Ogg Theora was pretty much a repeat of Ogg Vorbis, except that Theora is actually worse than the format it tries to replace and video is a much bigger patent minefield. Same open source fanatics, same arguments, very little corporate support, very little hardware support but somehow everything would be different this time around.
Hell, even the ogg container completely failed to compete with mkv. Just about the only thing coming from that direction I've seen in use is FLAC, which is apparently a quite good lossless format (size and complexity, lossless is of course lossless).
I think even Google will find that dethroning H.264 is a huge undertaking, if they didn't own YouTube I wouldn't think they stood a chance. If they can offer a superior WebM experience there, then maybe. But I'm still not going further than maybe.
It's really much simpler. Windows 7 and OS X has already licensed the codec, Microsoft has absolutely nothing to lose by pushing it. Firefox has problems with it, Linux has problems with it. When there's so few competitors, pushing them down is as good as lifting yourself up. Not to mention in public perception they don't want it to look like Google is leading the pack and Microsoft tagging along. There's so many political and strategical reasons to do it that far outweigh the minimal patent royalties they get.
A footnote in Mozilla's 2006 financial report states "Mozilla has a contract with a search engine provider for royalties. The contract originally expired in November 2006, however Google renewed the contract until November 2008 and has now renewed the contract through 2011.[8] Approximately 85% of Mozillaâ(TM)s revenue for 2006 was derived from this contract."
What is the status of the organization's contract with Google?
We have had a productive relationship with Google since 2004 and that relationship remains healthy. To date, we have renewed our contract three times, in 2005, 2006 and 2008. The current version extends through 2011.
So through 2011 Mozilla has a very good deal. But then Google didn't have a browser of their own and desperately needed Mozilla to break the IE monopoly. I suspect that these negotiations will go quite differently. I'm sure the deal will be extended but I doubt the terms will be anywhere near as favorable as they have been. Google has seen how easily they can now push their own browser into the market, they don't "need" Firefox that much anymore. And from a strict business point of view, where would they go? Bing? Yeah, I'm sure the open source community would love Microsoft as their default search engine. Not to mention that currently Chrome has targeted the IE holdouts. If they go their separate ways, Google will do their best to win Firefox users too. I'd put good money on the browser market looking completely changed in 2-3 years.
I think I remember seeing a fake dual vinyl player built just for this purpose in some magazine. Basically you hooked it up to the computer and controlled songs as if they played on vinyl, the driver also added suitable scratching if you wanted. Never mind all the other neat things you can do with the computer, it seemed like a superior solution. True, it's niche but so is doing that in the first place...
Sheldon isn't more of a living embodiment of OCD than he's a geek. On a TV show it makes for great comedy, in real life it'd drive you c-r-a-z-y. I have a buddy that's 1/10th of Sheldon and that's plenty...
This is a pretty good image from 2008 onwards at least. The predicted time has varied by a few months but the money has been on 2011 at least for the last three years.
Y2K was perfectly legitimate. It was only through heroic efforts that programmers were able to overcome years of managerial negligence and get the changes made in a knick of time. As is typical, since the herculean effort caused nothing to happen the world yawned and assumed the geeks were just moaning over nothing all along.
I'm sure that for many systems that was true. But to be honest, it passed with such a complete yawn that not only were all the important systems fixed, it seems all the "nice to have" systems were fixed as well. I don't think it's any exaggeration to say that many people and companies used y2k as a tool to sell companies services they didn't need or at rates that were far too high.
When you finally got the ball rolling so many IT companies had a direct profit motive in continuing the scare propaganda that it was blown out of proportion beyond the actual size of the problem. In retrospect I realized how many doomsday cult nutters got their say in the media, how many IT executives, tech pundits and others with all graveness of a funeral proclaimed this was the biggest crisis in computing ever and so on. You got the mainstream press to scare the average person in the street with what could happen, not just CEOs.
I'm not saying it wasn't real. It was big, but the crisis built around it was even bigger.
Except they were not stupid and they were not dumb. You look at your megabytes and gigabytes of RAM and think of course that's stupid. But a current era machine would be something like the Apple II with 4 kB - 4096 bytes - of RAM, where it really, really matter if an IP address takes up 4 bytes or 8 bytes. Or if you use 2 or 4 digits to store the year. By the time TCP/IP became official, cutting edge machines like the IBM PC and Spectrum Z80 had 16 kB.
You must remember that TCP/IP was designed only around the time people started to imagine the possibility of a personal computer, and even then it was for the few and rich. That we'd all get together in one big network was even further out, I used to dial BBS for many years before I got on the Internet, even though it already existed as such.
Even today when there's far more people and people are much richer than 30 years ago there's only about 2 billion people on the Internet, even if you assumed a PC for everyone we'd still be good for another while. But I have a PC at home and at work and in my pocket and it all adds up. But who had that crystal ball in the late 70s/early 80s and what if they did?
Sure, you could have just picked some impossibly huge number that'd obviously be enough for everything. But it would have had a huge and immediate impact on memory consumption and cost there and then. We're not talking about short sighted businessmen that only care about the next quarter here. We're talking about things that could only be a problem decades down the road if this becomes a megahit. Sure it's shitty for us, but that's not their fault. Particularly when people have been waving the warning flags for years and everybody's happily ignored it until we hit the brick wall.
Nehalem to Sandy Bridge? Not necessary, particularly since you need another new mobo. Is Sandy Bridge better than Nehlem? Overall yes, but not radically so. It's more a replacement for the 1156 than the 1366 platform, that refresh is coming later this year with a different socket and different processors. You can tell by the pricing, I don't remember the 2100 but the 2500 and 2600 are at $200 and $300 respectively. The "balls to the walls" segment is coming later probably with CPU prices up to $999 for the extreme editions.
Yep, which makes it sort of odd.... none of the reviewers caught it, none of the early adopters seem to have caught it... yet it's so critical it justifies halting production and starting over with fresh silicon. Granted they don't test the controller that much but at least some of the file tests would. I bet they're all scrambling to find out now though and we'll know in a day or two.
There is nothing that says that the laws of the USA automatically apply to people in other countries.
No, but a contract (subsets are terms of service and EULAs) work both ways. There is a choice of law and it's either in my jurisdiction or in theirs, not surprisingly every company I've ever dealt with state it's in their own as part of the contract. That means I can't use the laws of my country, because it won't have jurisdiction. I not only can, but I must use foreign law to sue as long as the contract is under foreign law. Even if the behavior against me has been criminal, I must use their criminal law not my country's criminal law. If the US refused to hear cases from foreigners on matters related to US contracts, all foreign trade would disappear faster than you could snap your fingers.
More and less are opposites, so is multiplication and division. Most people take ten times more to mean x * 10 and ten times less x / 10. Neither is mathematically correct, but it has a certain logical consistency.
You're assuming too much. For example there could be a low mass, low pressure and high mass, high pressure solution that both satisfy the conditions. Or even plateus or microscopic dips at different pressure levels as water atoms align in different ways. It can not be algebraicly solved as this is real physics, not idealized mathematical physics and there is no algebraic formula for the volume of water this would produce. You'd have to find one kg value, recalibrate your pressure measurement device then iteratively approach the right value. Not a very practical method.
Unless we create a magic battery, power consumption will always be a huge thing for laptops and cell phones. Data centers too certainly measure performance/watt. But I agree, for the regular desktop it's no longer a big deal, if it ever was.
This does not only concern energy use: contrary to a road or a railroad track, a cargo ropeway can be built straight through nature without harming animal and plant life (or, potentially, straight through a city without harming human life).
Then scroll down to the big ugly modern cargo ropeways/conveyor belts in the bottom of the article and you can see they're ugly as fuck and can be seen for many miles around. Compared to that a road or railroad is almost invisible. They also generously ignore that we've gotten a lot better at building bridges and tunnels than before, not worse.
I suppose it makes sense if you have a huge, stable amount of materials moving point-to-point, but for the most part such a cargoway will only add another exchange point where goods must be unloaded and reloaded which costs time and money. Also there's very little flexibility, with trucks or trains you can run more or less and even sell parts of it if things are slow. With this you have almost only fixed costs and if you hit the capacity limit it's a very hard limit.
This reminds me a little of the people that try to revive the zeppelins, it's only going to work in some really niche cases and those places usually already have one.
I based my cost statement on the higher number on the basis that MoS2 semiconductors would increase the demand.
Maybe, but that also depends on economics of scale. If it's more of a specialty product today it might go down with volume, unless you run into resource limitations.
I imagine MoS2 based semiconductors would only be cost effective if they can figure out how to use as little of it as possible, perhaps with MoS2 over some other substrate.
Near as I can tell it's dirt cheap. I figure the cost will be the same as current processors, getting it to ultra-pure quality and the etching process. You can get a kilo of not-so-very-pure MoS2 for about a buck. Even silicon good enough to make solar cells costs $67 dollars a kilo according to this 2009 article. The rest is for turning it from a lump of metal to a working processor.
Our best cheapish consumer clocks are based on GPS and they only know their internal time in the range of about 50 to 90 nanoseconds.
But why would you need to keep track of time as such? If you send a ping through the cable and back all modems can calculate ticks from the central and count ticks until the next sync. I figure the clock that drives my 3+ GHz processor has sub-nanosecond accuracy and putting one of those in a fiber optic modem doesn't sound unreasonable. My back-of-the-napkin protocol design would be:
1) Modem listens for clock ping-pong, finds delay in ticks 2) The central will for each cycle reserve a few time slots for new devices 3) Modem calculates central time and sends introduction in one of these slots with delay from central and checksum, which can collide but normally won't. If so the checksum will fail and they'll try again next cycle. 4) Central assigns modem a device id and starts assigning time slots. Since they know the delay of each modem they can make sure they don't collide. If they get crosstalk because one device has clock drift they can increase buffers on both sides for that one device.
I figure even with 1000 hosts you can afford to do this often, like once a second. It'll still be billions of ticks.between each sync. I'm probably missing something here, because this doesn't seem like that hard a problem.
You really should get that humor detector fixed. The point was that if you say you weigh 100 kg that really means "I weigh 100 times as much as a block of platinum-iridium alloy in France". Since the block always weighs the same as itself, it is always 1 kg while you would gain weight because you now weigh 100.00001 times as much as that block. So the whole universe has as such gained weight. He just suggested that the universe changed and not the stick we measure it by, which would require a Star Trek moment where it effected everything else in the universe except that block. That does seem to happen a lot to the Enterprise, not so much in reality.
Think of it like this: If the technology was readily available and cheap, why would I not want watch streamed movies in UHDTV resolution on a 100 inch 3D display at 100 Hz refresh rate? A single such stream would require bandwidth in the range of Gbit/s.
Well, we had 1600x1200 monitors back in the 1990s then we ran into pretty much a full stop, except for a few exotic 30" monitors and some specialty 2160p monitors for medical/military purposes. Oh and we added widescreen but most of those lost 120 pixels of horizontal resolution as well so 1920x1080 ~= 1600x1200.
I mean when I in 2011 can't get a better monitor than that, which I'm sure is due to lack of demand and not a technical limitation then what hope is there of UHDTV? Or even QHDTV, when it seems we can't see it staring at a monitor half a meter away? I predict to be an old, old man before UHDTV is common if ever.
What would you like to do with 100 times your own current network speed?
That would give me >2Gbit/s actual. I could stream what like 40 blurays simultaniously? Don't need it. Can't really imagine anyone who does, really. And I'd probably still be downloading from torrents because the TV/movie execs won't offer it here, no netflix, no hulu, no TV shows or movies on iTunes.
And for most things like series I follow my computer could just download it encrypted the night before in maximum quality, then deliver the key at release time. Bandwidth is really not a problem, at least the pirates seem able to deliver so it's strange if a big company couldn't. Sure I'd still take more if I could but it's no longer a bit deal. Before this is I had 2 Mbit down and that was horrible.
Yes, you generally only give your most trusted men the keys to the kingdom. But it doesn't mean it never, ever happens. Of course you can expect major chaos, backdoors, deleted data but it's nice if not everything goes up in flames. I'd say there's two things you need:
1) A backup system the admin doesn't have access to 2) A plan for a clean rebuild/restore of the core systems. 3) Don't tell him that's why you're doing it...
The backup can pretty much be explained by wanting to have an offsite backup with someone specializing in that, it's not core activity for you so you outsource it.
The plan for rebuild/restore could be part of some disaster recovery plan or something. "In case our data center goes *poof*, what would we need to start over on fresh hardware?
And if you're the religious type, you pray pretty damn hard you'll never need it.
Somehow I suspect a galaxy to be more a paradox of the heap kind of problem. A huge bunch of stars is a galaxy, remove one by one and at some point it stops being a galaxy. The question is what star turns a galaxy into a non-galaxy, I don't think there's a "fundamental property" to that.
Slashdot Mirror
Ogg Theora was pretty much a repeat of Ogg Vorbis, except that Theora is actually worse than the format it tries to replace and video is a much bigger patent minefield. Same open source fanatics, same arguments, very little corporate support, very little hardware support but somehow everything would be different this time around.
Hell, even the ogg container completely failed to compete with mkv. Just about the only thing coming from that direction I've seen in use is FLAC, which is apparently a quite good lossless format (size and complexity, lossless is of course lossless).
I think even Google will find that dethroning H.264 is a huge undertaking, if they didn't own YouTube I wouldn't think they stood a chance. If they can offer a superior WebM experience there, then maybe. But I'm still not going further than maybe.
It's really much simpler. Windows 7 and OS X has already licensed the codec, Microsoft has absolutely nothing to lose by pushing it. Firefox has problems with it, Linux has problems with it. When there's so few competitors, pushing them down is as good as lifting yourself up. Not to mention in public perception they don't want it to look like Google is leading the pack and Microsoft tagging along. There's so many political and strategical reasons to do it that far outweigh the minimal patent royalties they get.
Quoting WP:
A footnote in Mozilla's 2006 financial report states "Mozilla has a contract with a search engine provider for royalties. The contract originally expired in November 2006, however Google renewed the contract until November 2008 and has now renewed the contract through 2011.[8] Approximately 85% of Mozillaâ(TM)s revenue for 2006 was derived from this contract."
The financial FAQ dated November 18, 2010 says:
What is the status of the organization's contract with Google?
We have had a productive relationship with Google since 2004 and that relationship remains healthy. To date, we have renewed our contract three times, in 2005, 2006 and 2008. The current version extends through 2011.
So through 2011 Mozilla has a very good deal. But then Google didn't have a browser of their own and desperately needed Mozilla to break the IE monopoly. I suspect that these negotiations will go quite differently. I'm sure the deal will be extended but I doubt the terms will be anywhere near as favorable as they have been. Google has seen how easily they can now push their own browser into the market, they don't "need" Firefox that much anymore. And from a strict business point of view, where would they go? Bing? Yeah, I'm sure the open source community would love Microsoft as their default search engine. Not to mention that currently Chrome has targeted the IE holdouts. If they go their separate ways, Google will do their best to win Firefox users too. I'd put good money on the browser market looking completely changed in 2-3 years.
I think I remember seeing a fake dual vinyl player built just for this purpose in some magazine. Basically you hooked it up to the computer and controlled songs as if they played on vinyl, the driver also added suitable scratching if you wanted. Never mind all the other neat things you can do with the computer, it seemed like a superior solution. True, it's niche but so is doing that in the first place...
Sheldon isn't more of a living embodiment of OCD than he's a geek. On a TV show it makes for great comedy, in real life it'd drive you c-r-a-z-y. I have a buddy that's 1/10th of Sheldon and that's plenty...
I doubt that the distribution of geeks IQs is much different than that of most others. Unless you're saying geeks are defined by their IQ.
Most of the geeks I know, myself inclusive, have (...) reasonable analytical skills, and by associating ourselves with that sort of stuff.
Hint: That contributes a lot to IQ scores.
This is a pretty good image from 2008 onwards at least. The predicted time has varied by a few months but the money has been on 2011 at least for the last three years.
Y2K was perfectly legitimate. It was only through heroic efforts that programmers were able to overcome years of managerial negligence and get the changes made in a knick of time. As is typical, since the herculean effort caused nothing to happen the world yawned and assumed the geeks were just moaning over nothing all along.
I'm sure that for many systems that was true. But to be honest, it passed with such a complete yawn that not only were all the important systems fixed, it seems all the "nice to have" systems were fixed as well. I don't think it's any exaggeration to say that many people and companies used y2k as a tool to sell companies services they didn't need or at rates that were far too high.
When you finally got the ball rolling so many IT companies had a direct profit motive in continuing the scare propaganda that it was blown out of proportion beyond the actual size of the problem. In retrospect I realized how many doomsday cult nutters got their say in the media, how many IT executives, tech pundits and others with all graveness of a funeral proclaimed this was the biggest crisis in computing ever and so on. You got the mainstream press to scare the average person in the street with what could happen, not just CEOs.
I'm not saying it wasn't real. It was big, but the crisis built around it was even bigger.
Except they were not stupid and they were not dumb. You look at your megabytes and gigabytes of RAM and think of course that's stupid. But a current era machine would be something like the Apple II with 4 kB - 4096 bytes - of RAM, where it really, really matter if an IP address takes up 4 bytes or 8 bytes. Or if you use 2 or 4 digits to store the year. By the time TCP/IP became official, cutting edge machines like the IBM PC and Spectrum Z80 had 16 kB.
You must remember that TCP/IP was designed only around the time people started to imagine the possibility of a personal computer, and even then it was for the few and rich. That we'd all get together in one big network was even further out, I used to dial BBS for many years before I got on the Internet, even though it already existed as such.
Even today when there's far more people and people are much richer than 30 years ago there's only about 2 billion people on the Internet, even if you assumed a PC for everyone we'd still be good for another while. But I have a PC at home and at work and in my pocket and it all adds up. But who had that crystal ball in the late 70s/early 80s and what if they did?
Sure, you could have just picked some impossibly huge number that'd obviously be enough for everything. But it would have had a huge and immediate impact on memory consumption and cost there and then. We're not talking about short sighted businessmen that only care about the next quarter here. We're talking about things that could only be a problem decades down the road if this becomes a megahit. Sure it's shitty for us, but that's not their fault. Particularly when people have been waving the warning flags for years and everybody's happily ignored it until we hit the brick wall.
Nehalem to Sandy Bridge? Not necessary, particularly since you need another new mobo. Is Sandy Bridge better than Nehlem? Overall yes, but not radically so. It's more a replacement for the 1156 than the 1366 platform, that refresh is coming later this year with a different socket and different processors. You can tell by the pricing, I don't remember the 2100 but the 2500 and 2600 are at $200 and $300 respectively. The "balls to the walls" segment is coming later probably with CPU prices up to $999 for the extreme editions.
Yep, which makes it sort of odd.... none of the reviewers caught it, none of the early adopters seem to have caught it... yet it's so critical it justifies halting production and starting over with fresh silicon. Granted they don't test the controller that much but at least some of the file tests would. I bet they're all scrambling to find out now though and we'll know in a day or two.
There is nothing that says that the laws of the USA automatically apply to people in other countries.
No, but a contract (subsets are terms of service and EULAs) work both ways. There is a choice of law and it's either in my jurisdiction or in theirs, not surprisingly every company I've ever dealt with state it's in their own as part of the contract. That means I can't use the laws of my country, because it won't have jurisdiction. I not only can, but I must use foreign law to sue as long as the contract is under foreign law. Even if the behavior against me has been criminal, I must use their criminal law not my country's criminal law. If the US refused to hear cases from foreigners on matters related to US contracts, all foreign trade would disappear faster than you could snap your fingers.
More and less are opposites, so is multiplication and division. Most people take ten times more to mean x * 10 and ten times less x / 10. Neither is mathematically correct, but it has a certain logical consistency.
You're assuming too much. For example there could be a low mass, low pressure and high mass, high pressure solution that both satisfy the conditions. Or even plateus or microscopic dips at different pressure levels as water atoms align in different ways. It can not be algebraicly solved as this is real physics, not idealized mathematical physics and there is no algebraic formula for the volume of water this would produce. You'd have to find one kg value, recalibrate your pressure measurement device then iteratively approach the right value. Not a very practical method.
Unless we create a magic battery, power consumption will always be a huge thing for laptops and cell phones. Data centers too certainly measure performance/watt. But I agree, for the regular desktop it's no longer a big deal, if it ever was.
This does not only concern energy use: contrary to a road or a railroad track, a cargo ropeway can be built straight through nature without harming animal and plant life (or, potentially, straight through a city without harming human life).
Then scroll down to the big ugly modern cargo ropeways/conveyor belts in the bottom of the article and you can see they're ugly as fuck and can be seen for many miles around. Compared to that a road or railroad is almost invisible. They also generously ignore that we've gotten a lot better at building bridges and tunnels than before, not worse.
I suppose it makes sense if you have a huge, stable amount of materials moving point-to-point, but for the most part such a cargoway will only add another exchange point where goods must be unloaded and reloaded which costs time and money. Also there's very little flexibility, with trucks or trains you can run more or less and even sell parts of it if things are slow. With this you have almost only fixed costs and if you hit the capacity limit it's a very hard limit.
This reminds me a little of the people that try to revive the zeppelins, it's only going to work in some really niche cases and those places usually already have one.
I based my cost statement on the higher number on the basis that MoS2 semiconductors would increase the demand.
Maybe, but that also depends on economics of scale. If it's more of a specialty product today it might go down with volume, unless you run into resource limitations.
I imagine MoS2 based semiconductors would only be cost effective if they can figure out how to use as little of it as possible, perhaps with MoS2 over some other substrate.
Near as I can tell it's dirt cheap. I figure the cost will be the same as current processors, getting it to ultra-pure quality and the etching process. You can get a kilo of not-so-very-pure MoS2 for about a buck. Even silicon good enough to make solar cells costs $67 dollars a kilo according to this 2009 article. The rest is for turning it from a lump of metal to a working processor.
Our best cheapish consumer clocks are based on GPS and they only know their internal time in the range of about 50 to 90 nanoseconds.
But why would you need to keep track of time as such? If you send a ping through the cable and back all modems can calculate ticks from the central and count ticks until the next sync. I figure the clock that drives my 3+ GHz processor has sub-nanosecond accuracy and putting one of those in a fiber optic modem doesn't sound unreasonable. My back-of-the-napkin protocol design would be:
1) Modem listens for clock ping-pong, finds delay in ticks
2) The central will for each cycle reserve a few time slots for new devices
3) Modem calculates central time and sends introduction in one of these slots with delay from central and checksum, which can collide but normally won't. If so the checksum will fail and they'll try again next cycle.
4) Central assigns modem a device id and starts assigning time slots. Since they know the delay of each modem they can make sure they don't collide. If they get crosstalk because one device has clock drift they can increase buffers on both sides for that one device.
I figure even with 1000 hosts you can afford to do this often, like once a second. It'll still be billions of ticks.between each sync. I'm probably missing something here, because this doesn't seem like that hard a problem.
1 pascal (Pa) = 1 N/m^2 = 1 kg/(m*s^2). So you can't define 101.325 kPa without first defining the kilogram.
You really should get that humor detector fixed. The point was that if you say you weigh 100 kg that really means "I weigh 100 times as much as a block of platinum-iridium alloy in France". Since the block always weighs the same as itself, it is always 1 kg while you would gain weight because you now weigh 100.00001 times as much as that block. So the whole universe has as such gained weight. He just suggested that the universe changed and not the stick we measure it by, which would require a Star Trek moment where it effected everything else in the universe except that block. That does seem to happen a lot to the Enterprise, not so much in reality.
Think of it like this: If the technology was readily available and cheap, why would I not want watch streamed movies in UHDTV resolution on a 100 inch 3D display at 100 Hz refresh rate? A single such stream would require bandwidth in the range of Gbit/s.
Well, we had 1600x1200 monitors back in the 1990s then we ran into pretty much a full stop, except for a few exotic 30" monitors and some specialty 2160p monitors for medical/military purposes. Oh and we added widescreen but most of those lost 120 pixels of horizontal resolution as well so 1920x1080 ~= 1600x1200.
I mean when I in 2011 can't get a better monitor than that, which I'm sure is due to lack of demand and not a technical limitation then what hope is there of UHDTV? Or even QHDTV, when it seems we can't see it staring at a monitor half a meter away? I predict to be an old, old man before UHDTV is common if ever.
What would you like to do with 100 times your own current network speed?
That would give me >2Gbit/s actual. I could stream what like 40 blurays simultaniously? Don't need it. Can't really imagine anyone who does, really. And I'd probably still be downloading from torrents because the TV/movie execs won't offer it here, no netflix, no hulu, no TV shows or movies on iTunes.
And for most things like series I follow my computer could just download it encrypted the night before in maximum quality, then deliver the key at release time. Bandwidth is really not a problem, at least the pirates seem able to deliver so it's strange if a big company couldn't. Sure I'd still take more if I could but it's no longer a bit deal. Before this is I had 2 Mbit down and that was horrible.
Yes, you generally only give your most trusted men the keys to the kingdom. But it doesn't mean it never, ever happens. Of course you can expect major chaos, backdoors, deleted data but it's nice if not everything goes up in flames. I'd say there's two things you need:
1) A backup system the admin doesn't have access to
2) A plan for a clean rebuild/restore of the core systems.
3) Don't tell him that's why you're doing it...
The backup can pretty much be explained by wanting to have an offsite backup with someone specializing in that, it's not core activity for you so you outsource it.
The plan for rebuild/restore could be part of some disaster recovery plan or something. "In case our data center goes *poof*, what would we need to start over on fresh hardware?
And if you're the religious type, you pray pretty damn hard you'll never need it.
Somehow I suspect a galaxy to be more a paradox of the heap kind of problem. A huge bunch of stars is a galaxy, remove one by one and at some point it stops being a galaxy. The question is what star turns a galaxy into a non-galaxy, I don't think there's a "fundamental property" to that.