iPlanet had VERY little to do with the AOL you hate - it was Netscape people and Sun people selling Netscape and Sun software bundles.
Very LITTLE of AOL's infrastructure runs on Microsoft. The vast majority runs on either UNIX or on Tandem fault-tolerant minis. When I left last year, some folks were beginning to play with LINUX now that it was becoming more reliable. Only one thing I can think of runs on NT.
The only "AOL infrastructure" that relies on Microsoft is the word-processing infrastructure.
Very untrue. CompuServe was floundering when AOL bought them, and so instead of letting them go off on their own, their service was merged into the AOL infrastructure (via CS2000) and their developers have been pooled on many projects.
With Netscape, there's tight integration in some cases, where it makes sense (e-mail for NetCenter), and not others. And again, the development resources are often shared between groups when needed.
Sometimes the integration can be premature, at best. There were many articles in the press about trouble when AOL brought TW employees onto the AOL e-mail infrastructure, which just wasn't ready to support the type of groupware features TW was used to. I argued against forcing it down their throats, but the merger team had already decided what a Good Thing it was, and there was no fighting it. Long term, though, it'll be a big boon to the AOL back end, forcing some feature development. And I believe there are other such ways they've leveraged support staffs (staves?) and other infrastructure since I left.
In general, I think AOL's been fairly smart about when to integrate and when not to integrate.
When I was a hiring manager at AOL (which is based in the D.C. area), I was never impressed with candidates that were currently with government contractors and/or had little non-government experience - as esm wrote, those that had worked there "too long".
The reason was exactly BECAUSE of the discipline and formality of the government environment. Many of the programmers that excel there could never excel in an atmosphere of "Go add embedded image capability to e-mail and let QA know when it's ready to test." They needed formal design meetings and complete, pre-written UI specs to really do their best work; in the dot-com environment, those details are often decided, or at least coordinated, by the same programmer writing the code. And I was never sure how comfortable they'd be, coming to a place with Nerf-gun fights in the halls, where most of the best work happens AFTER 6pm. Also, at the time, UNIX and C/C++ experience was limited in the military; I'm sure that is changing today.
There are always exceptions, of course, and some of my best hires were ex-contractors. Some even brought their discipline to the team and helped DO more formal designs; they were guys who would think about every edge case. But I was generally more impressed with someone who worked on a 411 system or other high-rate transaction processing systems than with someone who'd done signal processing - systems engineering over computer science.
I suspect he meant their office in France, as opposed to their corporate office in Seoul, or their office here in the U.S., where the Internet is based... (I say, I say, that's sarcasm, boy)
There was a recent article in the NY Times about how Sony implicitly permits smuggling of their products in Pakistan, but even though I'm anti-smuggling, I wouldn't necessarily avoid all Sony products because of the actions of their South Asian division. It's a big world.
"St. George dedicated himself to the telecommunications industry over two decades ago working as an independent international technology consultant specializing in the delivery of bandwidth solutions to second and world countries throughout South America, South Pacific and Asia."
I thought a lot about stamped e-mail in a previous life as a mail systems developer. Our VP of development was really hot on the idea, since it would solve both the authentication problem and the no-incentive-for-targeting problem. You wouldn't even have to make it backwards-compatible; just create a new tier of "first-class" e-mail. Two big problems though:
1. Technical: It would be very, very expensive to process e-mail stamped with some form of digital cash. You're adding lots of crypto calculations, database lookups, and some sort of synchronization scheme that scales up to whole-Internet level. Large sites would likely have to have crypto plug-in hardware to do this at all efficiently.
2. Political: You'd have to get a significant number of ISPs on board, and these days most spam is NOT sent directly through the big ISP mail servers anyway.
It's a neat concept but there are too many problems. It ended up not being worth it.
I have the right to send anyone I want mail, to attempt to contact them at least once
No, actually you don't, at least not in bulk. Compuserve and AOL both settled that issue in the mid-90s in federal court - mail servers are property, and the property owner has the right to decide who may enter. If you enter without authority, you are trespassing. If they post a "sign" saying "no unsolicited bulk e-mail", and you violate it, you are trespassing.
I was involved in the first AOL spam lawsuits. Believe me.
In a legal sense, the definition will have to be much more specific and allow for that first attempt at contact before any law will be passed.
Think about it - "one bite at the apple" doesn't work. There are over 22 million small businesses in the US alone. If only 1% of them each tried to e-mail you only once over a one-year period, you'd get 600 messages a day.
You're wrong that most providers disagree with the "spam=UBE" definition; in fact, most Acceptable Use Policies do prohibit UBE. Trust us. We've been through this before. In 11 years at AOL I was on the winning end of quite a few anti-spam lawsuits in federal court. And there are several other Slashdotters here with even better credentials.
over 100 million users of AIM and you say, AOL won't be there during a holiday?
No; obviously SOMEONE is going to be there; there's a well-staffed 24x7 NOC. But we don't know what his reporting method was, other than that it was via e-mail. He presumably just e-mailed someone at AOL whose name he knew; that person could certainly be on vacation over the holidays.
Note that 5 work days would mean that the report would have been made around 2001-12-20.
And yet that isn't the case, according to the article - he notified "after Christmas" (even though he'd known for a few weeks) and announced on New Years. At best, that's three working days, even if you don't allow extra leniency for the holidays. So he didn't follow the guideline. It says Conover stated he wanted to release the exploit on 1/1 anyway because it was the anniversary of w00w00's previous announcement! Yes, I think that's irresponsible.
The BBC article mentions nothing about cross-pollination at all. A few links away (under Food: Under the Microscope) there is an article about cross-pollination in general that doesn't mention the terminator plant.
It's been too many years since I studied plant reproduction. Would the pollen necessarily carry the terminator gene? If so, would that make the pollen itself sterile? If so, would that mean that the pollen wouldn't successfully fertilize the crop, thus leaving it available for better pollen?
I'm no fan of either Monsanto or terminator genes, I just want to separate fact from FUD.
Re:Monsanto akin to evil corporations from the mov
on
Monsanto and PCBs
·
· Score: 1
with the full knowledge that these sterile seeds could spread and render entire regions infertile
Would you care to explain how sterile seeds can "spread"?
Not exactly. According to the Washington Post, "the group found the problem several weeks ago, but didn't contact AOL until after Christmas. The group didn't get any response from AOL through an e-mail during the holiday week...w00w00 set a New Year's deadline for sentimental reasons."
The potential problem is not so much with the company that bought the ad, but with all the parties that MADE the ad that may have explicitly retained Internet rebroadcast rights. Actors, producers, etc.
I used to live in the Washington area, which for some reason has a rich market for well-produced radio ads. I tried to get permission just to put a few on my personal web page, and each time was told that I couldn't, they didn't have the rights.
That was years ago, and I didn't pursue it too hard; I'm sure AdCritic managed to work something out legally. But it does take work.
Fascinating points! I never really read the book with a critical eye towards their methodology. What you're saying would certainly explain why Bob and Clippit are so universally reviled, despite the book's "predictions".
I'd love to see that paper if you get a chance to e-mail it to me.
Bob wasn't just a random experiment. Read _The Media Equation_, by Reeves and Nass. A very readable account of psychological experiments that show how people respond to computers, and to technology in general. Much of this research went into Bob, and later the Office Assistant. I'd love to know what went wrong.
Some examples: People watching a news program on a TV *set* labeled "news television" will rate that program as more informative and authoritative than those watching the same program on a TV labeled "general-purpose television". People using a computer program that praises another computer program will rate it smarter than a program that criticizes another program. Larger pictures will be better remembered and better liked than smaller pictures. People will rate a speaking tutorial program more honestly if the rating program uses a different voice!
Only popular domains that everyone KNOWS host millions of email accounts, get probed with a dictionary list.
Very much not true. There are only a handful of million-account domains. But "jim", "bob", etc. are valid e-mail addresses at nearly EVERY domain. Most of the dictionary attacks I've heard about happened at small business or personal domains.
Most email users do not pay extra for incoming emails, especially in the US.
You mean "at least" in the US. Most Internet access in other countries IS metered by the minute.
And you're forgetting about the ISP. Spam filtering saves recipient ISPs millions of dollars a year in hardware and network costs. Spam really is paid for by the recipient ISP, not the spammer.
Unfortunately, this turns out not to be the case in practice. Spam rarely includes real-world contact information these days. In fact, even the URLs are usually highly obfuscated through IE bugs. Once you find the real IP address, finding out who operates it is another bunch of legwork. And even then, tracing the money through off-shore accounts, foreign shell corporations, and the mob is a big job.
With the very low barrier to entry that spam presents, there's no way The Authorities could ever prosecute even a small percentage of spammers.
Legal solutions are important, because of the social mores and definitions they provide, and because they at least raise the spectre of punishment. But technology solutions must go hand-in-hand.
- When I worked on the AOL mail system, any time I met someone new - whether socially, in business, at the gas station, whatever - the first and only question they'd ask was how to stop the spam.
- During periods where the spamblocks are less effective (because the spammers are ahead of the game), spam is by far THE NUMBER ONE COMPLAINT to Steve Case's mailbox and to Customer Service.
And this is *after* scores of millions of spams have already been blocked each day.
The strong libertarian/individualist/techie pull of Slashdot notwithstanding, the average American e-mail user just doesn't want their spam.
I agree wtih others who said that ISPs should publicize the existence of their spamblocks, and it must be part of the Terms of Service. But to say that even if users agree to filtering, it should be illegal? I don't get it.
There are 22 million small businesses in the U.S. alone.
If one-tenth of one percent of them decided to send you one message this year, and if they coordinated to achieve load-balancing, you would still get over 200 pieces a DAY.
Slashdot Mirror
Part of the cachet of using a Red Hat distribution amongst the fringes of 'our little group' comes from its perceived independence
So what you're saying is that you hope AOL doesn't buy Red Hat cuz then it won't be cool?
What the heck are you talking about?!
iPlanet had VERY little to do with the AOL you hate - it was Netscape people and Sun people selling Netscape and Sun software bundles.
Very LITTLE of AOL's infrastructure runs on Microsoft. The vast majority runs on either UNIX or on Tandem fault-tolerant minis. When I left last year, some folks were beginning to play with LINUX now that it was becoming more reliable. Only one thing I can think of runs on NT.
The only "AOL infrastructure" that relies on Microsoft is the word-processing infrastructure.
Jay, the ex-Mail Guy
Very untrue. CompuServe was floundering when AOL bought them, and so instead of letting them go off on their own, their service was merged into the AOL infrastructure (via CS2000) and their developers have been pooled on many projects.
With Netscape, there's tight integration in some cases, where it makes sense (e-mail for NetCenter), and not others. And again, the development resources are often shared between groups when needed.
Sometimes the integration can be premature, at best. There were many articles in the press about trouble when AOL brought TW employees onto the AOL e-mail infrastructure, which just wasn't ready to support the type of groupware features TW was used to. I argued against forcing it down their throats, but the merger team had already decided what a Good Thing it was, and there was no fighting it. Long term, though, it'll be a big boon to the AOL back end, forcing some feature development. And I believe there are other such ways they've leveraged support staffs (staves?) and other infrastructure since I left.
In general, I think AOL's been fairly smart about when to integrate and when not to integrate.
Jay the ex-mail guy
When I was a hiring manager at AOL (which is based in the D.C. area), I was never impressed with candidates that were currently with government contractors and/or had little non-government experience - as esm wrote, those that had worked there "too long".
The reason was exactly BECAUSE of the discipline and formality of the government environment. Many of the programmers that excel there could never excel in an atmosphere of "Go add embedded image capability to e-mail and let QA know when it's ready to test." They needed formal design meetings and complete, pre-written UI specs to really do their best work; in the dot-com environment, those details are often decided, or at least coordinated, by the same programmer writing the code. And I was never sure how comfortable they'd be, coming to a place with Nerf-gun fights in the halls, where most of the best work happens AFTER 6pm. Also, at the time, UNIX and C/C++ experience was limited in the military; I'm sure that is changing today.
There are always exceptions, of course, and some of my best hires were ex-contractors. Some even brought their discipline to the team and helped DO more formal designs; they were guys who would think about every edge case. But I was generally more impressed with someone who worked on a 411 system or other high-rate transaction processing systems than with someone who'd done signal processing - systems engineering over computer science.
Jay the ex-Mail Guy
It's on-ly MOST-ly unbreakable.
If it was ALL unbreakable, there's not much you could do but check its pockets for loose change.
I suspect he meant their office in France, as opposed to their corporate office in Seoul, or their office here in the U.S., where the Internet is based... (I say, I say, that's sarcasm, boy)
There was a recent article in the NY Times about how Sony implicitly permits smuggling of their products in Pakistan, but even though I'm anti-smuggling, I wouldn't necessarily avoid all Sony products because of the actions of their South Asian division. It's a big world.
"St. George dedicated himself to the telecommunications industry over two decades ago working as an independent international technology consultant specializing in the delivery of bandwidth solutions to second and world countries throughout South America, South Pacific and Asia."
Translation: he was the Cable Guy.
I thought a lot about stamped e-mail in a previous life as a mail systems developer. Our VP of development was really hot on the idea, since it would solve both the authentication problem and the no-incentive-for-targeting problem. You wouldn't even have to make it backwards-compatible; just create a new tier of "first-class" e-mail. Two big problems though:
1. Technical: It would be very, very expensive to process e-mail stamped with some form of digital cash. You're adding lots of crypto calculations, database lookups, and some sort of synchronization scheme that scales up to whole-Internet level. Large sites would likely have to have crypto plug-in hardware to do this at all efficiently.
2. Political: You'd have to get a significant number of ISPs on board, and these days most spam is NOT sent directly through the big ISP mail servers anyway.
It's a neat concept but there are too many problems. It ended up not being worth it.
If a company posts their HR department's e-mail address on the web site, then obviously they ARE soliciting resumes. That wouldn't be UBE.
I have the right to send anyone I want mail, to attempt to contact them at least once
No, actually you don't, at least not in bulk. Compuserve and AOL both settled that issue in the mid-90s in federal court - mail servers are property, and the property owner has the right to decide who may enter. If you enter without authority, you are trespassing. If they post a "sign" saying "no unsolicited bulk e-mail", and you violate it, you are trespassing.
I was involved in the first AOL spam lawsuits. Believe me.
In a legal sense, the definition will have to be much more specific and allow for that first attempt at contact before any law will be passed.
Think about it - "one bite at the apple" doesn't work. There are over 22 million small businesses in the US alone. If only 1% of them each tried to e-mail you only once over a one-year period, you'd get 600 messages a day.
You're wrong that most providers disagree with the "spam=UBE" definition; in fact, most Acceptable Use Policies do prohibit UBE. Trust us. We've been through this before. In 11 years at AOL I was on the winning end of quite a few anti-spam lawsuits in federal court. And there are several other Slashdotters here with even better credentials.
over 100 million users of AIM and you say, AOL won't be there during a holiday?
No; obviously SOMEONE is going to be there; there's a well-staffed 24x7 NOC. But we don't know what his reporting method was, other than that it was via e-mail. He presumably just e-mailed someone at AOL whose name he knew; that person could certainly be on vacation over the holidays.
Note that 5 work days would mean that the report would have been made around 2001-12-20.
And yet that isn't the case, according to the article - he notified "after Christmas" (even though he'd known for a few weeks) and announced on New Years. At best, that's three working days, even if you don't allow extra leniency for the holidays. So he didn't follow the guideline. It says Conover stated he wanted to release the exploit on 1/1 anyway because it was the anniversary of w00w00's previous announcement! Yes, I think that's irresponsible.
Here's a copy of the AP article.
The BBC article mentions nothing about cross-pollination at all. A few links away (under Food: Under the Microscope) there is an article about cross-pollination in general that doesn't mention the terminator plant.
It's been too many years since I studied plant reproduction. Would the pollen necessarily carry the terminator gene? If so, would that make the pollen itself sterile? If so, would that mean that the pollen wouldn't successfully fertilize the crop, thus leaving it available for better pollen?
I'm no fan of either Monsanto or terminator genes, I just want to separate fact from FUD.
with the full knowledge that these sterile seeds could spread and render entire regions infertile
Would you care to explain how sterile seeds can "spread"?
They did wait; AOL ignored them.
Not exactly. According to the Washington Post, "the group found the problem several weeks ago, but didn't contact AOL until after Christmas. The group didn't get any response from AOL through an e-mail during the holiday week...w00w00 set a New Year's deadline for sentimental reasons."
Yeah, I'd say that's irresponsible.
The potential problem is not so much with the company that bought the ad, but with all the parties that MADE the ad that may have explicitly retained Internet rebroadcast rights. Actors, producers, etc.
I used to live in the Washington area, which for some reason has a rich market for well-produced radio ads. I tried to get permission just to put a few on my personal web page, and each time was told that I couldn't, they didn't have the rights.
That was years ago, and I didn't pursue it too hard; I'm sure AdCritic managed to work something out legally. But it does take work.
Nate,
Fascinating points! I never really read the book with a critical eye towards their methodology. What you're saying would certainly explain why Bob and Clippit are so universally reviled, despite the book's "predictions".
I'd love to see that paper if you get a chance to e-mail it to me.
Bob wasn't just a random experiment. Read _The Media Equation_, by Reeves and Nass. A very readable account of psychological experiments that show how people respond to computers, and to technology in general. Much of this research went into Bob, and later the Office Assistant. I'd love to know what went wrong.
Some examples: People watching a news program on a TV *set* labeled "news television" will rate that program as more informative and authoritative than those watching the same program on a TV labeled "general-purpose television". People using a computer program that praises another computer program will rate it smarter than a program that criticizes another program. Larger pictures will be better remembered and better liked than smaller pictures. People will rate a speaking tutorial program more honestly if the rating program uses a different voice!
Fascinating stuff.
I worked in AOL mail development from 1993 to 2001. I know firsthand.
Jay Levitt
Only popular domains that everyone KNOWS host millions of email accounts, get probed with a dictionary list.
Very much not true. There are only a handful of million-account domains. But "jim", "bob", etc. are valid e-mail addresses at nearly EVERY domain. Most of the dictionary attacks I've heard about happened at small business or personal domains.
Most email users do not pay extra for incoming emails, especially in the US.
You mean "at least" in the US. Most Internet access in other countries IS metered by the minute.
And you're forgetting about the ISP. Spam filtering saves recipient ISPs millions of dollars a year in hardware and network costs. Spam really is paid for by the recipient ISP, not the spammer.
And it's very easy to enforce
Unfortunately, this turns out not to be the case in practice. Spam rarely includes real-world contact information these days. In fact, even the URLs are usually highly obfuscated through IE bugs. Once you find the real IP address, finding out who operates it is another bunch of legwork. And even then, tracing the money through off-shore accounts, foreign shell corporations, and the mob is a big job.
With the very low barrier to entry that spam presents, there's no way The Authorities could ever prosecute even a small percentage of spammers.
Legal solutions are important, because of the social mores and definitions they provide, and because they at least raise the spectre of punishment. But technology solutions must go hand-in-hand.
Jay, the ex AOL mail guy
This has to be one of the most often grossly exagerrated problems anyone ever cites
30-50% of the messages going to the world's largest mail system are spam.
So obviously, "grossly" must mean less than 3x...
Jay, the ex AOL mail guy
I don't have numbers, but here's some data:
- When I worked on the AOL mail system, any time I met someone new - whether socially, in business, at the gas station, whatever - the first and only question they'd ask was how to stop the spam.
- During periods where the spamblocks are less effective (because the spammers are ahead of the game), spam is by far THE NUMBER ONE COMPLAINT to Steve Case's mailbox and to Customer Service.
And this is *after* scores of millions of spams have already been blocked each day.
The strong libertarian/individualist/techie pull of Slashdot notwithstanding, the average American e-mail user just doesn't want their spam.
I agree wtih others who said that ISPs should publicize the existence of their spamblocks, and it must be part of the Terms of Service. But to say that even if users agree to filtering, it should be illegal? I don't get it.
Jay, the ex AOL mail guy
There are 22 million small businesses in the U.S. alone.
If one-tenth of one percent of them decided to send you one message this year, and if they coordinated to achieve load-balancing, you would still get over 200 pieces a DAY.
Opt-out doesn't scale.