As another poster mentioned, Netscape Communicator (though not Mozilla) has been able to read your AOL mail via IMAP for lo these many years. Forgot about that.
I'd love to hear what "POP3 over IMAP" interface you say you saw. AOL's native mail system model is much closer to IMAP than POP, which is why we never developed a POP server. But when I was in charge of mail systems development, we did develop a fully functional IMAP interface, which was launched on CompuServe but never on AOL. It also provided the back end for many other services, including AOL Mail On The Web, since IMAP makes a nice, well-defined API. I'm glad marketing finally allowed it to be launched on AOL.
I'm still not clear how you knew it was him from the initial log - did your staff notice him coming in and leaving at the time the logs said, or was he just the most suspicious person there during that time? Surely the cafe wasn't otherwise empty.
That said, although you don't explicitly say so in the story, I'm assuming you were tracing only that particular MAC address, and so if it weren't him, your tcpdump wouldn't have shown anything at all. For some reason the story implied that you raced to the cafe and set up a trace on his particular Ethernet port, which is of course not how it's usually done unless you have a managed switch...
This is a story that starts with a sysadmin seeing a 419 scam, hearing that there was a black guy with a "suspicious" accent in his cafe, deciding that this must be our criminal, and deciding to read his e-mail to find out...
Yep, definitely - bouncing is a ridiculously expensive operation, especially when most of it can't be delivered and just sits there clogging up your queue. At AOL we had entire banks of separate machines just to handle the bounce queues - but of course moving mail from one machine to another is a load in itself.
I know we were always working on ways to push spamfighting back to the inbound relays, and so I imagine that what you'll see is SMTP rejects, not bounces.
Ha! I never cease to be amazed at what explanations floor-lore can come up with... reminds me of the Q-Link days when we used to tell customers to wrap their disk drives in tinfoil. (Which, to be fair, actually worked sometimes on Commodores, since they were usually hooked up to TVs which emitted huge amounts of radiation in close proximity to the poorly-shielded external floppy drive..)
I worked in development, not operations, so I don't remember specific episodes, but it sounds like Earthlink was getting caught up in AOL's spam filters. Here's a cite showing it happening at least once.
I'd love to hear who told you that one. There has never been anybody anywhere in the entire mail operations management chain, from Steve Case down to the guy sitting in the NOC at 3 a.m., who would randomly bounce mail to lower the load. That's quite an accusation.
AOL doesn't tell the senders or the intended recipients that it's dropping emails
This is actually starting to change; they are doing more bouncing and less blackholing these days. The spammers set up their own QA accounts anyway, so hiding the bounces doesn't really prevent the spammers from figuring out the filters.
AOL was known to block all mail from random domains to lower its server load when things got overloaded
You do realize that this is standard sendmail behavior, and not some sort of nefarious AOL plot to make your mail disappear? Refusing connections at a high load average just delays delivery of the mail until the next retry. It isn't related to the OP's mention of mail disappearing, which sounds like a spam-filtering problem.
That said, when I was at AOL, delivery times were usually in the seconds to minutes - I can't remember the last time we actually had a significant load average on the inbound relays.
alt.aol-sucks was not available to AOL users for quite a while.
Do you have a cite for that? I just did a bit of googling, and all I can find are posts that complained about the fact that AOL, by default, showed newsgroup titles from the config database instead of dotted names, so that alt.aol-sucks showed up as "Flames and complaints about AOL". Some folks assumed that AOL renamed the group for some nefarious reason and came up with the description themselves, and others said that AOL was censoring the group because they couldn't find it. IIRC, the option to switch the list display from descriptions to newsgroup names was either always present or was added fairly early on, making it a moot point.
I'm a pretty good Googler, I don't remember AOL ever not carrying the group, I've found several posts from that timeframe (not from AOL sycophants) explicitly stating that AOL *has* always carried the group, and I find it hard to believe that not one person complained to alt-aol-sucks about AOL not carrying that very newsgroup, but maybe I'm just not searching on the right phrases.
it is also worth noting that many web servers have multiple domains on one IP address using both virtual directories and virtual domains
I imagine that, since AOL also supplies DNS to the AOL client, they are blocking by name, not IP.
However, I wonder if blocking by IP would work too in this particular case. Spammers are selling something, and if you sell something it's almost always going to be with an SSL-encrypted link. SSL doesn't work with virtual domains, so either (a) all the spammers have non-virtual domains or (b) only the credit-card processing, which ends up being done by a few companies like CCBill, is SSL. If (a), blocking by IP would work too.
, AOL blocked newsgroups that were created to discuss (and flame of course) problems with AOL
Eh? Which newsgroups were those? alt.aol-sucks was certainly available from AOL, and I posted there frequently, often via AOL IIRC - in fact, although the flames were annoying and juvenile, some of us occasionally got useful bug reports there.
You should read Steven Pinker's The Language Instinct. You might also be interested in Hilary Putnam, as well as Saul Kripke (Naming and Necessity). I took a course in "Philosophy of Language" a few years ago, and this idea was discussed, among other things. Noam Chomsky might also be good reading, but unfortunately his works have never been translated into English.
One of them used a similar scenario to support the idea of an inner abstract language: If you know there's a word in English, but can't remember what the word is, what language are you using to think about that word?
I am fairly certain that the reason one looks like a dork when browsing the Internet on one's Palm is not the awkward angles required for infrared cellphone links.
I remember reading that Microsoft did not want to support it because it was too open as a standard.
Actually, Microsoft Hardware has already introduced the first products in a line of Bluetooth accessories to replace their old wireless keyboards and mice...
They have only ever sued people for sharing music files in excess of a certain number, and even then only if the person is sharing a lot of popular, contemporary music.
So, in essence, the RIAA itself is providing a strong incentive to share indie music? That's pretty cool of them.
- Pepsi pays full price for each song; - Pepsi expects x% of the caps to be redeemed, and has budgeted for the promotion accordingly; - This site could cause 2x% or 3x% to be redeemed instead, thus - Pepsi and other companies might think twice about running such a promotion in the future, because it costs a lot to do so.
However, as others have pointed out, if only a small percentage of Pepsi drinkers are either (a) current iTunes users or (b) likely to sign up just for the promotion, then an even smaller number are (c) neither a nor b and (d) aware of this site and willing to save/write down bottlecaps for it.
Yeah, I never had a problem with it under W2K either. The minute I upgraded to XP, it started crashing. Which is odd, because to a user application, XP and W2K aren't all THAT different, and I wouldn't expect QT to do any kernel-mode, driver-level stuff. Ironically, I upgraded to XP partly to solve problems I was having with RealPlayer; my old ServerWorks-based motherboard had a bad AGP driver in Win2K, and the only fix was to upgrade to XP. Why only RealPlayer suffered, I'll never know.
I just took a peek at a file that used to crash QuickTime, and it actually plays to the end with 6.3, so maybe they've finally fixed the issue, or maybe my clean install of XP this summer did it. I even paid for QuickTime Pro (or whatever their authoring add-in is called) at one point just to see if I could get support, but never had any luck with it. Most of the.MOV files from the original adcritic.com would crash QT.
On my Windows XP boxes, QuickTime has been remarkably unstable through three major and countless minor releases. Crashes, weird artifacts that linger for the duration of playback, "corrupted" files that played fine under Win2K...
I wouldn't use QuickTime for authoring unless my audience was primarily Mac-based.
Slashdot Mirror
As another poster mentioned, Netscape Communicator (though not Mozilla) has been able to read your AOL mail via IMAP for lo these many years. Forgot about that.
I'd love to hear what "POP3 over IMAP" interface you say you saw. AOL's native mail system model is much closer to IMAP than POP, which is why we never developed a POP server. But when I was in charge of mail systems development, we did develop a fully functional IMAP interface, which was launched on CompuServe but never on AOL. It also provided the back end for many other services, including AOL Mail On The Web, since IMAP makes a nice, well-defined API. I'm glad marketing finally allowed it to be launched on AOL.
Jay, the ex-AOL Mail Guy
I'm still not clear how you knew it was him from the initial log - did your staff notice him coming in and leaving at the time the logs said, or was he just the most suspicious person there during that time? Surely the cafe wasn't otherwise empty.
That said, although you don't explicitly say so in the story, I'm assuming you were tracing only that particular MAC address, and so if it weren't him, your tcpdump wouldn't have shown anything at all. For some reason the story implied that you raced to the cafe and set up a trace on his particular Ethernet port, which is of course not how it's usually done unless you have a managed switch...
This is a story that starts with a sysadmin seeing a 419 scam, hearing that there was a black guy with a "suspicious" accent in his cafe, deciding that this must be our criminal, and deciding to read his e-mail to find out...
Right?
Yep, definitely - bouncing is a ridiculously expensive operation, especially when most of it can't be delivered and just sits there clogging up your queue. At AOL we had entire banks of separate machines just to handle the bounce queues - but of course moving mail from one machine to another is a load in itself.
I know we were always working on ways to push spamfighting back to the inbound relays, and so I imagine that what you'll see is SMTP rejects, not bounces.
Ha! I never cease to be amazed at what explanations floor-lore can come up with... reminds me of the Q-Link days when we used to tell customers to wrap their disk drives in tinfoil. (Which, to be fair, actually worked sometimes on Commodores, since they were usually hooked up to TVs which emitted huge amounts of radiation in close proximity to the poorly-shielded external floppy drive..)
I worked in development, not operations, so I don't remember specific episodes, but it sounds like Earthlink was getting caught up in AOL's spam filters. Here's a cite showing it happening at least once.
I'd love to hear who told you that one. There has never been anybody anywhere in the entire mail operations management chain, from Steve Case down to the guy sitting in the NOC at 3 a.m., who would randomly bounce mail to lower the load. That's quite an accusation.
Ah. Then that's an operational problem, either at AOL's end or yours, and nothing to do with refusing mail when the system's busy.
Fascinating - what did he actually say when you talked to him?
AOL doesn't tell the senders or the intended recipients that it's dropping emails
This is actually starting to change; they are doing more bouncing and less blackholing these days. The spammers set up their own QA accounts anyway, so hiding the bounces doesn't really prevent the spammers from figuring out the filters.
AOL was known to block all mail from random domains to lower its server load when things got overloaded
You do realize that this is standard sendmail behavior, and not some sort of nefarious AOL plot to make your mail disappear? Refusing connections at a high load average just delays delivery of the mail until the next retry. It isn't related to the OP's mention of mail disappearing, which sounds like a spam-filtering problem.
That said, when I was at AOL, delivery times were usually in the seconds to minutes - I can't remember the last time we actually had a significant load average on the inbound relays.
Imagine a powered-by-human ATM cash machine.
You mean, like, a bank?
alt.aol-sucks was not available to AOL users for quite a while.
Do you have a cite for that? I just did a bit of googling, and all I can find are posts that complained about the fact that AOL, by default, showed newsgroup titles from the config database instead of dotted names, so that alt.aol-sucks showed up as "Flames and complaints about AOL". Some folks assumed that AOL renamed the group for some nefarious reason and came up with the description themselves, and others said that AOL was censoring the group because they couldn't find it. IIRC, the option to switch the list display from descriptions to newsgroup names was either always present or was added fairly early on, making it a moot point.
I'm a pretty good Googler, I don't remember AOL ever not carrying the group, I've found several posts from that timeframe (not from AOL sycophants) explicitly stating that AOL *has* always carried the group, and I find it hard to believe that not one person complained to alt-aol-sucks about AOL not carrying that very newsgroup, but maybe I'm just not searching on the right phrases.
it is also worth noting that many web servers have multiple domains on one IP address using both virtual directories and virtual domains
I imagine that, since AOL also supplies DNS to the AOL client, they are blocking by name, not IP.
However, I wonder if blocking by IP would work too in this particular case. Spammers are selling something, and if you sell something it's almost always going to be with an SSL-encrypted link. SSL doesn't work with virtual domains, so either (a) all the spammers have non-virtual domains or (b) only the credit-card processing, which ends up being done by a few companies like CCBill, is SSL. If (a), blocking by IP would work too.
And the spammer that owns 3-4K domains? Many do. There isn't an easy way to search for them all
Sure there is. If they're being spammed, they'll show up and can be blocked. If they're not, it doesn't matter.
, AOL blocked newsgroups that were created to discuss (and flame of course) problems with AOL
Eh? Which newsgroups were those? alt.aol-sucks was certainly available from AOL, and I posted there frequently, often via AOL IIRC - in fact, although the flames were annoying and juvenile, some of us occasionally got useful bug reports there.
Jay, the ex-AOL Mail Guy
You should read Steven Pinker's The Language Instinct. You might also be interested in Hilary Putnam, as well as Saul Kripke (Naming and Necessity). I took a course in "Philosophy of Language" a few years ago, and this idea was discussed, among other things. Noam Chomsky might also be good reading, but unfortunately his works have never been translated into English.
One of them used a similar scenario to support the idea of an inner abstract language: If you know there's a word in English, but can't remember what the word is, what language are you using to think about that word?
I am fairly certain that the reason one looks like a dork when browsing the Internet on one's Palm is not the awkward angles required for infrared cellphone links.
I remember reading that Microsoft did not want to support it because it was too open as a standard.
Actually, Microsoft Hardware has already introduced the first products in a line of Bluetooth accessories to replace their old wireless keyboards and mice...
I doubt clouds buffer sound much. (Foam doesn't, either, btw. It damps vibrations and reflections but doesn't prevent transmission.)
Mass buffers sound. Clouds have relatively little mass - in an example in Scientific American, a cloud might have 1g of water per cubic meter.
The first paint to go on sale will of course be white.
Followed closely by eggshell, ivory, ice, buff, linen, winter, antique, champagne, candlelight, ecru, snow, alabaster, bone, pearl, oyster, bisque, frost, china, dove, cotton, cloud, and cream.
They have only ever sued people for sharing music files in excess of a certain number, and even then only if the person is sharing a lot of popular, contemporary music.
So, in essence, the RIAA itself is providing a strong incentive to share indie music? That's pretty cool of them.
I think the premise of the question was that:
- Pepsi pays full price for each song;
- Pepsi expects x% of the caps to be redeemed, and has budgeted for the promotion accordingly;
- This site could cause 2x% or 3x% to be redeemed instead, thus
- Pepsi and other companies might think twice about running such a promotion in the future, because it costs a lot to do so.
However, as others have pointed out, if only a small percentage of Pepsi drinkers are either (a) current iTunes users or (b) likely to sign up just for the promotion, then an even smaller number are (c) neither a nor b and (d) aware of this site and willing to save/write down bottlecaps for it.
So, probably, no problem.
Yeah, I never had a problem with it under W2K either. The minute I upgraded to XP, it started crashing. Which is odd, because to a user application, XP and W2K aren't all THAT different, and I wouldn't expect QT to do any kernel-mode, driver-level stuff. Ironically, I upgraded to XP partly to solve problems I was having with RealPlayer; my old ServerWorks-based motherboard had a bad AGP driver in Win2K, and the only fix was to upgrade to XP. Why only RealPlayer suffered, I'll never know.
.MOV files from the original adcritic.com would crash QT.
I just took a peek at a file that used to crash QuickTime, and it actually plays to the end with 6.3, so maybe they've finally fixed the issue, or maybe my clean install of XP this summer did it. I even paid for QuickTime Pro (or whatever their authoring add-in is called) at one point just to see if I could get support, but never had any luck with it. Most of the
On my Windows XP boxes, QuickTime has been remarkably unstable through three major and countless minor releases. Crashes, weird artifacts that linger for the duration of playback, "corrupted" files that played fine under Win2K...
I wouldn't use QuickTime for authoring unless my audience was primarily Mac-based.