Yeah, while I personally don't use myspace nor do I envision myself doing so (I lost touch with those friends on purpose!) I have no issues with it. And I do think its bringing up peoples level of comfort with a collaborative, interactive web. Livejournal started the ball rolling, and myspace is the next generation.
Of course, the nerds bitched and moaned about LJ too as I recall.
Sadly no. MySpace seems to have achieved critical mass, and I know a disturbing number of otherwise reasonable adults who've become hooked. It combines everything they want: a profile, pictures, messaging, hooking up with people they haven't seen in years, and horny teenage girls.
Technically the Vatican, and ultimately the Pope, is the ultimate authority about Catholicism. Bishops have been excommunicated for doing things they weren't supposed to, and plenty of priests have been reassigned to shitholes for being out of line. There is also a mechanism for localizing Vatican doctrine to each region, namely the Conference of Catholic Bishops for a given country. Their latitude is limited, and is supposed to be finding ways to more effectively adhere to Vatican promulgated doctrine rather than changing it to suit the locale.
But yeah, original poster was way off base by saying "Christian Church" as if such an entity existed, let alone was part of the papacy.
Uhm, right. How exactly do you "exclude" people from a fake holiday perpetuated by geeks on the Internet? And for fucks sake, who DOES something about it?
My initial reaction was the same. But then I read http://lopsa.org/about_history and am considering killing my SAGE membership (sadly) and instead becoming a LOPSA member. I always did kinda wonder about USENIX as a parent organization to a generalist sysadmin focused group.
This seems like it wouldn't be that hard to reproduce the core functionality. Obviously the management friendly graphs, reports, and the config management portion are more work but simply writing a tool that shows you what your box actually does sounds rather straight forward. And it would also seem that the 90/10 rule would come into play, in that you'll get most of the benefit by simply getting the raw data and doing some basic analysis on it without needing to spend all that money.
Thanks for pointing this out, now I can point my engineering team at it and see how fast they can whip up an equivalent.
Yeah, dear friend of mine has OCD and has been bounced from med to med with each helping her for a period of time and then the side effects overtaking the benefits.
I'd be shocked if both seratonin and dopamine weren't both involved, along with several other factors we've probably yet to figure out. The thing I'm most sure of is that the brain is way more complex than we yet comprehend, and it'll be a long time before we have a decent handle on why it doesn't work the same all the time.
Please, please don't suggest people try and implement network fast-paths in fpgas. They're just too damn slow. Prototype in them all you want, but turn it into a real ASIC when you go production.
Also, you do have to be careful (as you probably know) with Cisco in how you configure their devices to keep the maximum amount of traffic on the fast-path or you will take a beating by pushing too many packets onto their slowass cpus.
Junipers are awesome, cuz they let us unix geeks onto the router and give us a familiar way to interact with it. Hell, just cuz grep actually works on their devices makes them hands down better than Cisco (imho).
A shop I worked in a while back tried replacing some juniper routers that were accepting GigE and OC-3 connections with some linux boxes using various PCI cards to take the same connections and do bgp, etc. Wow what a disaster. It wasn't linux's fault per se, but rather that we replaced a well designed, highly specialized box doing a highly specialized job with a general purpose machine. Sure we had a 2.4Ghz CPU instead of a sub 1ghz one, but we ended up being slower due to the lack of high speed silicon that's really good at the simple shit required to route 90+% of the packets it handled.
Probably so, however my guess (and its just that, also not being a psych or neuroscience guy) is that the obsessive-compulsive behavior is fueled by dopamine (or other chemical) mishandling by the brain. Not sure if this would also cover true OCD, way outside my knowledge.
I do know that as someone who is highly susceptible to addictive behavior I hope to see a lot more research in this area. Behavior control is damn hard, and its frustrating to sometimes have to abandon things early to avoid letting them run my life.
There's more and more research emerging to support the hypothesis that any addiction to a substance without physically addictive qualities (i.e. crack and its ilk) are all rooted in the same dopamine reactions. http://www.freerepublic.com/focus/f-news/1669601/p osts is a short synopsis of a story I read in long form in Chicago magazine about a woman who took a drug that affected how her brain handled dopamine and ended up with a massive gambling addiction. Stopping the meds brought back her original problem but allowed her to almost effortlessly quit gambling.
All of these non-chemical addictions seem to have the same core symptoms. People do something that makes them feel good. They do it often and begin to notice other things don't feel good anymore, then they notice they need to do this new thing more and more to keep the good feeling coming. Just because our brain makes a chemical doesn't mean it won't acquire a tolerance to it.
Well, granted we don't know if the whole "this man is dangerous" is a lie or not, however the complaint sure made it feel that way. Regardless of the truthfulness I'd say it smells like witness tampering to my non-lawyer brain.
We do know for sure this guy lied about his status as a federal agent (which as I mentioned is a crime by itself), so it also seems reasonable everything else he said would be a fabrication. Especially since he simply said "there are things you don't know" instead of giving real reasons the way real cops generally would.
We need to make sure we don't convict people of things that are convenient to convict them on, but instead make sure that the law was violated.
Absolutely. However if we convict them on things they actually did, that are actually illegal, and its convenient to do so I don't really have an issue with it. Consider Al Capone for example. He did a lot more than evade paying his taxes, but it made sense to prosecute him on that the minute they had a good case to get a conviction from because the bad stuff was so much harder to get a conviction on and this got him off the streets. Sure I'd be happier had Mr. Ness gotten him on the violent crimes he is known to have committed, but better in prison for something he did than on the streets doing more violence.
If the law wasn't violated, but the person did something reprehensible, then the law is obviously not keeping up with the times.
I'm extremely reluctant to agree with that, because reprehensible is such a mutable word. Something you consider reprehensible I may consider virtous. Illegal should be focused on preventing one person from harming another, not on making people be nice to each other or be upstanding citizens.
Anything that potentially induces a witness to fail to testify, or to alter their testimony, qualifies at witness tampering. In this case lying to the in-laws was intended to create consequences for the witness if he were to proceed as a witness for the prosecution. The expectation on the witness's part would be that if he continued with testifying that the consequences would escalate, possibly to violence.
I'm surprised they're not also charging him with impersonating a federal agent which is a serious crime in and of itself. Though they may still lay such charges against him, and it sounds like it'd be pretty damn impossible for him to beat the wrap on that.
You forgot that/dev/random blocks rather quickly./dev/urandom would much more rapidly yield a file large enough to potentially contain his 700+ site network.
Also, the file should end in.vsd for easier opening by colleagues.
Best efforts to stamp it out? What planet are you on, or more importantly what Internet? Spam filtering by content analysis is a piss poor means of eliminating it.
The major problem, which the article correctly identifies, with today's email system is the utter lack of enforced identity verification. Even if you want it, there's no mechanisms to support it. The only thing you can do is accept all of that email, and then only read the stuff that's PGP signed. Combine that with the lack of ease of use of most encryption solutions today.
We need to make the sender do some work to put all the info necessary to advertise the validity of their message, and then let the recipient MTA and MUA do a minimum of work to verify that they want this message.
I do think the original article is a bit ambitious on the thought of finding ways to make computing resources expensive enough to prevent spam but cheap enough to be feasible for users when sending under this new scheme. However you don't actually need to accomplish that, if you make it such that a spammer has to either prove who they are or pay a huge trust penalty for not doing so then you're way ahead of the current situation.
I'd love to back this up as well, we use our NetApps for a mixture of CIFS, iSCSI, NFS and soon FC as well.
Everything just works, their boxes are incredibly stable. Ours are pushing 900 days of uptime, with zero service interruptions during that time.
We've had one hard disk fail, I got a call from NetApp support while at lunch and I literally had to argue with them to get them to just drop ship the drive and not send a tech along to replace it.
NetApp is a fantastic product, and really offers surprisingly good price performance. Especially now that the FAS line allows SATA attached as well as FCAL.
Solaris rocks for stability and scalability, though its pretty heavy at the low end. FreeBSD is the shit for small webserver type usage.
Amusingly enough I was playing with osCommerce the other day and found it quite heavy.
Solaris would be your friend in trying to figure out your problems though, because its by far the most instrumentable environment I know of in the unix world.
No, I'm basing my commentary off analyst data from the big firms, all of whom are projecting HUGE HD set deployment in 2006 and into 2007.
I was actually highly skeptical of HD before acquiring a set that was HD capable, though that wasn't my primary consideration when buying it. I bought LCD due to space constraints, and bought HD because it didn't make sense to me to save $100 and get one of the very few EDTV LCDs I could find.
I also find it difficult to believe that the FCC will push the date, but instead that you'll see all of the set-top boxes downsampling to 480i for SDTV sets. EDTV is going to get completely ignored, as it deserves to.
Considering you can now buy an HDTV for under $500 USD, and that price continues to fall regularly, I'm thinking you're really exagerating what folks won't buy.
The industry, and their associated analysts, are predicting significantly larger growth than I'm talking about. I think my middle of the road prediction is fairly safe.
I do think the HD-DVD versus Blu-Ray fight is going to slow adoption somewhat, but with the FCC unlikely to back off again on the deadline for switching to HD broadcasts, and with HD channels rolling out at the speed they are, I do believe a significant swing over the next 24 months into the HD camp is inevitable.
"I've never once had a crash with the FreeBSD or Linux monolithic kernels. And unless you're a kernel developer, you haven't either."
Hahahahahaha. You make me laugh.
I run a moderate sized production environment (around 150 servers) and we have kernel panics and OOPSes. Not terribly frequently, but more frequently than zero. And nobody here has written kernel code in years.
My experience over the last 5 or so years as a full time Linux admin, has been that Linux stability is inversely related to the load on a system. Run your system at 80-90% load and you'll experience a lot less stability than a system running at 20-30% load.
Windows suffers from the same general issue, though it has shown some marked improvements in stability on the server side in the last couple releases. Linux, in my experience, can't say the same. Parity is rapidly approaching between the stability of the two, and for me neither is cutting it.
Solaris and FreeBSD are the only two systems I've run in production environments where high load didn't seem to significantly affect stability. Sadly I don't get to run them more often.
Slashdot Mirror
No no, use http://richard.jones.name/google-hacks/gmail-files ystem/gmail-filesystem.html
Much easier way to do this. Still not a good idea, but at least a better implementation.
Yeah, while I personally don't use myspace nor do I envision myself doing so (I lost touch with those friends on purpose!) I have no issues with it. And I do think its bringing up peoples level of comfort with a collaborative, interactive web. Livejournal started the ball rolling, and myspace is the next generation.
Of course, the nerds bitched and moaned about LJ too as I recall.
Sadly no. MySpace seems to have achieved critical mass, and I know a disturbing number of otherwise reasonable adults who've become hooked. It combines everything they want: a profile, pictures, messaging, hooking up with people they haven't seen in years, and horny teenage girls.
Technically the Vatican, and ultimately the Pope, is the ultimate authority about Catholicism. Bishops have been excommunicated for doing things they weren't supposed to, and plenty of priests have been reassigned to shitholes for being out of line. There is also a mechanism for localizing Vatican doctrine to each region, namely the Conference of Catholic Bishops for a given country. Their latitude is limited, and is supposed to be finding ways to more effectively adhere to Vatican promulgated doctrine rather than changing it to suit the locale.
But yeah, original poster was way off base by saying "Christian Church" as if such an entity existed, let alone was part of the papacy.
Uhm, right. How exactly do you "exclude" people from a fake holiday perpetuated by geeks on the Internet? And for fucks sake, who DOES something about it?
My initial reaction was the same. But then I read http://lopsa.org/about_history and am considering killing my SAGE membership (sadly) and instead becoming a LOPSA member. I always did kinda wonder about USENIX as a parent organization to a generalist sysadmin focused group.
Sure, but he didn't. One wonders what else he means to do, but doesn't.
Just avoid the states listed at http://www.unmarried.org/common.html
This seems like it wouldn't be that hard to reproduce the core functionality. Obviously the management friendly graphs, reports, and the config management portion are more work but simply writing a tool that shows you what your box actually does sounds rather straight forward. And it would also seem that the 90/10 rule would come into play, in that you'll get most of the benefit by simply getting the raw data and doing some basic analysis on it without needing to spend all that money.
Thanks for pointing this out, now I can point my engineering team at it and see how fast they can whip up an equivalent.
Yeah, dear friend of mine has OCD and has been bounced from med to med with each helping her for a period of time and then the side effects overtaking the benefits.
I'd be shocked if both seratonin and dopamine weren't both involved, along with several other factors we've probably yet to figure out. The thing I'm most sure of is that the brain is way more complex than we yet comprehend, and it'll be a long time before we have a decent handle on why it doesn't work the same all the time.
Please, please don't suggest people try and implement network fast-paths in fpgas. They're just too damn slow. Prototype in them all you want, but turn it into a real ASIC when you go production.
Also, you do have to be careful (as you probably know) with Cisco in how you configure their devices to keep the maximum amount of traffic on the fast-path or you will take a beating by pushing too many packets onto their slowass cpus.
Junipers are awesome, cuz they let us unix geeks onto the router and give us a familiar way to interact with it. Hell, just cuz grep actually works on their devices makes them hands down better than Cisco (imho).
A shop I worked in a while back tried replacing some juniper routers that were accepting GigE and OC-3 connections with some linux boxes using various PCI cards to take the same connections and do bgp, etc. Wow what a disaster. It wasn't linux's fault per se, but rather that we replaced a well designed, highly specialized box doing a highly specialized job with a general purpose machine. Sure we had a 2.4Ghz CPU instead of a sub 1ghz one, but we ended up being slower due to the lack of high speed silicon that's really good at the simple shit required to route 90+% of the packets it handled.
Probably so, however my guess (and its just that, also not being a psych or neuroscience guy) is that the obsessive-compulsive behavior is fueled by dopamine (or other chemical) mishandling by the brain. Not sure if this would also cover true OCD, way outside my knowledge.
I do know that as someone who is highly susceptible to addictive behavior I hope to see a lot more research in this area. Behavior control is damn hard, and its frustrating to sometimes have to abandon things early to avoid letting them run my life.
There's more and more research emerging to support the hypothesis that any addiction to a substance without physically addictive qualities (i.e. crack and its ilk) are all rooted in the same dopamine reactions. http://www.freerepublic.com/focus/f-news/1669601/p osts is a short synopsis of a story I read in long form in Chicago magazine about a woman who took a drug that affected how her brain handled dopamine and ended up with a massive gambling addiction. Stopping the meds brought back her original problem but allowed her to almost effortlessly quit gambling.
All of these non-chemical addictions seem to have the same core symptoms. People do something that makes them feel good. They do it often and begin to notice other things don't feel good anymore, then they notice they need to do this new thing more and more to keep the good feeling coming. Just because our brain makes a chemical doesn't mean it won't acquire a tolerance to it.
Oh, so true.
Well, granted we don't know if the whole "this man is dangerous" is a lie or not, however the complaint sure made it feel that way. Regardless of the truthfulness I'd say it smells like witness tampering to my non-lawyer brain.
We do know for sure this guy lied about his status as a federal agent (which as I mentioned is a crime by itself), so it also seems reasonable everything else he said would be a fabrication. Especially since he simply said "there are things you don't know" instead of giving real reasons the way real cops generally would.
Absolutely. However if we convict them on things they actually did, that are actually illegal, and its convenient to do so I don't really have an issue with it. Consider Al Capone for example. He did a lot more than evade paying his taxes, but it made sense to prosecute him on that the minute they had a good case to get a conviction from because the bad stuff was so much harder to get a conviction on and this got him off the streets. Sure I'd be happier had Mr. Ness gotten him on the violent crimes he is known to have committed, but better in prison for something he did than on the streets doing more violence.
I'm extremely reluctant to agree with that, because reprehensible is such a mutable word. Something you consider reprehensible I may consider virtous. Illegal should be focused on preventing one person from harming another, not on making people be nice to each other or be upstanding citizens.
Anything that potentially induces a witness to fail to testify, or to alter their testimony, qualifies at witness tampering. In this case lying to the in-laws was intended to create consequences for the witness if he were to proceed as a witness for the prosecution. The expectation on the witness's part would be that if he continued with testifying that the consequences would escalate, possibly to violence.
I'm surprised they're not also charging him with impersonating a federal agent which is a serious crime in and of itself. Though they may still lay such charges against him, and it sounds like it'd be pretty damn impossible for him to beat the wrap on that.
I agree entirely about right tool right job. Just make sure if its the wrong tool I don't have to throw my toolbox away to get to the right one.
You forgot that /dev/random blocks rather quickly. /dev/urandom would much more rapidly yield a file large enough to potentially contain his 700+ site network.
.vsd for easier opening by colleagues.
Also, the file should end in
Best efforts to stamp it out? What planet are you on, or more importantly what Internet? Spam filtering by content analysis is a piss poor means of eliminating it.
The major problem, which the article correctly identifies, with today's email system is the utter lack of enforced identity verification. Even if you want it, there's no mechanisms to support it. The only thing you can do is accept all of that email, and then only read the stuff that's PGP signed. Combine that with the lack of ease of use of most encryption solutions today.
We need to make the sender do some work to put all the info necessary to advertise the validity of their message, and then let the recipient MTA and MUA do a minimum of work to verify that they want this message.
I do think the original article is a bit ambitious on the thought of finding ways to make computing resources expensive enough to prevent spam but cheap enough to be feasible for users when sending under this new scheme. However you don't actually need to accomplish that, if you make it such that a spammer has to either prove who they are or pay a huge trust penalty for not doing so then you're way ahead of the current situation.
Or the parent is falling back on the good ole American defense "She asked for it!"
I'd love to back this up as well, we use our NetApps for a mixture of CIFS, iSCSI, NFS and soon FC as well.
Everything just works, their boxes are incredibly stable. Ours are pushing 900 days of uptime, with zero service interruptions during that time.
We've had one hard disk fail, I got a call from NetApp support while at lunch and I literally had to argue with them to get them to just drop ship the drive and not send a tech along to replace it.
NetApp is a fantastic product, and really offers surprisingly good price performance. Especially now that the FAS line allows SATA attached as well as FCAL.
Solaris rocks for stability and scalability, though its pretty heavy at the low end. FreeBSD is the shit for small webserver type usage.
Amusingly enough I was playing with osCommerce the other day and found it quite heavy.
Solaris would be your friend in trying to figure out your problems though, because its by far the most instrumentable environment I know of in the unix world.
No, I'm basing my commentary off analyst data from the big firms, all of whom are projecting HUGE HD set deployment in 2006 and into 2007.
I was actually highly skeptical of HD before acquiring a set that was HD capable, though that wasn't my primary consideration when buying it. I bought LCD due to space constraints, and bought HD because it didn't make sense to me to save $100 and get one of the very few EDTV LCDs I could find.
I also find it difficult to believe that the FCC will push the date, but instead that you'll see all of the set-top boxes downsampling to 480i for SDTV sets. EDTV is going to get completely ignored, as it deserves to.
Considering you can now buy an HDTV for under $500 USD, and that price continues to fall regularly, I'm thinking you're really exagerating what folks won't buy.
The industry, and their associated analysts, are predicting significantly larger growth than I'm talking about. I think my middle of the road prediction is fairly safe.
I do think the HD-DVD versus Blu-Ray fight is going to slow adoption somewhat, but with the FCC unlikely to back off again on the deadline for switching to HD broadcasts, and with HD channels rolling out at the speed they are, I do believe a significant swing over the next 24 months into the HD camp is inevitable.
"I've never once had a crash with the FreeBSD or Linux monolithic kernels. And unless you're a kernel developer, you haven't either."
Hahahahahaha. You make me laugh.
I run a moderate sized production environment (around 150 servers) and we have kernel panics and OOPSes. Not terribly frequently, but more frequently than zero. And nobody here has written kernel code in years.
My experience over the last 5 or so years as a full time Linux admin, has been that Linux stability is inversely related to the load on a system. Run your system at 80-90% load and you'll experience a lot less stability than a system running at 20-30% load.
Windows suffers from the same general issue, though it has shown some marked improvements in stability on the server side in the last couple releases. Linux, in my experience, can't say the same. Parity is rapidly approaching between the stability of the two, and for me neither is cutting it.
Solaris and FreeBSD are the only two systems I've run in production environments where high load didn't seem to significantly affect stability. Sadly I don't get to run them more often.