Problem is that components (both software and hardware) that can actually withstand 24/7 duty are orders of magitude more expensive than similar systems that will go down for a maintenance reboot regularly.
Linux for example can theoretically run for a long period of time but you will at least want to do the fsck-cycle once in a while because the common filesystem (speaking ext2 now, no real expirience with xfs) tends to get polluted slowly. That's at least by my expirience where I get occassional fsck-warnings on otherwise perfectly fine systems. I don't know whether there is a zero maintenance FS available for linux that can actually run for years straight without potentially summing up little problems that will eventually enforce an out-of-shedule reboot at an inconvenient time.
And that's only the FS. There sure are other components that simply have not ever been timetested long enough to be sure that they will last say, 5yrs straight, without running into some sort of unexpected problem.
So if you really want 24/7 for a long time you'll either go for really expensive hardware and/or some kind of failover solution, both of which adds a lot of expense and complexity to any project.
Often it's a lot cheaper when you design the system with a sheduled downtime (once a day, week, month, year..) in mind. As long as the regular service outage is torable you can not only save a lot of money that way but also ensure that checks that require an offline system (db integrity, hardware stuff) run regularly and provide early warnings. Doing all that on an always-hot system causes a lot more headache.
This article from 2002 claims that most fingerprint readers available to joe user by that time were easy to fool. Easy as in: press a plastic bag filled with warm water on it to replay the last print. Are we looking at a new, better generation of readers today or are they still as insecure as they used to be?
I won't comment about the other points. But what if you leave your computer running Seti@home or another distributed computing application. Don't those tend to use a lot of power during idle times?
Yep. Anything that keeps your CPU busy will drive your power bill up. The CPU doesn't really care whether its chewing on Quake III frames or seti chunks.
It describes what could happen were CO2 concentrations in the atmosphere to increase by 1% annually. I don't know whether this is actually the current trend.
I'd argue that "free" as in price is the major selling point of linux because hardly any computer-newbie ever starts with linux. Average joe will not use linux because he can't afford windows. It's way easier to pirate a copy of windows than to sit down and learn linux.
It's rather the "power users" who eventually become curious about what this "linux-thing" is all about and who've probably been annoyed one too many times by wintendo screwing up for no apparent reason.
I bet anyone who considers themselves a power user can tell a little story about an install of windows that just "stopped working" one day. Be it after installing some piece of software, a driver, or just out of the blue for no apparent reason.
I've had my share of that and eventually I was sick of doing the re-install every couple months without even knowing what exactly went south this time.
Maybe it's all better nowadays, though. I've seen a recent copy of windows that had "Repair"-buttons all over the place...;-)
Funny...this cat will pay $70 for a ram chip built by some guy in China, but won't pay $75 for an OEM copy of an OS he OBVIOUSLY wants to use, made in the USA?
Probably he doesn't want to use it but is forced to use it? Probably the ram chip won't be spyware infected within 10 minutes after installation?
This is insane. The only way to even get close to that size of an inbox would be a constant stream of huge attachments. Huge attachments are a horrible waste of ressources and the worst imaginable way to transfer a large file because the base64-encoding blows up your attachment by 1/3.
Attach a 450M file to an E-Mail and you'll be sending 600M, 150M wasted. Attach 1GB to an e-mail and you'll be wasting 333MB.
How can they encourage users to do such a stupid thing? Don't they have to pay for their bandwidth?
I really like that idea. The same mechanism could be used to lift the lion's cage at the same moment and lock all doors/windows! Lion must be trained to return to cage after dinner where a DNA-sensor can return everything back to normal when it detects lion-kaka mixed with human DNA.
The system would be pretty much zero maintenance (and no lion-food costs) as long as you have a steady supply of thieves.
the simplicity and ease of use that LaTeX provides
Sorry, I read that and have to kick in. Since I discovered LaTeX and spent the time to learn the basics I'm using it for just about any document that I have to deliver on dead tree. That's mainly letters, faxes and occassionally a documention or somesuch.
It's true that latex produces stunningly beautiful output that "magically" looks right most of the time (unless you're really screwing with things) and generally makes most.doc's (well, all I have seen) look cheap in comparison. I think you're stretching it a bit with your simplicity and ease of use statements, though. It took me a week only to get familar with the very basics; how to install the damn thing, where to put stuff so latex finds it, which commands must be called in what order to get anything out of it, basic LaTex-syntax, how to decipher LaTeX error messages to navigate through dependency-hell, which classes to use, how to set options and in what order and by which method (different classes use different styles), how to get proper pdf output without crippling the fonts, how to position images at least close to where i want them etc. etc.
With all that knowledge I'm still far from being able to do something supposedly simple as, say, design a custom letter head or footer. To do that I'd need to dive even deeper into latex-syntax and actually tamper with the classes that do the real work under the hood.
Someday I'll probably find the time and do it, but my point here is that simplicity is definately not something I'd attribute to LaTeX.
Quite a bit offtopic but maybe someone has had similar problems. Does anyone know how to make firefox properly display inline video (embedded into HTML pages)?
I tried the mplayer plugin but it doesn't work for me. Videos just pop up in a new window, play for a few secs, then close and won't come back.
Dude it's windows. It really doesn't matter. Just a few days ago I updated this officially licensed and paid for XP box at work via official windows update. You know, service pack 2 and stuff. Since then the virus scanner (antivir) is broken - the guard-service will pop up, complain about a missing dll and die. Uninstall/Reinstall of antivir didn't help (the dll-error seems gone but the symptom is the same). Also the poor soul who has to work at that box reports that apps randomly choke for up to one minute and everything seems horribly slow now. After a couple runs of the virusscanner, ad-aware and some other spyware-cleaning tools the box seems virus- and spyware-free. But the problems remain.
So, what exactly is funny about people pulling their broken patches from P2P? I think paying for this crap in first place really is the funny part.
I'd like to add that the LSB dictates questionable decisions in many areas (RPM, SysV-init,/srv-directory invented from thin air) and many refuse to adopt it for one or another.
Oh is that a new feature? The one I had just had this crackling sound fx from the PSU and the funny smell shortly after. Interestingly it would stay on for a couple more minutes, then I shut it down cleanly but it never came up again...
Well, admittedly yes. But a smarter worm could probably hook into fopen() and friends like realtime virus scanners do and trigger the payload when an attempt to remove the worm (launch/install of antivirus software?) is noticed. There could be a remote trigger so the worm writer can flip the switch as soon as patches/removal tools become available. And finally there might be worm writers whose malicious intend outweights the fiscal one. Those would be more interested in creating a big headline than in assembling zombies for whatever job. Heck, some annoyed spam-fighter might go for a clean sweep to wipe out all the damned spam drones - if only for a week... The possibilities are endless as long as MS is providing the playground.
Why don't you give them a chance? Yes, they should change their name ASAP (as in NOW you morons!). But other than that it seems like a reasonable approach to me.
Python allows for rapid developement, performance critical parts can be done in C. Postgres, while no performance wonder, seems like a reasonable stable foundation. I have not seen their code (in fact all I know is from a glimpse at their website) but with some skilled and dedicated coders they should be able to get somewhere in a couple months.
Ofcourse there's always the possibility of a premature death but why are you bashing them like that? Only for the name? Admittedly, if they don't change that they'll be crushed by MS-legal quicker than I can bunzip that tarball...
Linux remote-root exploits just happen rarely and kernel exploits even more so.
But what excuse does the biggest software company in the world have to not fix the gaping security holes in their two most used and probably most sensitive applications, explorer and outlook? We are watching this weekly windows exploit drama not for months but for years now. It's getting really old and its not funny at all anymore.
The worms we have seen were pretty harmless in my book, I'm still waiting for the one that carries some more serious payload. Like wiping out all accessible drives (network volumes), saturating all network cards with malicious packets, stuff like that. MS probably needs that kind of wake up call but are they really that bone-headed to not see it coming?
Thanks for the info. Now what I wonder is, would intel be able to make a larger, more expensive P4 that can run at the same speed as the regular P4 but needs less cooling?
I'd happily pay up to 50% more for a >2GHZ cpu that can be cooled with a slow spinning 100mm fan - or even passive (unrealistic I suppose..).
I'd probably make up for the higher price through my power bills over a couple months and building a silent PC would become much easier + cheaper, too.
Slashdot Mirror
If you are using Maildir a simple script should do the trick:
# delete files that were last accessed >30days ago
find . -type f -atime +30 -exec rm {} \;
Problem is that components (both software and hardware) that can actually withstand 24/7 duty are orders of magitude more expensive than similar systems that will go down for a maintenance reboot regularly.
Linux for example can theoretically run for a long period of time but you will at least want to do the fsck-cycle once in a while because the common filesystem (speaking ext2 now, no real expirience with xfs) tends to get polluted slowly. That's at least by my expirience where I get occassional fsck-warnings on otherwise perfectly fine systems.
I don't know whether there is a zero maintenance FS available for linux that can actually run for years straight without potentially summing up little problems that will eventually enforce an out-of-shedule reboot at an inconvenient time.
And that's only the FS. There sure are other components that simply have not ever been timetested long enough to be sure that they will last say, 5yrs straight, without running into some sort of unexpected problem.
So if you really want 24/7 for a long time you'll either go for really expensive hardware and/or some kind of failover solution, both of which adds a lot of expense and complexity to any project.
Often it's a lot cheaper when you design the system with a sheduled downtime (once a day, week, month, year..) in mind. As long as the regular service outage is torable you can not only save a lot of money that way but also ensure that checks that require an offline system (db integrity, hardware stuff) run regularly and provide early warnings.
Doing all that on an always-hot system causes a lot more headache.
This article from 2002 claims that most fingerprint readers available to joe user by that time were easy to fool. Easy as in: press a plastic bag filled with warm water on it to replay the last print.
Are we looking at a new, better generation of readers today or are they still as insecure as they used to be?
I won't comment about the other points. But what if you leave your computer running Seti@home or another distributed computing application. Don't those tend to use a lot of power during idle times?
Yep. Anything that keeps your CPU busy will drive your power bill up. The CPU doesn't really care whether its chewing on Quake III frames or seti chunks.
It describes what could happen were CO2 concentrations in the atmosphere to increase by 1% annually. I don't know whether this is actually the current trend.
1% annual increase of CO2 has indeed been the trend over the past 40 years.
It took me 15 secs to find that link via google. Maybe you should have spent these to avoid humiliating yourself...
I'd argue that "free" as in price is the major selling point of linux because hardly any computer-newbie ever starts with linux. Average joe will not use linux because he can't afford windows. It's way easier to pirate a copy of windows than to sit down and learn linux.
;-)
It's rather the "power users" who eventually become curious about what this "linux-thing" is all about and who've probably been annoyed one too many times by wintendo screwing up for no apparent reason.
I bet anyone who considers themselves a power user can tell a little story about an install of windows that just "stopped working" one day. Be it after installing some piece of software, a driver, or just out of the blue for no apparent reason.
I've had my share of that and eventually I was sick of doing the re-install every couple months without even knowing what exactly went south this time.
Maybe it's all better nowadays, though. I've seen a recent copy of windows that had "Repair"-buttons all over the place...
Funny...this cat will pay $70 for a ram chip built by some guy in China, but won't pay $75 for an OEM copy of an OS he OBVIOUSLY wants to use, made in the USA?
Probably he doesn't want to use it but is forced to use it?
Probably the ram chip won't be spyware infected within 10 minutes after installation?
This is insane. The only way to even get close to that size of an inbox would be a constant stream of huge attachments. Huge attachments are a horrible waste
of ressources and the worst imaginable way to transfer a large file because the base64-encoding blows up your attachment by 1/3.
Attach a 450M file to an E-Mail and you'll be sending 600M, 150M wasted.
Attach 1GB to an e-mail and you'll be wasting 333MB.
How can they encourage users to do such a stupid thing?
Don't they have to pay for their bandwidth?
I really like that idea. The same mechanism could be used to lift the lion's cage at the same moment and lock all doors/windows!
Lion must be trained to return to cage after dinner where a DNA-sensor can return everything back to normal when it detects lion-kaka mixed with human DNA.
The system would be pretty much zero maintenance (and no lion-food costs) as long as you have a steady supply of thieves.
the simplicity and ease of use that LaTeX provides
.doc's (well, all I have seen) look cheap in comparison.
Sorry, I read that and have to kick in.
Since I discovered LaTeX and spent the time to learn the basics I'm using it for just about any document that I have to deliver on dead tree. That's mainly letters, faxes and occassionally a documention or somesuch.
It's true that latex produces stunningly beautiful output that "magically" looks right most of the time (unless you're really screwing with things) and generally makes most
I think you're stretching it a bit with your simplicity and ease of use statements, though.
It took me a week only to get familar with the very basics; how to install the damn thing, where to put stuff so latex finds it, which commands must be called in what order to get anything out of it, basic LaTex-syntax, how to decipher LaTeX error messages to navigate through dependency-hell, which classes to use, how to set options and in what order and by which method (different classes use different styles), how to get proper pdf output without crippling the fonts, how to position images at least close to where i want them etc. etc.
With all that knowledge I'm still far from being able to do something supposedly simple as, say, design a custom letter head or footer.
To do that I'd need to dive even deeper into latex-syntax and actually
tamper with the classes that do the real work under the hood.
Someday I'll probably find the time and do it, but my point here is that simplicity is definately not something I'd attribute to LaTeX.
As usual, you get what you pay (lifetime) for.
Quite a bit offtopic but maybe someone has had similar problems.
Does anyone know how to make firefox properly display inline video (embedded into HTML pages)?
I tried the mplayer plugin but it doesn't work for me. Videos just pop up in a new window, play for a few secs, then close and won't come back.
the entire Microsoft organisation (which undoubtedly employs some of the world's finest software engineers and quality assurance experts)
What ever happened to judging people by their results...
Dude it's windows. It really doesn't matter.
Just a few days ago I updated this officially licensed and paid for XP box at work via official windows update. You know, service pack 2 and stuff.
Since then the virus scanner (antivir) is broken - the guard-service will pop up, complain about a missing dll and die.
Uninstall/Reinstall of antivir didn't help (the dll-error seems gone but the symptom is the same).
Also the poor soul who has to work at that box reports that apps randomly choke for up to one minute and everything seems horribly slow now.
After a couple runs of the virusscanner, ad-aware and some other spyware-cleaning tools the box seems virus- and spyware-free. But the problems remain.
So, what exactly is funny about people pulling their broken patches from P2P?
I think paying for this crap in first place really is the funny part.
So you mean it's... louder?
I'd like to add that the LSB dictates questionable decisions in many areas (RPM, SysV-init, /srv-directory invented from thin air) and many refuse to adopt it for one or another.
KHAAAAAAAAAAAAAAAAAN!
It's turtle-shaped.
Oh is that a new feature?
The one I had just had this crackling sound fx from the PSU and the funny smell shortly after. Interestingly it would stay on for a couple more minutes, then I shut it down cleanly but it never came up again...
So, uhm, now I've read two pages about how someone hacks up music in perl. But I still have no idea what it sounds like. Anyone have a link?
which box do you think they will reach for?
Well, the boxes look so similar, mom and dad might buy the Carbon by accident!
Well, admittedly yes. But a smarter worm could probably hook into fopen() and friends like realtime virus scanners do and trigger the payload when an attempt to remove the worm (launch/install of antivirus software?) is noticed. There could be a remote trigger so the worm writer can flip the switch as soon as patches/removal tools become available.
And finally there might be worm writers whose malicious intend outweights the fiscal one. Those would be more interested in creating a big headline than in assembling zombies for whatever job. Heck, some annoyed spam-fighter might go for a clean sweep to wipe out all the damned spam drones - if only for a week...
The possibilities are endless as long as MS is providing the playground.
Why don't you give them a chance?
Yes, they should change their name ASAP (as in NOW you morons !).
But other than that it seems like a reasonable approach to me.
Python allows for rapid developement, performance critical parts can be done in C. Postgres, while no performance wonder, seems like a reasonable stable foundation. I have not seen their code (in fact all I know is from a glimpse at their website) but with some skilled and dedicated coders they should be able to get somewhere in a couple months.
Ofcourse there's always the possibility of a premature death but why are you bashing them like that? Only for the name?
Admittedly, if they don't change that they'll be crushed by MS-legal quicker than I can bunzip that tarball...
Linux remote-root exploits just happen rarely and kernel exploits even more so.
But what excuse does the biggest software company in the world have to not fix the gaping security holes in their two most used and probably most sensitive applications, explorer and outlook?
We are watching this weekly windows exploit drama not for months but for years now. It's getting really old and its not funny at all anymore.
The worms we have seen were pretty harmless in my book, I'm still waiting for the one that carries some more serious payload. Like wiping out all accessible drives (network volumes), saturating all network cards with malicious packets, stuff like that. MS probably needs that kind of wake up call but are they really that bone-headed to not see it coming?
Thanks for the info.
Now what I wonder is, would intel be able to make a larger, more expensive P4 that can run at the same speed as the regular P4 but needs less cooling?
I'd happily pay up to 50% more for a >2GHZ cpu that can be cooled with a slow spinning 100mm fan - or even passive (unrealistic I suppose..).
I'd probably make up for the higher price through my power bills over a couple months and building a silent PC would become much easier + cheaper, too.
Depends. Can you play doom3 on your radiator?