Kid, I work with RF every day, troubleshooting all manner of wireless systems.
You can go on believing what ever you want. Until you understand it yourself, you are going to stand by whatever the Gods tell you. It's people like you that make troubleshooting problems such a pain in the ass as you consistently find things which are only superficially related, and then blame the world's problems on them, just before throwing so much cure at the symptoms that it's amazing anything even fucking works by the time they're done spending money and adding extra parts.
And, the funny part is: When you're done, the problem is still there. I just got done wasting six months of my life fixing problems which were caused by people who act like you.
I simplify wiring and antenna systems all the time which have so many fucking extra parts and points of failure and (yes) bad grounding that it's amazing they ever worked at all. And it's folks like you who are the problem, with their seat-of-the-pants reactions to what they perceive to be a problem instead of any attempt at best practice, because they have no fucking idea what best practice even IS most of the time.
And, yeah, I know: If I'd only seen it, it'd have made all the difference. Yeah. Duude. The RF, man, it's so like wild-looking, check it out! Did you see that? Whoah. Got any Cheetos left over there?
Now, then: You stick to your theories, telco guy. I'll stick to my spectrum analyzer.
No. I didn't miss a thing. I just failed to make any direct mention of the parabolic reflector next door, because I could not conceive of any way to do so without declaring you to be incompetent.
Just as the electricians should stick to electricity, and the networking guys should stick to networking, I figure you telco folks should stick to telco, and leave the RF stuff to people who are trained in the art.
You talk about grounding like folks used to talk about SCSI termination[1].
The reality is far simpler, and there's no goat's blood required: Use shielded wire where it makes a difference. Ground it at one end (or in the middle if that's what the situation dictates, but under no circumstances at more than one point).
And then, have a beer. While imbibing, rejoice in the fact that one's cell phone has ceased ringing with voices proclaiming connection issues.
[1]: SCSI buses should be terminated exactly twice; once at each end of the bus. If you have more than two terminators, they are in the middle of the bus, or your "bus" has more than two ends, you've done it all wrong and need to start over. Same with TV coax, radio coax, 10base2, or any other transmission line medium.
Grounding both ends of a signal-carrying wire in which the signal itself does not require a ground reference is a Bad Idea.
There are signals which require ground. RS-232, for example. And in those cases, there's a myriad of implements in existence which serve only to break that ground (opto isolators, "short-haul" modems, transformers for audio or video, etc). Or, even better: Use a signaling format which doesn't require ground, like RS-422.
The rest of the time? Ground one end if the wire is shielded, and none if it is not (and if the particular signaling interface doesn't require ground).
Because, simply, at the end of the day:
Would you rather have lightning shunted to ground via the flimsy shield and drain wire on some UTP Cat5e, or via the substantially-heftier 12 AWG wire in the electrical socket?
(As someone who has seen holes burned between the shield and the conductors of shielded wire, as well as the outer jacket into open space, I think the answer should be obvious.)
Sure, it's a little like an antenna, insofar as a Faraday cage is.
But grounding the shield at both ends creates ground loops. You might not notice them right away, but you sure will the first time the MOVs in the surge suppressors at one end or the other shunt a spike to ground, and some of that current decides that its preferred path to ground is over your STP Cat5.
Eventually, after you blow up enough switch ports, you'll stop doing it that way.
It's generally pretty bad form to ground both ends of any shielded wire that traverses any real length.
I even run UTP up radio towers for stuff like IP cameras. The stuff works fine, even in a fucking sea of RFI, with 40 to 100 Watt transmitting antennas blasting it from a few feet away.
I think it is, by sheer virtue of the absurdity of the whole thing.
Parody doesn't have to be funny. It doesn't have to be good. It doesn't have to be insulting.
The man simply made a good parody of a good Queen song. Not a remix, not a cover, not a flailing attempt to repetitiously exploit a riff or a bassline. Just a parody.
Oh, yeah - I glossed over the UPS issue. Don't even think about getting Google-esque integrated UPS at the moment; that particular design is highly encumbered by a very thorough patent, and (AFAIK) nobody yet offers anything like it on the open market.
I, myself, have a rack-mount Tripp-Lite UPS at home. Fits in 2 racks spaces. Not absurdly heavy. Full-time inversion to sine wave, for complete isolation from whatever nastiness exists on the power lines. There are other, similar things available from all of the usual suspects in UPS world. A rackmount UPS could fit into the same rack as your "server," or (perhaps ideally) be bolted into its own case to make moving things around a little less energetic.
Just pick yourself up a TRS-80 Model 100, and be done. It only uses about one Watt of power.
Or maybe an RS-232-equipped teletype; I have a TI Silent 700 here that uses thermal paper. It works fine after rescuing it from the hot equipment cabinet where it lived, powered up, for almost exactly twenty years. (Don't ask - you can't afford mine.)
More seriously: Perhaps looking to pro audio for the answer might be productive. You want physical durability, a fair bit of storage, and tolerance for hot environments, and dust. (I live in Ohio, where the humidity is rather high all summer, and I don't consider that to be a real concern since my stuff all seems to survive just fine.)
Nearly zero digging finds me this Ebay offering. I have no idea who really makes this case, but I've been seeing that particular design around for at least 15 years.
The important parts: Rack mount, lots of drive bays, ATX, dust filters, and (here's the clincher) hold-downs for the PCI cards.
Build the rest of the system as you see fit with whatever tradeoffs you want, and avoid copper CPU heatsinks due to the weight (and therefore stress) of them.
And then slide it all into a shock-mounted roadcase. Both Starcase and Anvil are capable of producing a shock-resistant, ATA-rated road case which will fit neatly with the aforementioned ATX case. Also, there is SKB. They don't do anywhere near as much custom work, and their stuff isn't (in my opinion) quite as durable as the others, but they're generally a lot cheaper and lighter, and they list several 4U shock-mounted racks as standard items here.
We're straying pretty far off topic here, I think, but...
Background: I've been married for a little more than five years. My wife is fixed. I have never, ever, had sex with this woman with any sort of protection. And I've never had anyone better.
That said, I was at Wal-Mart the other day, and I happened upon the condom section. I began reminiscing about how much fun it used to be to try out different condoms on different people.
I've thought about it a lot since then. It turns out that I really enjoy fucking without a condom, but I really also enjoyed the variety that they offered. And, sure, they dull (or at least modify) physical sensation, but that makes the experience more emotional and less about tactile sensation.
I didn't buy any that day. And the reason wasn't because I'd rather have sex without one, but was more about my wife's potential reaction to the idea that we use them -- I didn't like the negative ways in which the story line was playing out in my head.
But it's tradeoffs, I think. Not better one way or the other, really. Variety can be good -- especially if it means that you still get woken up with breakfast in bed every now and then.
I have a Dell laptop. 1.83GHz Pentium M, dedicated ATI video, biggest hard drive I can get in PATA, which unfortunately is not the most energy-conserving. Fancy WUXGA+ (1920x1200) 15.4" screen. 2 gigs of fast RAM. 802.11a/b/g. Bluetooth. DVD burner. It's old, now, but by the standards of the day it was a fucking hog -- more "desktop replacement" than "portable computer."
I get -huge- improvements in battery life by slowing the CPU. I can even tweak (with Rightmark's fantastic rmclock program) precisely under what circumstances the CPU will slow down and speed up. Further improvement is made by undervolting the CPU (again with rmclock) to use as little current as possible while maintaining stability.
And then, I use i8kfangui to modulate the system's fan based on parameters of my own choosing. By default, the laptop keeps the fan spinning all the time, and varies the speed from slow to gale based on CPU temperature. With i8kfangui, it is almost always off, unless I'm doing heavy computation or graphics with the machine, and even then it is only on intermittently at slow speed. This saves even more power, and I can measure a sizable gain in battery life by just not spinning the fan as much and letting convective cooling do a little more work.
Now, granted, it's a laptop, not a desktop with a big, inefficient power supply with cooling fans. I'd guess that, on a common desktop computer, the lousy efficiency of the power supply swamps any gains which can be made through these simple and effective measures.
It goes a few steps further than that, too: The sound chip is configured to turn off when not in use for a short period. The video card clocks down when on battery. I get about an extra hour of battery life out of this machine with these various tweaks.
But it's not so different from desktop, really: I still have a hard drive, and a display, and a cooling system. The CPU is little more than a riced-up, lower-voltage Pentium III. I still have peripherals to power.
On my main desktop machine at home (a quad-core Q6600) I keep Speedstep enabled because it's quieter that way: My motherboard modulates fan speeds based on various temperatures around the system, and the cooler I keep the CPU, the less noise I hear. And I don't need as much air conditioning in the summer. And I don't need to clean the cat hair out of the heatsinks as often. And the bearings in the fans last longer. Even the dual SLI 9800GTs in this box ramp up and down their fan speed based on graphics use - I have no idea what happens behind the scenes with clocks and such, but I only ever hear their fans when I'm running something like Crysis.
So, in summary: Just because you don't see any real improvement on your particular system with a flexibly-clocked system on a UPS Watt meter of unknown accuracy, doesn't mean that nobody ever will.
And further: Is there anything lost by enabling Speedstep (or the AMD equivalent)? If the answer is no, then there is only gain -- even if it's hard to spot.
I don't get bonuses for this stuff, but: Are you suggesting that I should take no pride at all in keeping a mail server up for over two years?
I even lubed the bearings in the CPU fan without interrupting anything.
It's only been down thrice in five years: Once, for an extended power outage following a tornado. Another because of a kernel exploit that needed patching. And lastly, when RAM got cheap like peanuts, I shut it down to stuff a bunch more in there.
I'm generally opposed to this sort of thing. The reason is obvious: I use a lot of bandwidth compared to most folks, and wish to pay as little as possible.
However, you raise an interesting point: "Lower off-peak rates." If I can only download torrents between 2 and 6AM without it costing me an arm and a leg, then so be it.
You don't understand. Which is normal: You're about the sixtieth person I've had to correct on this issue.
In synopsis: you're wrong.
Here's why:
RAM that is sitting there holding stuff you might need, sometime (ala Superfetch) is just as ready to be utilized as RAM which is doing nothing at all. Superfetch is a read caching system, and any RAM it has in use for itself can be used by other programs IMMEDIATELY if they need it instead. Nothing has to wait buffers to get pushed out to disk, there's no longstanding delay. It just gets repurposed, and overwritten with other stuff. It doesn't need zeroed first. It's RAM, ie Random Access Memory, ferfuck'ssake.
In other words:
A system with a gigabyte of free RAM is a system with a gigabyte of RAM that it's failed to use. An optimized system does not have unused RAM.
Linux systems also eventually use all available RAM for caching. Your UID is low enough that you've probably even seen discussions of this "problem" in *nix years and years ago, and you should understand by now that it's not a problem at all, for all of the same reasons (listed above) that it's not a problem with Windows.
I don't have a machine which (natively?) reads 3.5" floppies anymore, but I do have an external 3.5" USB floppy drive hanging on the shelf, which always works fine (largely because it hasn't been sucking up dust for the past decade). *shrug*
I work on computers for people, sometimes, as a side project.
For the past few years, every single computer that I have to nuke and reinstall Windows on gets the following treatment:
1. Google Updater with Firefox, set up to be as automatic and out-of-sight as possible 2. Avast antivirus, set up to be as automatic and out-of-sight as possible 3. Windows Update set to always install every update, all by itself
I then set Firefox as the default browser, and get rid of most of the IE icons in the system. People take about 0.3 seconds to get used to Firefox, and are happy to hear that it will keep itself updated and reasonably free of unintentional badness.
Before I started doing these things, computers would come back to me pretty quickly after a clean reinstall of Windows, because they'd trash them in no time.
Now, it usually takes years before I see the same PC again. And it's not that the customer is mad at me for installing the EVIL GOOGLE UPDATER and don't want to give me any more business, it's just that their shit is STILL WORKING JUST FINE.
I run into customers from time to time at the grocery store or wherever, and always ask how their computer is doing. "Oh, it's been great since you last had it," is a typical response.
I use Google Updater myself. Of course I want the latest Firefox. And why not the newest Google Earth, too? I see no harm in this.
All of the data I've created on a computer in the past 20 years is readable by modern machines -- it's on 3.5" floppy. Stored properly, and read on a clean drive (NOT the one which has been sucking up dust for the past six years, otherwise unused), this stuff still works fine.
I've thrown almost all of of it away, though. That's the part you missed in your synopsis of media history: The human aspect.
Some of the stuff that I've tossed, I'd like to get back, but it's in a landfill somewhere.
Some of the sheepskin documents survive; but the unimportant ones (as determined by the people of the day) are mostly gone, having been discarded.
Another poster says that the tapes are helican scan, which does make it a little more difficult... But even then, armed only with the original heads and an educated guess of what the results should look like, it should be doable with far less than 2,000 pounds of additional gear. We don't need a bunch of fancy, twiddly, analog feedback sections with failing discrete components to keep things in check anymore, as this is a job better suited to a fast microcontroller and some software. The demodulation of the signal, once things are scanning right, can be done completely in software after a simple preamp and A/D stage.
Would it cost less? It'd certainly be cheaper to reinvent most of the wheel if they wanted to create a lot of these readers, but for one machine? Who knows...
Meanwhile, I'm just happy they've accomplished something.
Every try cut and pasting this LONG URL from e-mail to the browser if you're using a small monitor, i.e. laptop?
My 15.4" laptop monitor runs at 1920x1200, you insensitive clod. That URL fits just fine. The screen has enough pixels to fit about two of those URLs on one line. This paragraph doesn't even wrap on my monitor.
Slashdot Mirror
Books? Theories?
Kid, I work with RF every day, troubleshooting all manner of wireless systems.
You can go on believing what ever you want. Until you understand it yourself, you are going to stand by whatever the Gods tell you. It's people like you that make troubleshooting problems such a pain in the ass as you consistently find things which are only superficially related, and then blame the world's problems on them, just before throwing so much cure at the symptoms that it's amazing anything even fucking works by the time they're done spending money and adding extra parts.
And, the funny part is: When you're done, the problem is still there. I just got done wasting six months of my life fixing problems which were caused by people who act like you.
I simplify wiring and antenna systems all the time which have so many fucking extra parts and points of failure and (yes) bad grounding that it's amazing they ever worked at all. And it's folks like you who are the problem, with their seat-of-the-pants reactions to what they perceive to be a problem instead of any attempt at best practice, because they have no fucking idea what best practice even IS most of the time.
And, yeah, I know: If I'd only seen it, it'd have made all the difference. Yeah. Duude. The RF, man, it's so like wild-looking, check it out! Did you see that? Whoah. Got any Cheetos left over there?
Now, then: You stick to your theories, telco guy. I'll stick to my spectrum analyzer.
No. I didn't miss a thing. I just failed to make any direct mention of the parabolic reflector next door, because I could not conceive of any way to do so without declaring you to be incompetent.
Just as the electricians should stick to electricity, and the networking guys should stick to networking, I figure you telco folks should stick to telco, and leave the RF stuff to people who are trained in the art.
I was just trying to be nice.
You talk about grounding like folks used to talk about SCSI termination[1].
The reality is far simpler, and there's no goat's blood required: Use shielded wire where it makes a difference. Ground it at one end (or in the middle if that's what the situation dictates, but under no circumstances at more than one point).
And then, have a beer. While imbibing, rejoice in the fact that one's cell phone has ceased ringing with voices proclaiming connection issues.
[1]: SCSI buses should be terminated exactly twice; once at each end of the bus. If you have more than two terminators, they are in the middle of the bus, or your "bus" has more than two ends, you've done it all wrong and need to start over. Same with TV coax, radio coax, 10base2, or any other transmission line medium.
Either way, you learn:
Grounding both ends of a signal-carrying wire in which the signal itself does not require a ground reference is a Bad Idea.
There are signals which require ground. RS-232, for example. And in those cases, there's a myriad of implements in existence which serve only to break that ground (opto isolators, "short-haul" modems, transformers for audio or video, etc). Or, even better: Use a signaling format which doesn't require ground, like RS-422.
The rest of the time? Ground one end if the wire is shielded, and none if it is not (and if the particular signaling interface doesn't require ground).
Because, simply, at the end of the day:
Would you rather have lightning shunted to ground via the flimsy shield and drain wire on some UTP Cat5e, or via the substantially-heftier 12 AWG wire in the electrical socket?
(As someone who has seen holes burned between the shield and the conductors of shielded wire, as well as the outer jacket into open space, I think the answer should be obvious.)
*shrug*
Sure, it's a little like an antenna, insofar as a Faraday cage is.
But grounding the shield at both ends creates ground loops. You might not notice them right away, but you sure will the first time the MOVs in the surge suppressors at one end or the other shunt a spike to ground, and some of that current decides that its preferred path to ground is over your STP Cat5.
Eventually, after you blow up enough switch ports, you'll stop doing it that way.
It's generally pretty bad form to ground both ends of any shielded wire that traverses any real length.
Indeed.
I even run UTP up radio towers for stuff like IP cameras. The stuff works fine, even in a fucking sea of RFI, with 40 to 100 Watt transmitting antennas blasting it from a few feet away.
I think it is, by sheer virtue of the absurdity of the whole thing.
Parody doesn't have to be funny. It doesn't have to be good. It doesn't have to be insulting.
The man simply made a good parody of a good Queen song. Not a remix, not a cover, not a flailing attempt to repetitiously exploit a riff or a bassline. Just a parody.
IMHO.
One word: Parody.
That's almost as funny as an Irish man walking out of a bar.
Oh, yeah - I glossed over the UPS issue. Don't even think about getting Google-esque integrated UPS at the moment; that particular design is highly encumbered by a very thorough patent, and (AFAIK) nobody yet offers anything like it on the open market.
I, myself, have a rack-mount Tripp-Lite UPS at home. Fits in 2 racks spaces. Not absurdly heavy. Full-time inversion to sine wave, for complete isolation from whatever nastiness exists on the power lines. There are other, similar things available from all of the usual suspects in UPS world. A rackmount UPS could fit into the same rack as your "server," or (perhaps ideally) be bolted into its own case to make moving things around a little less energetic.
Ah! The truth comes out!
Just pick yourself up a TRS-80 Model 100, and be done. It only uses about one Watt of power.
Or maybe an RS-232-equipped teletype; I have a TI Silent 700 here that uses thermal paper. It works fine after rescuing it from the hot equipment cabinet where it lived, powered up, for almost exactly twenty years. (Don't ask - you can't afford mine.)
More seriously: Perhaps looking to pro audio for the answer might be productive. You want physical durability, a fair bit of storage, and tolerance for hot environments, and dust. (I live in Ohio, where the humidity is rather high all summer, and I don't consider that to be a real concern since my stuff all seems to survive just fine.)
Nearly zero digging finds me this Ebay offering. I have no idea who really makes this case, but I've been seeing that particular design around for at least 15 years.
The important parts: Rack mount, lots of drive bays, ATX, dust filters, and (here's the clincher) hold-downs for the PCI cards.
Build the rest of the system as you see fit with whatever tradeoffs you want, and avoid copper CPU heatsinks due to the weight (and therefore stress) of them.
And then slide it all into a shock-mounted roadcase. Both Starcase and Anvil are capable of producing a shock-resistant, ATA-rated road case which will fit neatly with the aforementioned ATX case. Also, there is SKB. They don't do anywhere near as much custom work, and their stuff isn't (in my opinion) quite as durable as the others, but they're generally a lot cheaper and lighter, and they list several 4U shock-mounted racks as standard items here.
Does any of this come close to the mark?
Not to be pedantic, but: How do you tear down a hole in the ground?
We're straying pretty far off topic here, I think, but...
Background: I've been married for a little more than five years. My wife is fixed. I have never, ever, had sex with this woman with any sort of protection. And I've never had anyone better.
That said, I was at Wal-Mart the other day, and I happened upon the condom section. I began reminiscing about how much fun it used to be to try out different condoms on different people.
I've thought about it a lot since then. It turns out that I really enjoy fucking without a condom, but I really also enjoyed the variety that they offered. And, sure, they dull (or at least modify) physical sensation, but that makes the experience more emotional and less about tactile sensation.
I didn't buy any that day. And the reason wasn't because I'd rather have sex without one, but was more about my wife's potential reaction to the idea that we use them -- I didn't like the negative ways in which the story line was playing out in my head.
But it's tradeoffs, I think. Not better one way or the other, really. Variety can be good -- especially if it means that you still get woken up with breakfast in bed every now and then.
I have a Dell laptop. 1.83GHz Pentium M, dedicated ATI video, biggest hard drive I can get in PATA, which unfortunately is not the most energy-conserving. Fancy WUXGA+ (1920x1200) 15.4" screen. 2 gigs of fast RAM. 802.11a/b/g. Bluetooth. DVD burner. It's old, now, but by the standards of the day it was a fucking hog -- more "desktop replacement" than "portable computer."
I get -huge- improvements in battery life by slowing the CPU. I can even tweak (with Rightmark's fantastic rmclock program) precisely under what circumstances the CPU will slow down and speed up. Further improvement is made by undervolting the CPU (again with rmclock) to use as little current as possible while maintaining stability.
And then, I use i8kfangui to modulate the system's fan based on parameters of my own choosing. By default, the laptop keeps the fan spinning all the time, and varies the speed from slow to gale based on CPU temperature. With i8kfangui, it is almost always off, unless I'm doing heavy computation or graphics with the machine, and even then it is only on intermittently at slow speed. This saves even more power, and I can measure a sizable gain in battery life by just not spinning the fan as much and letting convective cooling do a little more work.
Now, granted, it's a laptop, not a desktop with a big, inefficient power supply with cooling fans. I'd guess that, on a common desktop computer, the lousy efficiency of the power supply swamps any gains which can be made through these simple and effective measures.
It goes a few steps further than that, too: The sound chip is configured to turn off when not in use for a short period. The video card clocks down when on battery. I get about an extra hour of battery life out of this machine with these various tweaks.
But it's not so different from desktop, really: I still have a hard drive, and a display, and a cooling system. The CPU is little more than a riced-up, lower-voltage Pentium III. I still have peripherals to power.
On my main desktop machine at home (a quad-core Q6600) I keep Speedstep enabled because it's quieter that way: My motherboard modulates fan speeds based on various temperatures around the system, and the cooler I keep the CPU, the less noise I hear. And I don't need as much air conditioning in the summer. And I don't need to clean the cat hair out of the heatsinks as often. And the bearings in the fans last longer. Even the dual SLI 9800GTs in this box ramp up and down their fan speed based on graphics use - I have no idea what happens behind the scenes with clocks and such, but I only ever hear their fans when I'm running something like Crysis.
So, in summary: Just because you don't see any real improvement on your particular system with a flexibly-clocked system on a UPS Watt meter of unknown accuracy, doesn't mean that nobody ever will.
And further: Is there anything lost by enabling Speedstep (or the AMD equivalent)? If the answer is no, then there is only gain -- even if it's hard to spot.
I don't get bonuses for this stuff, but: Are you suggesting that I should take no pride at all in keeping a mail server up for over two years?
I even lubed the bearings in the CPU fan without interrupting anything.
It's only been down thrice in five years: Once, for an extended power outage following a tornado. Another because of a kernel exploit that needed patching. And lastly, when RAM got cheap like peanuts, I shut it down to stuff a bunch more in there.
I'm generally opposed to this sort of thing. The reason is obvious: I use a lot of bandwidth compared to most folks, and wish to pay as little as possible.
However, you raise an interesting point: "Lower off-peak rates." If I can only download torrents between 2 and 6AM without it costing me an arm and a leg, then so be it.
No, no. I'm not loving Superfetch - it's turned off on my systems (for reasons other than those being discussed).
But Superfetch is the topic, since that is what essentially ensures that Vista (and 7) have roughly zero free RAM at all times.
I recognize that Window's memory management sucks. But I also recognize that unused RAM is wasted RAM, which is my main point.
You don't understand. Which is normal: You're about the sixtieth person I've had to correct on this issue.
In synopsis: you're wrong.
Here's why:
RAM that is sitting there holding stuff you might need, sometime (ala Superfetch) is just as ready to be utilized as RAM which is doing nothing at all. Superfetch is a read caching system, and any RAM it has in use for itself can be used by other programs IMMEDIATELY if they need it instead. Nothing has to wait buffers to get pushed out to disk, there's no longstanding delay. It just gets repurposed, and overwritten with other stuff. It doesn't need zeroed first. It's RAM, ie Random Access Memory, ferfuck'ssake.
In other words:
A system with a gigabyte of free RAM is a system with a gigabyte of RAM that it's failed to use. An optimized system does not have unused RAM.
Linux systems also eventually use all available RAM for caching. Your UID is low enough that you've probably even seen discussions of this "problem" in *nix years and years ago, and you should understand by now that it's not a problem at all, for all of the same reasons (listed above) that it's not a problem with Windows.
a, b, c, d--
Riddles and nonsense. I've said my piece. Dispute it with facts, or move on.
Thanks!
I don't have a machine which (natively?) reads 3.5" floppies anymore, but I do have an external 3.5" USB floppy drive hanging on the shelf, which always works fine (largely because it hasn't been sucking up dust for the past decade). *shrug*
Why do you care so much more about my data than I do?
I work on computers for people, sometimes, as a side project.
For the past few years, every single computer that I have to nuke and reinstall Windows on gets the following treatment:
1. Google Updater with Firefox, set up to be as automatic and out-of-sight as possible
2. Avast antivirus, set up to be as automatic and out-of-sight as possible
3. Windows Update set to always install every update, all by itself
I then set Firefox as the default browser, and get rid of most of the IE icons in the system. People take about 0.3 seconds to get used to Firefox, and are happy to hear that it will keep itself updated and reasonably free of unintentional badness.
Before I started doing these things, computers would come back to me pretty quickly after a clean reinstall of Windows, because they'd trash them in no time.
Now, it usually takes years before I see the same PC again. And it's not that the customer is mad at me for installing the EVIL GOOGLE UPDATER and don't want to give me any more business, it's just that their shit is STILL WORKING JUST FINE.
I run into customers from time to time at the grocery store or wherever, and always ask how their computer is doing. "Oh, it's been great since you last had it," is a typical response.
I use Google Updater myself. Of course I want the latest Firefox. And why not the newest Google Earth, too? I see no harm in this.
All of the data I've created on a computer in the past 20 years is readable by modern machines -- it's on 3.5" floppy. Stored properly, and read on a clean drive (NOT the one which has been sucking up dust for the past six years, otherwise unused), this stuff still works fine.
I've thrown almost all of of it away, though. That's the part you missed in your synopsis of media history: The human aspect.
Some of the stuff that I've tossed, I'd like to get back, but it's in a landfill somewhere.
Some of the sheepskin documents survive; but the unimportant ones (as determined by the people of the day) are mostly gone, having been discarded.
Another poster says that the tapes are helican scan, which does make it a little more difficult... But even then, armed only with the original heads and an educated guess of what the results should look like, it should be doable with far less than 2,000 pounds of additional gear. We don't need a bunch of fancy, twiddly, analog feedback sections with failing discrete components to keep things in check anymore, as this is a job better suited to a fast microcontroller and some software. The demodulation of the signal, once things are scanning right, can be done completely in software after a simple preamp and A/D stage.
Would it cost less? It'd certainly be cheaper to reinvent most of the wheel if they wanted to create a lot of these readers, but for one machine? Who knows...
Meanwhile, I'm just happy they've accomplished something.
Every try cut and pasting this LONG URL from e-mail to the browser if you're using a small monitor, i.e. laptop?
My 15.4" laptop monitor runs at 1920x1200, you insensitive clod. That URL fits just fine. The screen has enough pixels to fit about two of those URLs on one line. This paragraph doesn't even wrap on my monitor.
*shrug*