Not to hijack this rant, but that's one of the problems with Google Talk: words have meanings. For instance:
* talk * voice * chat
See how someone might get a little confused by what "Google Talk" is? They've got a branding problem with the product, because many people still don't know what it is (as a product). Remember, we're talking about a population who probably, by and large, just thinks that a pop-up with a friend sending them a message in their email is just another email interface.
Now ask yourselves, do you think "What do you mean you've got to put this @jabber.org email address at the end of your hangout to talk?" would go over well? I've heard people (younger people, granted) communicate with each other and say "what's your gmail address?" and they just share the prefix in return. Many of these people don't seem to realize there is anything out there aside from gmail for mail, or that gmail is in no way exclusive. For younger people, it's all there has been for as long as they remember. (People who were 8 when Google came out are now 18...)
"97% of scientiss agree" makes a nice sound bite and article headline, but reality is somewhat different than this assertion. This assertion in and of itself is pretty demonstrative of the kind of science we can frequently see out of climatologists...
Look, it's pretty simple: a consensus of opinion does not mean a consensus or correctness of fact. You can have 3 people in a room of 100 saying things which get drowned out and castigated by the ignorant 97 - and this happens quite often.
What's ironic is that there's actually a lot of push back against all the PowerShell exclusivity MS is peddling. "Why would I pay huge gobs of money to have to use a command line? Wasn't that the whole point of Windows in the first point?" I'm sorry, but there is little point in removing functionality from a GUI just to put it back into an arbitrary and vague typed command you've got to dig through documentation to find. That's what MS has been doing.
The fact that PS "pipes" (not technically) arbitrary objects between themselves is interesting, but the implementation of the language is, for the most part, a clusterfuck. (And people say perl is difficult to learn/use/etc.).
Some of us saw this coming since the day it was announced.
a) This is big government we're talking about. We're lucky the costs aren't higher. b) Since when has anything orchestrated by this administration been anything but the opposite of what claimed? c) We're lucky they're actually telling us at this point.
$2k will buy you a pretty decent system, not just a "home NAS" box. If you've got 10TB of data already, chances are your use case is somewhat more esoteric than the "home NAS user" - for instance, you deal a lot in video. I understand your point, but 12 disks is quite a lot; you're looking at at least $3k for that, realistically.
Personally, I've got a lot of recycled equipment for my storage needs. I don't have more than 4TB of hot data, but I do have it duplicated across 3 systems right now using zfs snapshots - and the machines dual purpose as VirtualBox virtualization hosts, as well.
W7 was more an incremental upgrade from Vista. More like a parallel-developed version of Vista that never got released initially. It's just too different.
I'd say the curve should be closer to 5 or 6 years, depending on what it is we're talking about. If we're talking about a storage system, I'll start replacing disks sooner than later in the hopes of offsetting a catastrophic EOL failure.
Video cards are about 3 years for me, personally. But general computing equipment? I don't have a problem running it into the ground, as long as there isn't important data on it. 5 years is a good benchmark. If it starts to fail after 4, I just chuck it and start over. It's all about the cost potential involved.
Why are "IT decision makers" listening to anyone outside their organization when it comes to actual decisions?
I'm sorry, but they have many people in their organizations who have informed opinions on equipment - such as the people who have to work with them. Things like, "These NetApps are shit, let's go with someone else" or "we need new switches, these are dropping packets and are totally fabric saturated". Employees tell their bosses this stuff all the time; they know it amongst themselves as well.
What's more, it takes what, 15 minutes to get a feel for how bad a product is online before purchasing it - 30-60 if you don't really know what you're looking for or aren't too familiar with the technology?
Why are purchasing deciders making decisions in a vacuum when there is more than enough information available?
I suspect that this trick would still work today on a great number of unconnected campus and library cards. In 2000, my campus had us use magstrip cards for lunch and lab printing - it was a shared pool you paid for once a semester, and when you ran out, you ran out. The printers weren't network connected, and in retrospect, that'd have to make the cafeteria "readers" also writers, unless they didn't actually count printing (I don't remember - it was an arbitrarily large amount for food, maybe $2500, and $0.05 per printed page didn't hurt that...)
I don't remember what happened if anyone lost a card, I never heard of it happening. But if it did happen, I'm guessing that the user, er student would receive a new card with total original amount minus whatever could possibly be used eating lunch every day until that point, if the above is true.
I suspect that number is wildly conservative. That's crazy, when you consider the costs associated with:
* Multiple FT "Exchange Admins" * Needing people on-staff who actually understand email * If they were using something like Forefront and/or additional spam services as well (additional $$$) * Dozens of servers they no longer need to maintain maintain and replace * Tens of terabytes of fast, redundant storage they no longer need to keep on-premises
Due to the cost of such a large migration (will they be migrating existing mail, I wonder, or just keeping it on a network-mapped share for archival access?) I have to wonder how long this will take.
I'd have thought the per-year savings would be closer to a million than a quarter mil, personally.
Don't fool yourself. What do you think the whole 'point' behind making smokeless powder the explosive during the Boston Bombing is? They're going to push up smokeless power (and thus firearm ammo) regulation on the agenda, because they're quite obviously failing on firearm regulation. It's the same reason that DHS has been buying up all available domestic ammunition (at over 260 million rounds now!) and they're trying to ban importation of ammunition outright.
Doesn't matter how many guns you've got if you can't get ammo for 'em.
You're thinking in the context of mass attacks. That isn't going to happen.
The NK DMZ is 2.5 miles wide. It has no easy passage for vehicles and directly passing through the forest would be all but impossible. Reaching the other side, on foot - if you avoid being eaten or otherwise killed by wildlife or AP mines - would instantly result in being turned into hamburger and/or fine mist by a mix of automated turrets, mortars, etc. Any massing of troops in the forest, as detected by airborne infrared sensors, would immediately result in shelling of the area.
So really, a land passage isn't exactly tenable. There are small passages through these jungles and those are likewise guarded. They'd get shelled out within seconds of any indication of a convoy rolling down the road. (I don't care if they are well trained soldiers, they've got to either walk or ride vehicles, and it takes a long time to move even a fraction of a million people, well trained or not.)
So really, the only tenable way for NK to get actual troops and their associated Chinese vehicles to SK is by sea or air. How well do you think that will work?
Here's a hint: NK uses 1950s-1970s Soviet technology for pretty much everything they do that's "advanced". That means most of what they do is one-off and poorly assembled; they are easily 70 years behind the West at this point in basic industrialization, and they're even further behind if you consider what they are able to produce domestically. NK would almost instantly fall apart internally if they expended the time, energy, and resources to engage in a war - in the matter of days, people would be dying of starvation in high numbers. Posturing alone is likely too much for them to sustain for long.
Within institution's, who's problem is this to fix?
Obviously, this is the developer's (and PM's) fault. They're horrible at their jobs and write lazy, insecure software.
But this is probably going to fall on the shoulders of Google's in-house IT department to get resolved (likely by pushing at Niagara's support channel). Meanwhile, the IT department is also answerable to and for everyone else's snafus in-house in most organizations.
Developers - as individuals and departments - really need to pull their shit together and take more responsibility for their products. This is pretty disheartening and, on a daily basis, frustrating as fuck.
In more metropolitan areas, you can get same day shipping from Amazon already. Order it at work at noon and it's sitting on your doorstep at home when you get there. It's pretty crazy!
So how is it that you build your 'unstable' applications, using the Debian-approved package system (deb-src packages) from later releases? Or do you just download tarballs?
I've noticed that the latter of those two methods is something devs are particularly prone to, often for no apparent reason. Just last year I found a php package tarball overlay a developer had been using, with the PHP release date being from 2005-2006. He used it on every project he manned, and sadly, he wasn't the exception IME. In this case, he just tarred it out over whatever the distribution packaged.
That's where debian really shines. Second to stable software is the ability to cleanly and easily upgrade without too many hickups; Debian is, in my experience, better than pretty much everything else in this regard.
I've personally got a physical machine that's been kicking around since Debian 2.0 but is now on Debian 7.0 as of last week. Despite running quite a few different services and having some pretty ugly alien'd packages installed, the upgrade process has gone smoothly every single time (since 5.0 at least, I can't recall too much further back but the package repo just works). This system does a LOT on my network still (mostly things I can lose w/o crying if the thing dies, but it's still doing things).
In theory (I've not checked to see if old release repositories are actually still available online; I think most are, or could be easily recreated locally), I could do a tiered upgrade from an old 2.0 machine (over 10 years old) today and get it 'current', if I wanted to. I can't do that with anything else: we've got 2-3 year old RHEL/Cent/Fedora/Gentoo machines which are, in their current form, all but un-upgradeable already because a release was missed/not performed when it needed to be. So instead you're stuck doing a full 'migration' to a new host, and that's always a bucket of fun...
I think the tiered distribution approach is Wrong - eg. 'stable, testing, unstable' - regardless of which distribution we're talking about.
What I would like to see is something of the following:
* Each distribution has a 'core' - a set of packages which varies very little, but remains stable. * each distribution has stable 'release points' for different application groups.
Basically, I might have a server which needs a well-tried kernel but I need up-to-date packages. I could always use backports, which fits 90% of the scenarios, but it still feels like a bit of a hack/inelegant. And then you end up with upgrade problems down the line, if you're not careful (not a big problem in my experience - not unless you're using some fubar 3rd party repos which don't properly document depenencies and requirements).
Granted, this would be slightly recombinant in implementation and exceedingly difficult, but being able to say "I'm up to date on Debian core desktop stable, with the Desktop testing package release" would be kind of nice.
No, Ubuntu does not pull from Debian "stable". Even a cursory look would tell you this is false.
I believe what they do is pull from either Debian experimental or debian sid (most likely). Even Debian 'testing' is more likely than 'stable', which is improbable at best...
Governmental immunity is made possible through simple denial or claimed culpability of the fault.
"You may have evidence of an insurmountable nature, but it's a conspiracy theory, and as everyone knows, conspiracy theories are false and to be disdained as quackery. So we're just going to dismiss it out of hand, because we control the entire system at this point, top down."
Incorrect. Here is one good reason (which expounds to several good reasons) why Mariadb is not better, regardless. (You're thinking like a developer, not someone who looks at the broader picture, like a sysadmin).
You're both absolutely wrong. There are very good reasons for using MySQL; namely:
* It is distribution supported. That probably means shit to most developers, who will usually write code on top of any hot new steaming pile of shit (whether a database, an application, or a framework), but to people who have to work with these things for a living, this is significant. * A lot of 3rd party frameworks and applications not only work preferentially with MySQL, many also require it. We're talking about things like, oh, MediaWiki, which may or may not also have a number of add-ons and other things which modify the MySQL database. * Much of the documentation still targets MySQL specifically. * Many of the tools for db administration target MySQL.
Yeah, I know: there are many reasons why people would want to use Postgres or MariaDB. But this is not an invalid list.
Slashdot Mirror
Not to hijack this rant, but that's one of the problems with Google Talk: words have meanings. For instance:
* talk
* voice
* chat
See how someone might get a little confused by what "Google Talk" is? They've got a branding problem with the product, because many people still don't know what it is (as a product). Remember, we're talking about a population who probably, by and large, just thinks that a pop-up with a friend sending them a message in their email is just another email interface.
Now ask yourselves, do you think "What do you mean you've got to put this @jabber.org email address at the end of your hangout to talk?" would go over well? I've heard people (younger people, granted) communicate with each other and say "what's your gmail address?" and they just share the prefix in return. Many of these people don't seem to realize there is anything out there aside from gmail for mail, or that gmail is in no way exclusive. For younger people, it's all there has been for as long as they remember. (People who were 8 when Google came out are now 18...)
"97% of scientiss agree" makes a nice sound bite and article headline, but reality is somewhat different than this assertion. This assertion in and of itself is pretty demonstrative of the kind of science we can frequently see out of climatologists...
Look, it's pretty simple: a consensus of opinion does not mean a consensus or correctness of fact. You can have 3 people in a room of 100 saying things which get drowned out and castigated by the ignorant 97 - and this happens quite often.
What's ironic is that there's actually a lot of push back against all the PowerShell exclusivity MS is peddling. "Why would I pay huge gobs of money to have to use a command line? Wasn't that the whole point of Windows in the first point?" I'm sorry, but there is little point in removing functionality from a GUI just to put it back into an arbitrary and vague typed command you've got to dig through documentation to find. That's what MS has been doing.
The fact that PS "pipes" (not technically) arbitrary objects between themselves is interesting, but the implementation of the language is, for the most part, a clusterfuck. (And people say perl is difficult to learn/use/etc.).
Some of us saw this coming since the day it was announced.
a) This is big government we're talking about. We're lucky the costs aren't higher.
b) Since when has anything orchestrated by this administration been anything but the opposite of what claimed?
c) We're lucky they're actually telling us at this point.
$2k will buy you a pretty decent system, not just a "home NAS" box. If you've got 10TB of data already, chances are your use case is somewhat more esoteric than the "home NAS user" - for instance, you deal a lot in video. I understand your point, but 12 disks is quite a lot; you're looking at at least $3k for that, realistically.
Personally, I've got a lot of recycled equipment for my storage needs. I don't have more than 4TB of hot data, but I do have it duplicated across 3 systems right now using zfs snapshots - and the machines dual purpose as VirtualBox virtualization hosts, as well.
W7 was more an incremental upgrade from Vista. More like a parallel-developed version of Vista that never got released initially. It's just too different.
I'd say the curve should be closer to 5 or 6 years, depending on what it is we're talking about. If we're talking about a storage system, I'll start replacing disks sooner than later in the hopes of offsetting a catastrophic EOL failure.
Video cards are about 3 years for me, personally. But general computing equipment? I don't have a problem running it into the ground, as long as there isn't important data on it. 5 years is a good benchmark. If it starts to fail after 4, I just chuck it and start over. It's all about the cost potential involved.
Why are "IT decision makers" listening to anyone outside their organization when it comes to actual decisions?
I'm sorry, but they have many people in their organizations who have informed opinions on equipment - such as the people who have to work with them. Things like, "These NetApps are shit, let's go with someone else" or "we need new switches, these are dropping packets and are totally fabric saturated". Employees tell their bosses this stuff all the time; they know it amongst themselves as well.
What's more, it takes what, 15 minutes to get a feel for how bad a product is online before purchasing it - 30-60 if you don't really know what you're looking for or aren't too familiar with the technology?
Why are purchasing deciders making decisions in a vacuum when there is more than enough information available?
No, it's a direct hit to taxpayers.
Banks are insured FDIC. Do you know what that means?
I suspect that this trick would still work today on a great number of unconnected campus and library cards. In 2000, my campus had us use magstrip cards for lunch and lab printing - it was a shared pool you paid for once a semester, and when you ran out, you ran out. The printers weren't network connected, and in retrospect, that'd have to make the cafeteria "readers" also writers, unless they didn't actually count printing (I don't remember - it was an arbitrarily large amount for food, maybe $2500, and $0.05 per printed page didn't hurt that...)
I don't remember what happened if anyone lost a card, I never heard of it happening. But if it did happen, I'm guessing that the user, er student would receive a new card with total original amount minus whatever could possibly be used eating lunch every day until that point, if the above is true.
I suspect that number is wildly conservative. That's crazy, when you consider the costs associated with:
* Multiple FT "Exchange Admins"
* Needing people on-staff who actually understand email
* If they were using something like Forefront and/or additional spam services as well (additional $$$)
* Dozens of servers they no longer need to maintain maintain and replace
* Tens of terabytes of fast, redundant storage they no longer need to keep on-premises
Due to the cost of such a large migration (will they be migrating existing mail, I wonder, or just keeping it on a network-mapped share for archival access?) I have to wonder how long this will take.
I'd have thought the per-year savings would be closer to a million than a quarter mil, personally.
Don't fool yourself. What do you think the whole 'point' behind making smokeless powder the explosive during the Boston Bombing is? They're going to push up smokeless power (and thus firearm ammo) regulation on the agenda, because they're quite obviously failing on firearm regulation. It's the same reason that DHS has been buying up all available domestic ammunition (at over 260 million rounds now!) and they're trying to ban importation of ammunition outright.
Doesn't matter how many guns you've got if you can't get ammo for 'em.
What does this have to do with 'arms treaty exports'? As CAD files, absolutely nothing.
This is CAD files, blueprints. Don't let them fool you: it very much is about controlling firearm dissemination.
You're thinking in the context of mass attacks. That isn't going to happen.
The NK DMZ is 2.5 miles wide. It has no easy passage for vehicles and directly passing through the forest would be all but impossible. Reaching the other side, on foot - if you avoid being eaten or otherwise killed by wildlife or AP mines - would instantly result in being turned into hamburger and/or fine mist by a mix of automated turrets, mortars, etc. Any massing of troops in the forest, as detected by airborne infrared sensors, would immediately result in shelling of the area.
So really, a land passage isn't exactly tenable. There are small passages through these jungles and those are likewise guarded. They'd get shelled out within seconds of any indication of a convoy rolling down the road. (I don't care if they are well trained soldiers, they've got to either walk or ride vehicles, and it takes a long time to move even a fraction of a million people, well trained or not.)
So really, the only tenable way for NK to get actual troops and their associated Chinese vehicles to SK is by sea or air. How well do you think that will work?
Here's a hint: NK uses 1950s-1970s Soviet technology for pretty much everything they do that's "advanced". That means most of what they do is one-off and poorly assembled; they are easily 70 years behind the West at this point in basic industrialization, and they're even further behind if you consider what they are able to produce domestically. NK would almost instantly fall apart internally if they expended the time, energy, and resources to engage in a war - in the matter of days, people would be dying of starvation in high numbers. Posturing alone is likely too much for them to sustain for long.
Within institution's, who's problem is this to fix?
Obviously, this is the developer's (and PM's) fault. They're horrible at their jobs and write lazy, insecure software.
But this is probably going to fall on the shoulders of Google's in-house IT department to get resolved (likely by pushing at Niagara's support channel). Meanwhile, the IT department is also answerable to and for everyone else's snafus in-house in most organizations.
Developers - as individuals and departments - really need to pull their shit together and take more responsibility for their products. This is pretty disheartening and, on a daily basis, frustrating as fuck.
In more metropolitan areas, you can get same day shipping from Amazon already. Order it at work at noon and it's sitting on your doorstep at home when you get there. It's pretty crazy!
Same here. It doesn't hurt that I sound quite a bit like The Bruce when I say it. :P
Now, "woot"? That one needs to die. It was conceived when I was a kid, but I hear kids and pretty damn near everyone else saying it these days.
So how is it that you build your 'unstable' applications, using the Debian-approved package system (deb-src packages) from later releases? Or do you just download tarballs?
I've noticed that the latter of those two methods is something devs are particularly prone to, often for no apparent reason. Just last year I found a php package tarball overlay a developer had been using, with the PHP release date being from 2005-2006. He used it on every project he manned, and sadly, he wasn't the exception IME. In this case, he just tarred it out over whatever the distribution packaged.
Right, but at some point you have to upgrade.
That's where debian really shines. Second to stable software is the ability to cleanly and easily upgrade without too many hickups; Debian is, in my experience, better than pretty much everything else in this regard.
I've personally got a physical machine that's been kicking around since Debian 2.0 but is now on Debian 7.0 as of last week. Despite running quite a few different services and having some pretty ugly alien'd packages installed, the upgrade process has gone smoothly every single time (since 5.0 at least, I can't recall too much further back but the package repo just works). This system does a LOT on my network still (mostly things I can lose w/o crying if the thing dies, but it's still doing things).
In theory (I've not checked to see if old release repositories are actually still available online; I think most are, or could be easily recreated locally), I could do a tiered upgrade from an old 2.0 machine (over 10 years old) today and get it 'current', if I wanted to. I can't do that with anything else: we've got 2-3 year old RHEL/Cent/Fedora/Gentoo machines which are, in their current form, all but un-upgradeable already because a release was missed/not performed when it needed to be. So instead you're stuck doing a full 'migration' to a new host, and that's always a bucket of fun...
I think the tiered distribution approach is Wrong - eg. 'stable, testing, unstable' - regardless of which distribution we're talking about.
What I would like to see is something of the following:
* Each distribution has a 'core' - a set of packages which varies very little, but remains stable.
* each distribution has stable 'release points' for different application groups.
Basically, I might have a server which needs a well-tried kernel but I need up-to-date packages. I could always use backports, which fits 90% of the scenarios, but it still feels like a bit of a hack/inelegant. And then you end up with upgrade problems down the line, if you're not careful (not a big problem in my experience - not unless you're using some fubar 3rd party repos which don't properly document depenencies and requirements).
Granted, this would be slightly recombinant in implementation and exceedingly difficult, but being able to say "I'm up to date on Debian core desktop stable, with the Desktop testing package release" would be kind of nice.
No, Ubuntu does not pull from Debian "stable". Even a cursory look would tell you this is false.
I believe what they do is pull from either Debian experimental or debian sid (most likely). Even Debian 'testing' is more likely than 'stable', which is improbable at best...
Governmental immunity is made possible through simple denial or claimed culpability of the fault.
"You may have evidence of an insurmountable nature, but it's a conspiracy theory, and as everyone knows, conspiracy theories are false and to be disdained as quackery. So we're just going to dismiss it out of hand, because we control the entire system at this point, top down."
Incorrect. Here is one good reason (which expounds to several good reasons) why Mariadb is not better, regardless. (You're thinking like a developer, not someone who looks at the broader picture, like a sysadmin).
$ aptitude search mariadb
$ echo $?
0
You're both absolutely wrong. There are very good reasons for using MySQL; namely:
* It is distribution supported. That probably means shit to most developers, who will usually write code on top of any hot new steaming pile of shit (whether a database, an application, or a framework), but to people who have to work with these things for a living, this is significant.
* A lot of 3rd party frameworks and applications not only work preferentially with MySQL, many also require it. We're talking about things like, oh, MediaWiki, which may or may not also have a number of add-ons and other things which modify the MySQL database.
* Much of the documentation still targets MySQL specifically.
* Many of the tools for db administration target MySQL.
Yeah, I know: there are many reasons why people would want to use Postgres or MariaDB. But this is not an invalid list.
As a single man who is unable to become pregnant, I find it offensive that I am excluded from paternal leave.
Should I not get conception leave?