Reliable data storage is possible, but at what price?
Paper is very easy to store. You use good stock and you throw it in a fire-resistant safe, preferably high off the ground and outside of buildings that are prone to fire. My parents keep their important papers in a safe on a high shelf in the steel shed in their back yard. It's been there for decades. The papers in there look nearly as good as the day they put them in - some of which have been there for nearly four decades (marriage certificate is what I'm specifically thinking of). The deed to the place has been in there for thirty years, and also looks just like new.
Data is hard to store. I'm an engineer by day, but I'm also a semi-professional photographer (meaning I make a few grand a year at it, not enough to be a full-time job, but enough to pay for some kickass gear for my pseudo-hobby). For the last five years, I've been shooting digital. When you have nearly 3/4 of a terabyte of images to store, maintenance becomes a real headache. You change disks in the RAID, you make enormous backups, you monitor for corruption. I've never lost an image... yet. I've come dangerously close, however. The last time a disk in the array failed, I discovered another one was having problems while the array was rebuilt on a new disk. It didn't fail until after things were complete, but I had two weddings I'd shot that day that hadn't been properly integrated into a real backup yet. I've had backups come up bad during attempted restores, including one on supposedly "archival" quality media where the dye layer had started to delaminate (and yes, it was stored in a cool, dark location with a constant humidity of about 40% - my basement). Basically, yes, you can build a reliable system, but it takes considerable effort to do archivist-style work.
Papers? Yup, still the way to go for the important stuff, and I'm only 29. For that matter, I also think they're easier to read, particularly when I'm sitting out on my deck. Sun really washes out the old LCD on the laptop. I'm actually having a tough time editing this because I can barely read through the glare. I also like printed copies for datasheets when I'm building stuff. I can mark pages with post-its, scribble notes to myself, and eventually the doc just falls open to the pages I frequently use.
It's like they introduce new "standards" for the sake of having new standards, not because they address any particular problem. I have bluetooth crap all over the place (scanners, headsets, printers, telemetry nodes), and I have no intention of changing it all over. Much like anything that takes a Sony memory stick, it's just another reason to avoid certain vendors and products.
Agreed - "random" is a batch of BS. I'll admit, I'm sometimes a little rough looking - a razor and I can be strangers for a couple weeks at a time. Usually it's jeans, t-shirt, and boots for flying attire if I'm not on business travel or interline standby. I tend to do a lot of one way tickets, and I don't tend to check a whole lot (particularly if traveling standby - the I usually ship my luggage home, since I never know what flight I'll actually get). I do, however, carry a large carry-on full of cameras, lenses, and electronic gear. Oh yeah, and I'm an airline employee with a ramp badge, and my employer has put me through a number of background checks. Still, every time, "Sir, you've been randomly selected..." Random, my ass. I don't look or act like the usual traveler, and you picked up on that. Just admit it.
Personally I'd rather drive anyway. There's nothing like a warm fall day, the open road, and a convertible to relax the nerves. I only fly when it's the only option left.
At least you're not my old roommate from college. He shares a name (a very common one) with a guy on one of the watch lists. He gets the *extra* special screening every time he sets foot in an airport.
Personally, I can't wait for the day when they find that warning sticker glue is somehow very mildly carcinogenic. Then we'll get warning stickers on warning stickers on warning stickers on...
I advocate we go down to a single warning sticker on everything - "Please know what the fuck you're doing, or else return this product."
Say it with me now: RAID is not a backup solution.
In my life, I've managed to blow two RAID arrays. The first was in our departmental webserver at work, where a fan ate through a bundle of drive power wires over the weekend, shorting +12 to +5 and really f@#$ing up the entire 9 disk RAID-5 array. Every drive controller board was dead. The better part of that day was when we found the backup group had kept all of our backups on the same DLT tape, because they fit so nicely. Too bad the drive ate the only backup tape when it was put in for restore... Wound up buying an identical drive on eBay, placing it on each disk, and pulling an image. With all that done, I got nine new drives and pushed the images back onto them, and recovered most of it...
The second time was due to a screwy driver upgrade on my desktop machine. Long story short, it mangled large disk transfers. Since I was running software RAID-1 at the time, it mangled both disks in identical ways. I had growing corruption across the array and didn't know it until too late...
That said, I do run RAID-1 at home as a short-term strategy to protect against individual disk failure. That doesn't take the place of my weekly full backup, however. I did cut out the incrementals every night, though. They don't buy me much for my particular style of usage - YMMV.
Jeez, buck up. I mean I know it has neurotoxic effects and it tends to build up in organic tissue over time, but considering all the stupid things a lot of us did with mercury as kids, the miniscule amount in one broken light bulb (4 mg, typically) isn't going to kill you... or probably even have measurable effects.
Actually there are dimmable CFLs. I have three of them in my kitchen over my table - they're some off-breed, but I see that Philips has recently started making name-brand dimmables as well.
I have two Samsung ML-1740s, and am thrilled with both of them. Cheap (one was $120 new, the other $100 when I purchased it later), advertises Linux support on the box and it actually does work like a charm. I eventually put one on my fileserver box, which makes a network printer, but my dad's is attached to his network via a $35 DLink network print server. Again, works like a charm. Not terribly fast, but it is cheap, reliable, very low power when idled, and does a nice job printing.
Oh hell, let's just call it a day and turn the WWW into the W3D - World Wide Word Doc. That's what many corporate intranets have become anyway - talk about incompatibility and complete lack of usefulness. I've spent hours and hours trying to convince people that a Wiki is a fairly good tool for collaboratively building documentation. My own group didn't need much prodding, since we're mostly a Unix and embedded group, but everyone else has been a challenge.
The argument from others was, "Why can't I just use Word? Why can't I make this text use blue *insert stupid looking font here*?" It's nearly impossible to convey to them that when it comes to documentation, it's about organization and content, and there's nothing that beats writing and publishing in essentially plain text. It forces the author to think more about clear, logical content, and in the end, allows it to be used on any platform and to be easily searched.
Flash is much the same way in my book. Most of the websites I've seen that use it extensively have poor content or organization and are trying to make up for it with whiz bang neato bits made with Flash. Need a menu? Guess what, there are great ways to either do that with pure HTML or a combination of DHTML and Javascript. Why do people feel obcessed to implement simple things in complex and incompatible ways?
Yours is the only proposal I've heard so far that makes any sense - what ECU designer in their right mind would want an OS in the way? A quick look through Google shows that WinCE makes claims of being an RTOS, but I can't seem to believe it, nor can I find actual specs on what exactly are their definitions of "real time". Quite frankly, I just don't believe it, nor would I personally trust any mission-critical, timing-critical functionality to an operating system that wasn't specifically written for it (think pSOS, etc). Personally, I wouldn't use an RTOS at all in an ECU (basically because there's just no need for it), and yes, I have tinkered a bit with engine computer code.
Like many of the other responses here, this might actually seem fairly reasonable to me. Depends on what the benefit is. If this means that FF can be smaller and faster by dropping a bunch of crappy legacy API calls in favor of better, newer ones, then so be it. If it makes changes to code that is frequently touched less error-prone, then good. If it's just because somebody thought the old calls were "ugly", then it's dumb and the idea should be trashed. I don't know, I'm not a Win32 programmer or FF dev and can't reasonably evaluate the changes.
Someone who is perfectly happy to sit on Win98 is probably perfectly happy to use FF 1.5, because it "just works". I used 2000 up until about eight months ago, because there was no compelling reason to upgrade. Then my wife's laptop met an unfortunate end, and I recycled the XP license onto my desktop. That's the only reason I upgraded, since 2000 is still actively being patched and supported with new drivers.
It's not necessarily a representative sample of the whole world, but just looking at the OS breakdown of poeple who've hit my website since the first of the month, AWStats shows me the following (and these exclude me, btw):
Windows XP 56647 74.5 % Windows NT 286 0.3 % Windows Me 897 1.1 % Windows Vista (Longhorn) 4 0 % Windows CE 10 0 % Windows 98 2163 2.8 % Windows 95 84 0.1 % Windows 2003 117 0.1 % Windows 2000 8084 10.6 %
Quite simply, because I use a radio for more than just the same music over and over... It's nice to be in the middle of BFE (or at least west-central British Columbia) and still be able to keep up on news from home, whatever NPR has on that day, etc. For being hundreds of miles north of the target area for Sirius (US lower 48), the darn thing worked very well. I probably only had it cut out once or twice on the whole two week trip.
I wouldn't live without satellite radio these days, because it works wherever I am 99.9% of my life. Will I touch HDR with a 10 foot pole? Not remotely likely. It's not sound quality that made me stop listening to AM/FM broadcast radio, it's the small coverage areas and the usual, rebundled crap content. About the only broadcast radio I listen to any more is my local NPR station for some programming not carried on the satellite NPR channels.
Hours getting a modem to work? Sheesh, but then again that's why I only did external modems. RS232 is so simple it's really damn hard for any manufacturer to screw it up. Internal modems always seemed like a kluge, except for the ones that literally had a UART connected to the bus, and then that was connected to the modem guts. Software-based modems? Don't get me started - total ugly kluge to save a few bucks. It's like anything else - if you buy a serious kluge, you're going to have trouble.
I probably build an average of a system every two months (friends, family, etc.) for other people. In the last six years, I haven't had any components that "just didn't like each other". Most of them go off without a hitch, and wind up being very easy to upgrade (future cost savings) because they all use robust, standard components, not some bizarre crap an OEM/VAR decided was a good idea. I find that most of the ways that VARs add "value" is really adding nonstandard crap that can't be fixed, supported, or upgraded later.
The savings come in when I can do selective upgrades. Upgrade my vid board because I want better framerates for a new game I just picked up? No problem. Need a dual core because of some new project? No problem - planned for that when I bought the motherboard. Want to upgrade the PS to an 80% efficient one, or an ultra-quiet one? No problem. Consider that I haven't upgraded my actual case in five years, nor my power supply in the last three, yet I've gone through several rounds of motherboards/procs. Each time I just upgrade pieces, I save over either having to a) do without or b) go buy a whole new box.
Bottom line is decide what fits your needs. If you know what you're doing, build a machine if you want. If you're willing to invest time in learning, build a machine if you want. If you have no clue and/or don't want to invest the time to learn, pay somebody else to do it for you. Ain't specialization a great economic concept?
Seriously, you might lose some power. Ethanol has only 60% of the energy density of gasoline. Assuming a 10% mix, that means you're losing about 4% of your net energy concentration. There's some evidence that not all of this is a loss, however, since ethanol burns somewhat faster, leading what amounts to really about a 2-3% drop.
As far as destroying engines, I highly doubt it. I used to live in Iowa, and you pretty much can't get gas without EthOH. I've owned several cars that ran their entire lives on ethanol-blended gas, and all of them were at nearly 200k miles before I sold them, usually still running fine with original engines and fuel systems. My current everyday car is a 1995 Honda that's pushing 250k, and ran blended gas for its first 200k miles. (I bought it at 100k, and know the previous owner and what they did with it.) Unfortunately, the last 50k have been on non-blended gas, and I've more problems with injector fouling than I ever did when pushing ethanol through the system. Difference in running on straight and blended fuel (it's blended here in the summer, and I get back to Iowa for a week or two at a time)? Net wash in perceptable MPG, and yes, I keep records on every tank.
I realize that my experience does not a peer-reviewed study make, so I won't say authoritatively that it does not. However, my experience is such that leads me to think that you have no idea what you're talking about, unless you can back it up with some sort of evidence, preferably real studies.
For those of us who would rather cut off our heads rather than deal with a cable company again, I'm perfectly happy to buy without giving a second thought to cablecard. I went to satellite TV about five years ago and will never go back.
I agree, common sense dictates a UPS. But the guy was beyond that, his filesystem was trashed, and he needed to fix it.
Quite frankly, arguing that ext2 is fine because you should have had a UPS is much like arguing that if you have airbags, it doesn't matter if the seatbelt in your car only works about 80% of the time. Both are making up for a non-robust system by throwing more crap at it. That only improves your odds, it doesn't improve the overall robustness. Eventually the complexity of system-on-system-on-system will bring you down. A UPS guards against power failures, but not against a failed power supply. So you throw more power supplies at it, but then what about a voltage regulator on a proc that suddenly dies? Where does it stop, this heaping on more crap to make up for an underlying flaw?
The best solutions are nearly always eerily elegant. You fix the underlying problem - that the filesystem isn't recoverable from one of the most likely things that can happen, an unclean shutdown. Journalling filesystems are one such elegant solution.
I agree that an attitude problem is common in IT, but I don't think it's limited. We have a lot of management with an equal problem - they think they're on top of the world, and run the company on their own. They fail to realize that most of them don't even have a technical background, and if the engineers walked out tomorrow in frustration, they'd take the company with them. Not only that, they fail to recognize that a lot of the intellectual capital the company has is locked up in a handful of very talented people, not spread equally across all the underlings. So firing Bill the Worthless and replacing him with an intern might be fine, but firing Ted the Chief Engineer, who designed half the stuff the company uses, might not be such a good idea because there's nobody who can fill his shoes.
Let's face it, no matter how awesome the management team is, it's the technical people that create the products that keep the company in business. Likewise, the technical staff usually needs effective guidance from management so that technology is targetted at business goals, not just firing off willy-nilly at whatever captures their interest today.
Definitely yes. X10 may suck, but it's way better than not having it.
A few examples:
- My current house has a very stupid wiring setup. Three of the bedrooms have a ceiling fan and light in the center of the room, which serves as the main light for the room. These were installed when the house was built. Each of these bedrooms has a light switch by the door. Switch doesn't connect to fan at all. In fact, they're on separate circuits on opposite phases - the switch runs an outlet which usually winds up behind the bed. Running new wire would involve a great deal of work (breaking out drywall, running cable, and then sealing the whole mess back up). Solution? Throw an X-10 transmitter into the socket (my personal choice being Keypadlincs), put two wire-in relay modules in the fixture box above the fan, do a little programming, and the whole thing just works.
- Since I currently have a 6-switch Keypadlinc in each bedroom (along with another 8 throughout the rest of the house), I can control more stuff. In particular, I find myself leaving a downstairs lamp on quite frequently. So I programmed it into one of the spare buttons. Likewise, my front outside house lights (small and tasteful) and the backyard light (affectionately known as the Death Ray or the Daybringer) are also connected in. If I hear a noise I want to check out, I just hit a button on the wall rather than having to go downstairs.
- Over Christmas, my outdoor Christmas lights were run via X-10. I could turn them on and off at will, or my firewall box would shut them off after a certain time of night, in case I forgot. Since I wanted to control them from the bedroom so I could kill them as I went to bed, I just temporarily reassigned one of the keypad buttons to run them.
- My wife and I both travel a lot and leave the house unattended, sometimes for a week or two at a time. Using the X-10 interface on my firewall box, I can log in, turn on the lights, check the security cameras scattered about the house, and then return the lights to their previous state. Also, I don't just have a timer on a lamp to make the house look occupied. I have my firewall box set up to create a log of when I turn lights on and off, since the whole house is under X-10 control. When I'm gone, I just tell it to play back the previous week's log. It's virtually indistinguishable whether I'm home or not, at least from the lights.
Damn, the people you run into on/. How ya doin', Steve? (It's Nathan)
It is possible to get X-10 to work reliably, it just takes some work, a little tinkering, and quality components. My house is now almost exclusively based on X-10, largely on Smarthome Keypadlincs. Whoever wired this place initially was a moron, and pulling new romex was way more trouble than it was worth. Light switches that don't control the lights, places where you have to wander into dark rooms to find the lightswitches, etc. Enter powerline carrier gear. I've found that as long as I stay away from the cheapo X-10 crap (the stuff largely marketed by x10.com) and stick with either Smarthome or Leviton-manufactured bits, things work fairly well. Oh, and a whole-house filter and active phase-to-phase repeater. I have one TV that soaks up signal and thus must be filtered at the outlet, but otherwise everything has "just worked" for about 18 months now.
Insteon looks like a quantum leap forward, but I haven't embraced it because it's a single-source system. Once 3-4 vendors make products, I might consider upgrading. The protocol is much, much, much more solid (acknowledgements, checksums, more data bytes), and I definitely wouldn't complain about better response times. Open source support is a deal-breaker, however. Like Spidey pointed out, Neil Cherry seems to not think this will be an issue based on his conversations, but I'm taking a very wait-and-see attitude. Even if SHM never officially supports it, if at least someone gets it working and they don't get sued, I'll consider it workable. Their closed attitude does rather concern me, however - exactly who do they think they're selling to if it's not the tinkerer market?
I'll point out that Smarthome was never really helpful about documenting the original X-10 USB Powerlinc protocol, either. I messed with it for a while by sniffing with the Windows driver, and then decided it was easier to stick with good ol' RS232. Eventually I just started using WiSH, and the open source community eventually worked out the details on the USB PLv1. (I continue to use serial, however)
Call me pragmatic, but I've always evaluated technology on its suitability to the task and rather it accomplishes my goals (price, support, stability, security, documentation, language du jour, whatever), unless I have specific moral objections to the supplier (CEO eats third-world children for desert, and buying said product would give him funds to import eight more children, etc.) Otherwise, if my component was written by Nepalese nudists but it was free, full-featured, stable, and fast, I'll take it. I don't care who wrote it, as long as it works for me.
Those who procure technology based on what color Armani suit the sales guy wore are going to wind up buying overpriced junk. Those who fairly judge technology on its merits and suitability to an application will succeed. Eventually group 1 will catch on once they see the example of group 2.
Slashdot Mirror
> TA-DA.. the COMPUTER!!! [Applause]
Awesomely fucking funny, awesomely fucking true.
Cue the cluebat on ProtectDisc.
Reliable data storage is possible, but at what price?
Paper is very easy to store. You use good stock and you throw it in a fire-resistant safe, preferably high off the ground and outside of buildings that are prone to fire. My parents keep their important papers in a safe on a high shelf in the steel shed in their back yard. It's been there for decades. The papers in there look nearly as good as the day they put them in - some of which have been there for nearly four decades (marriage certificate is what I'm specifically thinking of). The deed to the place has been in there for thirty years, and also looks just like new.
Data is hard to store. I'm an engineer by day, but I'm also a semi-professional photographer (meaning I make a few grand a year at it, not enough to be a full-time job, but enough to pay for some kickass gear for my pseudo-hobby). For the last five years, I've been shooting digital. When you have nearly 3/4 of a terabyte of images to store, maintenance becomes a real headache. You change disks in the RAID, you make enormous backups, you monitor for corruption. I've never lost an image... yet. I've come dangerously close, however. The last time a disk in the array failed, I discovered another one was having problems while the array was rebuilt on a new disk. It didn't fail until after things were complete, but I had two weddings I'd shot that day that hadn't been properly integrated into a real backup yet. I've had backups come up bad during attempted restores, including one on supposedly "archival" quality media where the dye layer had started to delaminate (and yes, it was stored in a cool, dark location with a constant humidity of about 40% - my basement). Basically, yes, you can build a reliable system, but it takes considerable effort to do archivist-style work.
Papers? Yup, still the way to go for the important stuff, and I'm only 29. For that matter, I also think they're easier to read, particularly when I'm sitting out on my deck. Sun really washes out the old LCD on the laptop. I'm actually having a tough time editing this because I can barely read through the glare. I also like printed copies for datasheets when I'm building stuff. I can mark pages with post-its, scribble notes to myself, and eventually the doc just falls open to the pages I frequently use.
It's like they introduce new "standards" for the sake of having new standards, not because they address any particular problem. I have bluetooth crap all over the place (scanners, headsets, printers, telemetry nodes), and I have no intention of changing it all over. Much like anything that takes a Sony memory stick, it's just another reason to avoid certain vendors and products.
>I figure you'd need Linux for that, right? Java too, probably?
:)
Well, if you use Java, that's probably bordering on underpowered for either of your suggested applications.
Agreed - "random" is a batch of BS. I'll admit, I'm sometimes a little rough looking - a razor and I can be strangers for a couple weeks at a time. Usually it's jeans, t-shirt, and boots for flying attire if I'm not on business travel or interline standby. I tend to do a lot of one way tickets, and I don't tend to check a whole lot (particularly if traveling standby - the I usually ship my luggage home, since I never know what flight I'll actually get). I do, however, carry a large carry-on full of cameras, lenses, and electronic gear. Oh yeah, and I'm an airline employee with a ramp badge, and my employer has put me through a number of background checks. Still, every time, "Sir, you've been randomly selected..." Random, my ass. I don't look or act like the usual traveler, and you picked up on that. Just admit it.
Personally I'd rather drive anyway. There's nothing like a warm fall day, the open road, and a convertible to relax the nerves. I only fly when it's the only option left.
At least you're not my old roommate from college. He shares a name (a very common one) with a guy on one of the watch lists. He gets the *extra* special screening every time he sets foot in an airport.
Personally, I can't wait for the day when they find that warning sticker glue is somehow very mildly carcinogenic. Then we'll get warning stickers on warning stickers on warning stickers on...
I advocate we go down to a single warning sticker on everything - "Please know what the fuck you're doing, or else return this product."
Say it with me now: RAID is not a backup solution.
In my life, I've managed to blow two RAID arrays. The first was in our departmental webserver at work, where a fan ate through a bundle of drive power wires over the weekend, shorting +12 to +5 and really f@#$ing up the entire 9 disk RAID-5 array. Every drive controller board was dead. The better part of that day was when we found the backup group had kept all of our backups on the same DLT tape, because they fit so nicely. Too bad the drive ate the only backup tape when it was put in for restore... Wound up buying an identical drive on eBay, placing it on each disk, and pulling an image. With all that done, I got nine new drives and pushed the images back onto them, and recovered most of it...
The second time was due to a screwy driver upgrade on my desktop machine. Long story short, it mangled large disk transfers. Since I was running software RAID-1 at the time, it mangled both disks in identical ways. I had growing corruption across the array and didn't know it until too late...
That said, I do run RAID-1 at home as a short-term strategy to protect against individual disk failure. That doesn't take the place of my weekly full backup, however. I did cut out the incrementals every night, though. They don't buy me much for my particular style of usage - YMMV.
Jeez, buck up. I mean I know it has neurotoxic effects and it tends to build up in organic tissue over time, but considering all the stupid things a lot of us did with mercury as kids, the miniscule amount in one broken light bulb (4 mg, typically) isn't going to kill you... or probably even have measurable effects.
There are high power factor CFLs (>0.9pf), they're just not as common because they cost a bit more.
Actually there are dimmable CFLs. I have three of them in my kitchen over my table - they're some off-breed, but I see that Philips has recently started making name-brand dimmables as well.
>Is everybody here retarded?
/., aren't I?
Ummm, I think the answer is rather obvious... I'm posting on
I have two Samsung ML-1740s, and am thrilled with both of them. Cheap (one was $120 new, the other $100 when I purchased it later), advertises Linux support on the box and it actually does work like a charm. I eventually put one on my fileserver box, which makes a network printer, but my dad's is attached to his network via a $35 DLink network print server. Again, works like a charm. Not terribly fast, but it is cheap, reliable, very low power when idled, and does a nice job printing.
Oh hell, let's just call it a day and turn the WWW into the W3D - World Wide Word Doc. That's what many corporate intranets have become anyway - talk about incompatibility and complete lack of usefulness. I've spent hours and hours trying to convince people that a Wiki is a fairly good tool for collaboratively building documentation. My own group didn't need much prodding, since we're mostly a Unix and embedded group, but everyone else has been a challenge.
The argument from others was, "Why can't I just use Word? Why can't I make this text use blue *insert stupid looking font here*?" It's nearly impossible to convey to them that when it comes to documentation, it's about organization and content, and there's nothing that beats writing and publishing in essentially plain text. It forces the author to think more about clear, logical content, and in the end, allows it to be used on any platform and to be easily searched.
Flash is much the same way in my book. Most of the websites I've seen that use it extensively have poor content or organization and are trying to make up for it with whiz bang neato bits made with Flash. Need a menu? Guess what, there are great ways to either do that with pure HTML or a combination of DHTML and Javascript. Why do people feel obcessed to implement simple things in complex and incompatible ways?
Yours is the only proposal I've heard so far that makes any sense - what ECU designer in their right mind would want an OS in the way? A quick look through Google shows that WinCE makes claims of being an RTOS, but I can't seem to believe it, nor can I find actual specs on what exactly are their definitions of "real time". Quite frankly, I just don't believe it, nor would I personally trust any mission-critical, timing-critical functionality to an operating system that wasn't specifically written for it (think pSOS, etc). Personally, I wouldn't use an RTOS at all in an ECU (basically because there's just no need for it), and yes, I have tinkered a bit with engine computer code.
Like many of the other responses here, this might actually seem fairly reasonable to me. Depends on what the benefit is. If this means that FF can be smaller and faster by dropping a bunch of crappy legacy API calls in favor of better, newer ones, then so be it. If it makes changes to code that is frequently touched less error-prone, then good. If it's just because somebody thought the old calls were "ugly", then it's dumb and the idea should be trashed. I don't know, I'm not a Win32 programmer or FF dev and can't reasonably evaluate the changes.
Someone who is perfectly happy to sit on Win98 is probably perfectly happy to use FF 1.5, because it "just works". I used 2000 up until about eight months ago, because there was no compelling reason to upgrade. Then my wife's laptop met an unfortunate end, and I recycled the XP license onto my desktop. That's the only reason I upgraded, since 2000 is still actively being patched and supported with new drivers.
It's not necessarily a representative sample of the whole world, but just looking at the OS breakdown of poeple who've hit my website since the first of the month, AWStats shows me the following (and these exclude me, btw):
Windows XP 56647 74.5 %
Windows NT 286 0.3 %
Windows Me 897 1.1 %
Windows Vista (Longhorn) 4 0 %
Windows CE 10 0 %
Windows 98 2163 2.8 %
Windows 95 84 0.1 %
Windows 2003 117 0.1 %
Windows 2000 8084 10.6 %
Quite simply, because I use a radio for more than just the same music over and over... It's nice to be in the middle of BFE (or at least west-central British Columbia) and still be able to keep up on news from home, whatever NPR has on that day, etc. For being hundreds of miles north of the target area for Sirius (US lower 48), the darn thing worked very well. I probably only had it cut out once or twice on the whole two week trip.
I wouldn't live without satellite radio these days, because it works wherever I am 99.9% of my life. Will I touch HDR with a 10 foot pole? Not remotely likely. It's not sound quality that made me stop listening to AM/FM broadcast radio, it's the small coverage areas and the usual, rebundled crap content. About the only broadcast radio I listen to any more is my local NPR station for some programming not carried on the satellite NPR channels.
Hours getting a modem to work? Sheesh, but then again that's why I only did external modems. RS232 is so simple it's really damn hard for any manufacturer to screw it up. Internal modems always seemed like a kluge, except for the ones that literally had a UART connected to the bus, and then that was connected to the modem guts. Software-based modems? Don't get me started - total ugly kluge to save a few bucks. It's like anything else - if you buy a serious kluge, you're going to have trouble.
I probably build an average of a system every two months (friends, family, etc.) for other people. In the last six years, I haven't had any components that "just didn't like each other". Most of them go off without a hitch, and wind up being very easy to upgrade (future cost savings) because they all use robust, standard components, not some bizarre crap an OEM/VAR decided was a good idea. I find that most of the ways that VARs add "value" is really adding nonstandard crap that can't be fixed, supported, or upgraded later.
The savings come in when I can do selective upgrades. Upgrade my vid board because I want better framerates for a new game I just picked up? No problem. Need a dual core because of some new project? No problem - planned for that when I bought the motherboard. Want to upgrade the PS to an 80% efficient one, or an ultra-quiet one? No problem. Consider that I haven't upgraded my actual case in five years, nor my power supply in the last three, yet I've gone through several rounds of motherboards/procs. Each time I just upgrade pieces, I save over either having to a) do without or b) go buy a whole new box.
Bottom line is decide what fits your needs. If you know what you're doing, build a machine if you want. If you're willing to invest time in learning, build a machine if you want. If you have no clue and/or don't want to invest the time to learn, pay somebody else to do it for you. Ain't specialization a great economic concept?
Seriously, you might lose some power. Ethanol has only 60% of the energy density of gasoline. Assuming a 10% mix, that means you're losing about 4% of your net energy concentration. There's some evidence that not all of this is a loss, however, since ethanol burns somewhat faster, leading what amounts to really about a 2-3% drop.
As far as destroying engines, I highly doubt it. I used to live in Iowa, and you pretty much can't get gas without EthOH. I've owned several cars that ran their entire lives on ethanol-blended gas, and all of them were at nearly 200k miles before I sold them, usually still running fine with original engines and fuel systems. My current everyday car is a 1995 Honda that's pushing 250k, and ran blended gas for its first 200k miles. (I bought it at 100k, and know the previous owner and what they did with it.) Unfortunately, the last 50k have been on non-blended gas, and I've more problems with injector fouling than I ever did when pushing ethanol through the system. Difference in running on straight and blended fuel (it's blended here in the summer, and I get back to Iowa for a week or two at a time)? Net wash in perceptable MPG, and yes, I keep records on every tank.
I realize that my experience does not a peer-reviewed study make, so I won't say authoritatively that it does not. However, my experience is such that leads me to think that you have no idea what you're talking about, unless you can back it up with some sort of evidence, preferably real studies.
God, I'm glad I'm not the only one who immediately went to that quote when seeing the name...
For those of us who would rather cut off our heads rather than deal with a cable company again, I'm perfectly happy to buy without giving a second thought to cablecard. I went to satellite TV about five years ago and will never go back.
I agree, common sense dictates a UPS. But the guy was beyond that, his filesystem was trashed, and he needed to fix it.
Quite frankly, arguing that ext2 is fine because you should have had a UPS is much like arguing that if you have airbags, it doesn't matter if the seatbelt in your car only works about 80% of the time. Both are making up for a non-robust system by throwing more crap at it. That only improves your odds, it doesn't improve the overall robustness. Eventually the complexity of system-on-system-on-system will bring you down. A UPS guards against power failures, but not against a failed power supply. So you throw more power supplies at it, but then what about a voltage regulator on a proc that suddenly dies? Where does it stop, this heaping on more crap to make up for an underlying flaw?
The best solutions are nearly always eerily elegant. You fix the underlying problem - that the filesystem isn't recoverable from one of the most likely things that can happen, an unclean shutdown. Journalling filesystems are one such elegant solution.
I agree that an attitude problem is common in IT, but I don't think it's limited. We have a lot of management with an equal problem - they think they're on top of the world, and run the company on their own. They fail to realize that most of them don't even have a technical background, and if the engineers walked out tomorrow in frustration, they'd take the company with them. Not only that, they fail to recognize that a lot of the intellectual capital the company has is locked up in a handful of very talented people, not spread equally across all the underlings. So firing Bill the Worthless and replacing him with an intern might be fine, but firing Ted the Chief Engineer, who designed half the stuff the company uses, might not be such a good idea because there's nobody who can fill his shoes.
Let's face it, no matter how awesome the management team is, it's the technical people that create the products that keep the company in business. Likewise, the technical staff usually needs effective guidance from management so that technology is targetted at business goals, not just firing off willy-nilly at whatever captures their interest today.
Definitely yes. X10 may suck, but it's way better than not having it.
A few examples:
- My current house has a very stupid wiring setup. Three of the bedrooms have a ceiling fan and light in the center of the room, which serves as the main light for the room. These were installed when the house was built. Each of these bedrooms has a light switch by the door. Switch doesn't connect to fan at all. In fact, they're on separate circuits on opposite phases - the switch runs an outlet which usually winds up behind the bed. Running new wire would involve a great deal of work (breaking out drywall, running cable, and then sealing the whole mess back up). Solution? Throw an X-10 transmitter into the socket (my personal choice being Keypadlincs), put two wire-in relay modules in the fixture box above the fan, do a little programming, and the whole thing just works.
- Since I currently have a 6-switch Keypadlinc in each bedroom (along with another 8 throughout the rest of the house), I can control more stuff. In particular, I find myself leaving a downstairs lamp on quite frequently. So I programmed it into one of the spare buttons. Likewise, my front outside house lights (small and tasteful) and the backyard light (affectionately known as the Death Ray or the Daybringer) are also connected in. If I hear a noise I want to check out, I just hit a button on the wall rather than having to go downstairs.
- Over Christmas, my outdoor Christmas lights were run via X-10. I could turn them on and off at will, or my firewall box would shut them off after a certain time of night, in case I forgot. Since I wanted to control them from the bedroom so I could kill them as I went to bed, I just temporarily reassigned one of the keypad buttons to run them.
- My wife and I both travel a lot and leave the house unattended, sometimes for a week or two at a time. Using the X-10 interface on my firewall box, I can log in, turn on the lights, check the security cameras scattered about the house, and then return the lights to their previous state. Also, I don't just have a timer on a lamp to make the house look occupied. I have my firewall box set up to create a log of when I turn lights on and off, since the whole house is under X-10 control. When I'm gone, I just tell it to play back the previous week's log. It's virtually indistinguishable whether I'm home or not, at least from the lights.
Damn, the people you run into on /. How ya doin', Steve? (It's Nathan)
It is possible to get X-10 to work reliably, it just takes some work, a little tinkering, and quality components. My house is now almost exclusively based on X-10, largely on Smarthome Keypadlincs. Whoever wired this place initially was a moron, and pulling new romex was way more trouble than it was worth. Light switches that don't control the lights, places where you have to wander into dark rooms to find the lightswitches, etc. Enter powerline carrier gear. I've found that as long as I stay away from the cheapo X-10 crap (the stuff largely marketed by x10.com) and stick with either Smarthome or Leviton-manufactured bits, things work fairly well. Oh, and a whole-house filter and active phase-to-phase repeater. I have one TV that soaks up signal and thus must be filtered at the outlet, but otherwise everything has "just worked" for about 18 months now.
Insteon looks like a quantum leap forward, but I haven't embraced it because it's a single-source system. Once 3-4 vendors make products, I might consider upgrading. The protocol is much, much, much more solid (acknowledgements, checksums, more data bytes), and I definitely wouldn't complain about better response times. Open source support is a deal-breaker, however. Like Spidey pointed out, Neil Cherry seems to not think this will be an issue based on his conversations, but I'm taking a very wait-and-see attitude. Even if SHM never officially supports it, if at least someone gets it working and they don't get sued, I'll consider it workable. Their closed attitude does rather concern me, however - exactly who do they think they're selling to if it's not the tinkerer market?
I'll point out that Smarthome was never really helpful about documenting the original X-10 USB Powerlinc protocol, either. I messed with it for a while by sniffing with the Windows driver, and then decided it was easier to stick with good ol' RS232. Eventually I just started using WiSH, and the open source community eventually worked out the details on the USB PLv1. (I continue to use serial, however)
Call me pragmatic, but I've always evaluated technology on its suitability to the task and rather it accomplishes my goals (price, support, stability, security, documentation, language du jour, whatever), unless I have specific moral objections to the supplier (CEO eats third-world children for desert, and buying said product would give him funds to import eight more children, etc.) Otherwise, if my component was written by Nepalese nudists but it was free, full-featured, stable, and fast, I'll take it. I don't care who wrote it, as long as it works for me.
Those who procure technology based on what color Armani suit the sales guy wore are going to wind up buying overpriced junk. Those who fairly judge technology on its merits and suitability to an application will succeed. Eventually group 1 will catch on once they see the example of group 2.