You don't know what the other random numbers correspond to, but if yours was 1928787 you know that your vote is
there and was counted as 'A'.
This works for me to verify that my 'A' vote counted, and I could download
the whole database and verify the math (do my own count of the posted votes)
but how do you know that dirty cheating 'B' didn't add an extra 10,000
votes for himself with random numbers? Is there a way to know the total
number of votes? (I don't think so.)
The article points out that there are problems with RBLs, and that is true.
On the other hand, they're very useful in blocking spam.
This is why I like spamassassin. It lets you look up DNSBLs, and include those in a mail's score. It combines these and distributed spam reporing services like razor (which could be abused, too, but only on a per-message basis, not whole sites or netblocks)
with its own content-based checks and an automated whitelist facility.
The personalization algorythems aren't all that smart, and don't have very much information to work on, anyway. So, they work a little bit, but people anthroporphise the personalizations, and get worked up about not much.
(My tastes and moods vary so much I mostly get random stuff, so these are rarely useful to me.)
On my servers, it rarely takes >1s to scan mails. If it takes longer on yours, the most likely reason is that it's timing out on some service. Check to see if your DNSBLs are working, and if you're using razor, make sure you aren't firewalling off UDP echos.
I hope this helps, even though its off-topic for this discussion.
Hey! i got it.. if we all clicked on all the ads and links in spam email would we/. them? if so.. would it technically be a dos attack? hehe... another way to stop spam i guess..
Except that about half of the spam I get has links that include identifiers to
verify which email address resulted in loading the page. (As far as I can tell, that's the only reason for some of the Korean spam.) So, doing that would make spam more efficient, and also be a selling point for the spamware sellers ("Be carefull! You might get so much responce it takes down your server!")
When spam outnumbers real posts, most users stop reading (and posting).
I remember some groups decided to "move" somewhere else in the tree to try to avoid spam -- this worked for a while, but it really wansn't practical to move a community over and over.
Simularly, many users cycle through email addresses when the old ones get on too many spam lists, but that means that friends can't keep up with them. (And really, email is the first and biggest p2p application.)
Using multiple p2p applications/transports/networks might make it harder to degrade the overall system, than using any one. A smart client might be able to automate that process.
Spam is a distributed problem, and will require
a distributed solution. Courts, law-enforcement, education (social engineering), and technical measures will each chip away at the problem.
Source-based filtering work best when the sources are concentrated and not moving (like when Sanford Wallace was making most of the noise.) This still works a little, and is the premise that all the various RBLs and DNS-BLs are based upon.
Content-based filtering works only when the content of the spam is either identical for a large number of victims over time (which is how razor works) or contains patterns that are very unlikely to appear in legitimate email. (Tools like spamassassin work well against these.)
If these technical measures against (obvious) spams were effective and universially applied, it would cut down on the volume of spam, but the spammers would get more subtle, and start sending spam that is very hard to detect.
Since most spammers do it only once (but there are a lot of them) it would likely help to educate the public that the spamware-salesmen are essentially con-artists. If it were illegal to send spam, this would be a lot easier.
Legal measures alone would likely be unenforcable, because of the sheer numbers of spammers, and the fact that its easier for them to get new accounts and other services than it is to track them down.
If I my offer an analogy, this is like people burguling my house. I can stop most of them by putting locks on my front door. For those that are determined enough to defeat those locks, the police will will stop them by sending lots of men and women with guns and handcuffs. It also helps if parents and schools teach their children that it's not right to steal.
I bought a replyTV when they first came out. Now they've been bought out, and the new company is wanting to do subscriptions -- after I already paid $995 (or so) for 7 hours + lifetime subscription. I thought that meant the lifetime of the device, but now I wonder if it means the lifetime of SonicBlue's patience with Replay's old marketing scheme.I've already put up with the addition of ads instead of content on my pause button; I wonder what other changes are coming.
Of course, I could just buy a new PVR, but why? It's not like there's enough "good" TV to justify me getting on a 2 or 3 year upgrade treadmill.
Haven't the courts already ruled that forcing somoene to label their speach is an infringement on free speach? (otherwise, it'd be much simpler to require and "adult" meta-tag.)
I think a TLD specifically for porn is a good idea, just like we have r- and x- rated movies. (Of course, those are run by industry groups, not mandated by congress.)
It seems to me this is entirely a user preference.
Sites should never open new windows, but users
who like new windows should just use button 2
(or whatever your browser requires.)
I know a guy who did this -- IPaq + linux (I forgot which distro) + 802.11 card + NFS gives him a very portable mp3 player. (Kinda bulky for my taste, but still a neat hack...)
The brutal truth is that, if the internet
were connected as an arbritary directed graph, routing would be impractical, and growth could not
have been sustained (and could not be sustained now). A heirarcical internet can be routed with N routers holding log(N) routes, but if everybody has independant (unagregatable) addreses, N routers must each track N routes. I suspect this would also imply quadradic flap rates as well.
The simplest way of building heirarcy into the network is to have provider-dependant addressing, which can be aggregated by the provider. For local regions, you can advertise your prefix through multiple providers, but expect that to also be aggregated when it hits the first expensive (eg, trans-oceanic) link.
I understand that the IPv6 group looked at geography-based heirarcical routing. I'm not sure why this was dropped (although one could guess), but it had two interesting side effects: First, it put small providers on more even ground than big providers; and second, it also made the senders of traffic pay for the long-haul / backbone bandwidth (as opposed to traffic usually finding the backbone closest to the reciever of the traffic).
I've had one of the first-generation replayTVs
for a couple years, and found that many comercials
are doing just that -- 35 seconds long, and with
the product name promently displayed for the last
five seconds. Of course, the 7 second backwards
jump makes it reasonably quick to fix any over-skips.
Anyway, it's very sweet to watch half-hour shows
in 25 miniutes or so, but I haven't seen anything
that *makes* me want to get a second unit (or even just upgrade) yet.
They seem to be expecting that everyone is just going to switch over this weird system over current net-speak.
Well, I once thought that this http protocol
would never catch on (since it was less versitile
than ftp), and all that extra typing of a URL
was silly because you really just had to give
the hostname of the ftp site and, if it wasn't
obvious, the directory to cd to in order to find
all your files.
It seems to be to be totally unfair and obnoxious for managers to try manipulating employees with arbitrary assignments instead of straighforwardly telling the employee what the problem is. It is impossible to control another human being. The best (only?) way to solve such problems is to gain that person's cooperation. Once that is done, exercises in responsibilty have a chance to be really effective.
I admit to having been there, and a conversation starting with "Dave, you're a real smart guy, but you're an asshole." showed me what was wrong, and
let *me* fix *my* problem. Maybe that won't work for everyone, but basic respect for people demands that you give them that chance.
The difference is, that BN secretly
tested prospective employees. That is useful
only for helping the company cherry-pick
the genetically least-risk employees.
Maybe this would be a good thing, if they just
provided that new employees be screened, and
let the employee make an informed choice. But
in this case, the BN employees loose
the ability to choose, and even loose
the right to be informed.
I am amazed that Microsoft still holds a dominant
position, while other OSs with greater features
(uptime, scalabilty, remote access etc) with
lower costs (about $20 for a FreeBSD or slackware CD vs over $400 for a Win2K license). The obvious differece is marketing -- in which Microsoft
plays songs that contain lyrics like "It makes
a grown man cry" and (something like -- I don't
know any latin) 'The confused are damned to hell'
Okay, I understand that fluff editors sometimes
let old news in (especially if it sounds scary)
but did Hemos read the article? It even says
it's a well-known vulnerability which has been addressed by recent software.
So, what did they gain by this landing,
besides showing they could do it? And
what do they hope to gain by lifting off
again?
I had assumed they were going to study
rotational movements (maybe even vibrations,
using phase shifts of the transmitter's signal?)
when they landed.
It's cool that they can do stuff like this,
and I understand that yet-more-pictures of
a rock in space wouldn't be much fun, but I
would have thought that even the "bonus science"
at the end of the mission would have been
better planned.
This would imply that wireless connections are
made "by arrangement" as opposed to someone
doing a search and discovering that I'm the
closest node to their bus stop.
I'd be happy to arrange semi-permanent connectivity with anyone who wants to, but I'd
also be very interested if anyone
has a way to share with the bus-stop user as well.
Just how do you deal with "drive by
spammers," or crackers using your network to
attack other internet sites? Are you just
hoping noone will?
(This is a serious question -- I'd like to
get involved and contribute a couple of APs
to the world, but only if I have some kind
of tracability/accountability to make sure
what I add is an asset to the network, not
a liability.)
I generally change a key only when a system (root)
is compromised.
Some people like to change
keys periodically (every 2 years or so) on
the pretext that doing so limits the vunerability
of having a host key brute-forced. (ie, it might
be changed by the time the cracker broke it, or
at least the cracker would be locked out eventually.) I don't like this, as it increases
the complexity of using the software, and trains
users to ignore key changes.
As other posters mentioned, ssh doesn't nicely
deal with multiple meanchines all being referenced by the same DNS name. (It looks like the key changes, when you're really just getting a different machine.)
Slashdot Mirror
This works for me to verify that my 'A' vote counted, and I could download the whole database and verify the math (do my own count of the posted votes) but how do you know that dirty cheating 'B' didn't add an extra 10,000 votes for himself with random numbers? Is there a way to know the total number of votes? (I don't think so.)
This is why I like spamassassin. It lets you look up DNSBLs, and include those in a mail's score. It combines these and distributed spam reporing services like razor (which could be abused, too, but only on a per-message basis, not whole sites or netblocks) with its own content-based checks and an automated whitelist facility.
(My tastes and moods vary so much I mostly get random stuff, so these are rarely useful to me.)
I hope this helps, even though its off-topic for this discussion.
--David Burns
Except that about half of the spam I get has links that include identifiers to verify which email address resulted in loading the page. (As far as I can tell, that's the only reason for some of the Korean spam.) So, doing that would make spam more efficient, and also be a selling point for the spamware sellers ("Be carefull! You might get so much responce it takes down your server!")
I remember some groups decided to "move" somewhere else in the tree to try to avoid spam -- this worked for a while, but it really wansn't practical to move a community over and over. Simularly, many users cycle through email addresses when the old ones get on too many spam lists, but that means that friends can't keep up with them. (And really, email is the first and biggest p2p application.)
Using multiple p2p applications/transports/networks might make it harder to degrade the overall system, than using any one. A smart client might be able to automate that process.
Why would you divide by altitude?
Source-based filtering work best when the sources are concentrated and not moving (like when Sanford Wallace was making most of the noise.) This still works a little, and is the premise that all the various RBLs and DNS-BLs are based upon. Content-based filtering works only when the content of the spam is either identical for a large number of victims over time (which is how razor works) or contains patterns that are very unlikely to appear in legitimate email. (Tools like spamassassin work well against these.) If these technical measures against (obvious) spams were effective and universially applied, it would cut down on the volume of spam, but the spammers would get more subtle, and start sending spam that is very hard to detect.
Since most spammers do it only once (but there are a lot of them) it would likely help to educate the public that the spamware-salesmen are essentially con-artists. If it were illegal to send spam, this would be a lot easier. Legal measures alone would likely be unenforcable, because of the sheer numbers of spammers, and the fact that its easier for them to get new accounts and other services than it is to track them down. If I my offer an analogy, this is like people burguling my house. I can stop most of them by putting locks on my front door. For those that are determined enough to defeat those locks, the police will will stop them by sending lots of men and women with guns and handcuffs. It also helps if parents and schools teach their children that it's not right to steal.
I bought a replyTV when they first came out. Now they've been bought out, and the new company is wanting to do subscriptions -- after I already paid $995 (or so) for 7 hours + lifetime subscription. I thought that meant the lifetime of the device, but now I wonder if it means the lifetime of SonicBlue's patience with Replay's old marketing scheme.I've already put up with the addition of ads instead of content on my pause button; I wonder what other changes are coming.
Of course, I could just buy a new PVR, but why? It's not like there's enough "good" TV to justify me getting on a 2 or 3 year upgrade treadmill.
IANAL, but I think this varies state-to-state;
This would make it illegal everywhere in the USA.
Haven't the courts already ruled that forcing somoene to label their speach is an infringement on free speach? (otherwise, it'd be much simpler to require and "adult" meta-tag.)
I think a TLD specifically for porn is a good idea, just like we have r- and x- rated movies. (Of course, those are run by industry groups, not mandated by congress.)
It seems to me this is entirely a user preference. Sites should never open new windows, but users who like new windows should just use button 2 (or whatever your browser requires.)
I know a guy who did this -- IPaq + linux (I forgot which distro) + 802.11 card + NFS gives him a very portable mp3 player. (Kinda bulky for my taste, but still a neat hack...)
The simplest way of building heirarcy into the network is to have provider-dependant addressing, which can be aggregated by the provider. For local regions, you can advertise your prefix through multiple providers, but expect that to also be aggregated when it hits the first expensive (eg, trans-oceanic) link.
I understand that the IPv6 group looked at geography-based heirarcical routing. I'm not sure why this was dropped (although one could guess), but it had two interesting side effects: First, it put small providers on more even ground than big providers; and second, it also made the senders of traffic pay for the long-haul / backbone bandwidth (as opposed to traffic usually finding the backbone closest to the reciever of the traffic).
"So, it's about heavy breathing and smoochin'?"
I really hope this doesn't turn out to be a soap opera...
Anyway, it's very sweet to watch half-hour shows in 25 miniutes or so, but I haven't seen anything that *makes* me want to get a second unit (or even just upgrade) yet.
Well, I once thought that this http protocol would never catch on (since it was less versitile than ftp), and all that extra typing of a URL was silly because you really just had to give the hostname of the ftp site and, if it wasn't obvious, the directory to cd to in order to find all your files.
Net-speak changes (and changes very rapidly).
Why does the frank conversation have to be last?
It seems to be to be totally unfair and obnoxious for managers to try manipulating employees with arbitrary assignments instead of straighforwardly telling the employee what the problem is. It is impossible to control another human being. The best (only?) way to solve such problems is to gain that person's cooperation. Once that is done, exercises in responsibilty have a chance to be really effective.
I admit to having been there, and a conversation starting with "Dave, you're a real smart guy, but you're an asshole." showed me what was wrong, and
let *me* fix *my* problem. Maybe that won't work for everyone, but basic respect for people demands that you give them that chance.
Maybe this would be a good thing, if they just provided that new employees be screened, and let the employee make an informed choice. But in this case, the BN employees loose the ability to choose, and even loose the right to be informed.
Hello
I am amazed that Microsoft still holds a dominant position, while other OSs with greater features (uptime, scalabilty, remote access etc) with lower costs (about $20 for a FreeBSD or slackware CD vs over $400 for a Win2K license). The obvious differece is marketing -- in which Microsoft plays songs that contain lyrics like "It makes a grown man cry" and (something like -- I don't know any latin) 'The confused are damned to hell'
So, how do you do it?
So, where's the story?
I had assumed they were going to study rotational movements (maybe even vibrations, using phase shifts of the transmitter's signal?) when they landed.
It's cool that they can do stuff like this, and I understand that yet-more-pictures of a rock in space wouldn't be much fun, but I would have thought that even the "bonus science" at the end of the mission would have been better planned.
--David Burns
This would imply that wireless connections are made "by arrangement" as opposed to someone doing a search and discovering that I'm the closest node to their bus stop.
I'd be happy to arrange semi-permanent connectivity with anyone who wants to, but I'd also be very interested if anyone has a way to share with the bus-stop user as well.
Just how do you deal with "drive by spammers," or crackers using your network to attack other internet sites? Are you just hoping noone will?
(This is a serious question -- I'd like to get involved and contribute a couple of APs to the world, but only if I have some kind of tracability/accountability to make sure what I add is an asset to the network, not a liability.)
Some people like to change keys periodically (every 2 years or so) on the pretext that doing so limits the vunerability of having a host key brute-forced. (ie, it might be changed by the time the cracker broke it, or at least the cracker would be locked out eventually.) I don't like this, as it increases the complexity of using the software, and trains users to ignore key changes.
As other posters mentioned, ssh doesn't nicely deal with multiple meanchines all being referenced by the same DNS name. (It looks like the key changes, when you're really just getting a different machine.)