, that's quite true. Woefully, the chance of their gathering it incorrectly, taking no useful advantage of it and incidentally screwing over thousands of people's lives is pretty huge based on prior track records....
What people always forget is that most of the damage caused by large beauracracies is not caused by the focused, well-managed efforts of sinister authority figures. It's usually the broken bungling of incompetent peons who have been given a pointless role to serve and are terrified that someone will realize that fact.
So, if I'm prohibited to use NAT-based firewalling, who's going to take responsibility for securing my home LAN?
Home LAN?! Was that expressly provided for in your agreement with your service provider?!
I think you're going to have to rip out those cables before the FBI gets to your house, buddy!
Yeah, I meant 1 not 2... stupid off-by-one.
However, the kernel headers do disagree with the rest of your statement. __PAGE_OFFSET determines the TASK_SIZE and that's 3GB by default on an x86. You'll find that even with CONFIG_HIMEM64GB, you are limited to a 3GB address space per process due to the fact that the x86 only addresses 4GB of the 64 at any given time.
According to page.h the kernel address space (which is always present, you don't page out the kernel address space) is the remaining 1GB of physmem.
If you search for "3GB linux address space x86" I think you'll find lots of discussion around this. Ours is not the only company that's been bitten by this particularly ugly problem that's symptomatic of the fact that the x86 simply wasn't designed for serious applications.
So, I would expect that an Itanium could address more than 4GB of RAM, but I have no data to suggest that. Someone want to chime in who has worked on one?
Can we please just start blocking all SMTP from 0/0? It would really make the point to all these damn spammers.
:-(
Grrrr... stupid, flailing at windmills considered harmful. Film at 11.
Have you reported them through Spamcop and/or sent mail to abuse@rackspace.com? You can't blame RS for failing to eliminate a spammer if you don't tell them. In most ISP's cases you can't even rely on telling them once, since that's when they send the warning.
;-)
The right thing to do is block the spammer, not Rackspace. Rackspace probably deserves an entry in an ISP-RBL somewhere, but most systems will weight that fairly low.
The really funny thing is that I got Slashdot's email-notification of your post, and there was a Rackspace ad in it
I do this too, but the problem is that I only want my home system to become so much of a server. At some point, I would end up having to buy more hardware and get a decent pipe, which would be a major pain.
If I ever start thinking that way, I'll probably grab a server out at Rackspace and blow the $300/month or whatever it is. But, when/if I do that, I'll have to start charging my friends and that will require that I provide them with support...
There *are* good ISPs out there, but it takes time and effort to find them. Perhaps I should start doing that research. Sigh.
They are in a position to trump any "standards" because of their saturation.
No they're not. But if you think that way they are.
Imagine if you will that AOL had tens of thousands of support calls asking why friends and loved ones were getting bounces. I have a friend at work whose wife was asking why she couldn't send mail to their priest on AOL. Another friend can't get my mail, and I told him to just call AOL until they fix it. I've suggested to other friends and familiy that they switch.
My hunch is that most of the people that AOL is blocking this way are the technically savvy folks who their friends and familly go to for help. If we all start telling ten or so of our friends to call up AOL and ask why they can't get mail from us.... AOL's spam problems will begin to seem less important.
Understand this: I'm not suggesting spamming their phones. I'm not suggesting that anyone "get revenge". It's simply a matter that the service, as advertized, is broken. They don't actually accept mail from large chunks of the net, and that needs to get to all of their customers.
The customers will decide....
Correct, and what's more they issue that 550 ending with "550 Goodbye" and then a connection reset (TCP-"R") packet, which is also in violation of the RFC.
If you run SpamAssassin, I highly recommend adding:to your
It's not punative so much as showing them the right way to have solved this problem. Yes, AOL gets a lot of mail; yes, filtering spam out of it is hard; but if they simply weighted blacklists based on how accurate they are (as SA does) and then combined the results of several lists from dynips to rfci to relays with those weights, then they could make an accurate assessment, inform the sites that are blacklisted appropriately (in conformance with the RFC).
Ultimately, even after issuing that 554, if someone pushes on with a "RCPT To: postmaster@aol.com", they should accept it so that the site has a usable route for delivering mail to assert that the problem has been solved, but that would be a rare occurance if the lists were public and used/maintained correctly.
Bah.
If you actually *break* the protocol on the otherhand, then things will probably get a little more ugly
Then it's time for it to get ugly. AOL breaks the protocol by issuing at 550 (not a 554) and not leaving the session open until timeout or client issues "QUIT" (you are allowd to say "553 Get bent" to every command issued, but you're not allowed to disconnect).
Let the blacklisting of AOL begin!
RFCs aside, though, they're blacklisting folks for getting an address assigned by a protocol. This is arbitrary and foolish. It also eliminates a lot of good mail.
I'll keep running my mail server, and AOL can keep ignoring me, but I'm going to start sending my friends and familly to AOL's competition, must as I hate to because that's mostly folks like MSN and the regional phone companies.