Yep. So far those G5s have been comparable (some times a hair faster, sometimes a hair slower) to Opterons of the same clock speed, with slightly less power draw. So, my answer is, I want as little intel-based hardware around the lab as possible, at least until they get the floating-point performance up there on the EM64Ts. I had so hoped that the XServes would have gone to a modified Power5 architecture, rather than right off the map.
Every now and then I get a butterfly in the gut that Intel/Apple is going to spring a low-power Itanium on the high-end users, as an attempt to finally find a market for all that investment. In a way, it would be quite in form. 1984: we use the technologically superior to Intel 8086 Motorola M68K family; 1994: we use the technologically superior to Intel 386/486 IBM/Motorola PowerPC family; 2007 we use the technologically superior to Intel EM64T Intel Itanium-2 family.
Repeat after me: My startup budget went to the cluster of G5s, and it sees ridiculous to put $2100 machines in front of grad students (don't forget they need a screen), when I can buy $1200 -$1800 machines for users (I'm running my University-sanctioned G5 on my desktop), and put any other funds into memory, software, network interfaces, or more compute nodes.
I actually am running 64-bit programs (quantum chemical / quantum materials science simulations), and the advantage of the 64-bit desktop isn't the memory, it's the ability to test the numerical stability of the code on the identical processor/OS combination it's eventually going to be used on in production. It would have been nice to afford PowerMacs right out of the door, but that would have seriously cut into the budget for the cluster, where most of the work gets done.
Given that desks are small and the lab small as well, self-contained 17 to 20" machines for use by developers were a definite feature of the Apple line-up when I started. This isn't a fatal change, just an annoying one. The real annoying one is going to be if the xServes go to EM64T. I'll live, but may at that point move the cluster to either real Opterons, or Power5s for any subsequent nodes.
Not to overly complain (and for the laptop users, this point is moot), but the previous version of the iMac was 64-bit, G5 based. This meant that you could buy an iMac as a development platform, then push the package up to PowerMacs or XServes. The new iMac, while I'm sure quite snappy and all that, can only do that for 32-bit code. I know, you can cross-compile, and test it on the remote system, but that's less convenient than local compiling, testing, and debugging.
In the end, maybe this doesn't matter, but it does seem rather inconvenient. I hope that this trend isn't carried into the xServes (or iServes, or whatever they become).
Just idly curious: I was brought up on old sci-fi like Niven, with the various Bussard Ramjets as the plot device to get us to interstellar space. However, one story had a more reasonable design of a scoop made of some ultra-light mesh, which was magnetized.
So...
Since much of the junk is fairly small, and it should be carrying a charge from the solar wind, could you run a dragnet across some suitable orbit, charged to the same polarity, and electrostatically bump everything below it at a certain range back into the atmosphere? Is this idea (it's late) on the scale of "in a few years with some engineering", "space-elevator time-frame", or "we'll see Bussard ramjets leave for Tau Ceti first?"
I presume the alternative "try to live with it", is going to become untenable soon.
As I remember from running simulations at the time, the top of the game was IBM's Power, with DEC's Alpha close behind, then followed by SGI/MIPS. The MIPS R8000 was the first hard-core contender, but they were already having trouble keeping up with DEC/IBM. Sparc was on anyone's radar because it was cheap (relatively), and all of the software written for the previous 3/XX generation could still run.
We used to be really psyched that the PowerMacs had a version of IBM's workstation chip inside (PPC 601/604 chips were in both PowerMacs and AIX workstations). A lot of people bought them for Mathematica as a result.
Actually, I come from one of those towns. I put up with years of "what are you going to do with all that education", "sitting in an office isn't real work", and "but that's UnGodly!" I've had people from my home church tell me that my major was thought up by "Pot Smoking Wierdos grasping at straws". I had friends who went to college but returned to the area, married locally, and became worried about the unchristian influences being beamed in from the big city.
So, yes, I have some stereotypes about those parts of the country, but I've come by them honestly. I really would like to see my home town (lovely area, good outdoors activities, not unreasonably far from urban areas) start attracting telecommuters, but those attitudes are real, and some adjustments are going to have to be made before there is an exodus to such places.
Not to rain on this particular parade, as I'd love to see certain areas I've lived in remain viable, but one of the issues for knowledge-economy is intellectual openness. How many small towns are going to put up with educated outsiders full of "Ideeers" coming in and changing things? If they have some experience (i.e. upstate NY, which used to have Kodak, Xerox, etc), then it's a return to a more profitable era, but for other regions, it's going to be "you dress funny, eat the wrong foods, don't worship our God often enough and we won't even get started on your foreign car". The school systems are also generally in need of upgrading to attract the type of workers that IT or other high-tech needs, and that starts even more conflicts. In modern societies with functioning educational systems, this idea might work. In many parts of the US, it's probably not worth the trouble.
Look at places such as Binghamton/Owego NY (I'm sure you have your local equivalents); even with a moderate-sized public university present, approximately 3 hours from NYC and Philly, very reasonable property, and a skilled workforce downsized from IBM, you can't attract enough investment to do better than limp along here. No local tech business of any size has been started to replace what's been lost, and the local governments aren't willing to take any meaningful steps to either encourage entrepeneurs or relocation by established businesses. Extrapolate this experience to some former wheat depot in Kansas, and you begin to see the problem.
I would put more money on relocation to the inner-city, gentrification, and reuse of brownfields than I would outsourcing to rural america. A cleaned-up Joiliette or Gary, IN, would be far more attractive than Snakenavel, KS.
That's true, which is why when a local company donated a pile of parts from a failed project to my college, I took the IBM 360 Emulator board. I lost 2 XT slots, but it backfilled my ram to 640K.
Too bad the software we had never worked right, as it would have been college-age nerd cool to be the only OS/360 user in a VMS shop.
A PC-XT, Hercules (real) MGA adapter, 40M HD, running Wordstar. It may still be the least obtrusive computer I ever owned.
You'll have to confirm this with the Folding@Home people, but the versions I saw previously were spitting out Gromacs http://www.gromacs.org/ messages, which is written in C++. Depending on what version they use will depend on whether the Gromacs team included the PPC optimizations.
Incidentally, IBM's XL Fortran simply screams on G5s. I've compiled ab-initio codes (Quantum Chemistry) with it, and get speeds identical to Opterons running Linux, compiled with Pathscale or IFC (in em64T mode). IBM's fortran may be why my cluster gets rolled into PPC Linux, when I can afford memory upgrades, rather than remain under OS-X. I like OS-X for the ease of use and programming, but it's hard to beat XLF in the field.
People are lazy, and the viewfinders of most dSLR are like looking through a toilet-paper tube. In other words, they're awful to focus, especially as the split-image or microprism aids are often missing, as they cause problems for the autofocus.
Oh well, there's enough manual-focus stuff on the used market to keep me set for the rest of my life.
Given evidence from the era of Microsoft hacking with Win 3.11 to make sure that it broke Windows compatibility, OS/2s demise was only partly that IBM couldn't market eternal life in 1993.
We ran it too, used it to multitask DOS programs, run Win3.1 apps more stabily than Win3.1 did, and to run native apps that needed the 32-bit address space. It was great to be able to recompile our VAX apps with Watcom Fortran, run them (and get a speed-boost over the VAX), and still be able to use the computer for other apps. Other research groups had it powering their Mass-Spectrometers, and similar fussy hardware.
More importantly, we never had a problem with frequent crashes. We bought good memory and standard hardware, and made sure that we had 8-16 meg, which seemed to be the sweet spot. It just ran. I didn't leave it behind until NT 4 had a service pack or two behind it, and I'd acquired a PowerIndigo2 with the Cray-derived Fortran compiler at work, pretty much eliminating why I was still running OS/2.
We're still paying for the mistake of not adopting it, as many of the security problems in Windows stem from single-user, insecure, Windows95 getting released and established first, rather than VMS|OS/2 derived NT.
So as long as he's careful to leave before about 4:00 p.m., and avoid the bottlenecks, he should be at Mount Doom a little after rushhour, throw the Ring in, and still be home in time to catch RoTK on cable. Should take no time at all, being as the Dark Lord will be too busy trying to stop pop-ups from obscuring his web-cam.
If it currently works on AS/400, and you really need to upgrade the system, then I'd change one variable (the hardware), and keep running the current ERP. If they later wish to transition to a Windows-only solution, make sure that you get a generous time-table and enough technical support to ensure that it's running smoothly before the old system is turned off.
I'm sure that's being done, but sometimes executives get bit by a buzzword-compliant vendor, and lose sight of what's actually at risk; your entire business. Remember when Hershey shot itself in the foot over a several-month period when their SAP upgrade didn't work as well as it should have.
Note, I have no particular love for AS/400s, but I do believe in being cautious when potentially screwing up my entire environment and calling months of unpleasant work down on my head.
You'll remember that one of the issues that brought Newt distress was that he was conversing freely over a cellphone, and didn't think that anyone was listening. There's also the old Chicago saying, "First guy on the bus gets the best seat"; i.e. when someone's going down, you never know what they've been saving and would be willing to reveal in return for better treatment.
Both LBJ and RMN taped their own conversations, which turned around to bite them later. Never underestimate hubris on the part of the powerful.
Good for you. I don't give myself privs on the system (I have a separate account for root-access), and I'm certainly not giving people who aren't familiar with all of the ins and outs of a production system access. I am most certainly not giving developers, who have a tendency to muck with libraries and paths to solve problems, access, even if its logged. Being able to yell at the specific miscreant later really is poor compensation for having to take a production system down, repair their handiwork, and deal with the rest of the angry users.
I used to deal with this on our production cluster all of the time. I implemented a pretty rigid policy early on, which was no root for users on anything that was (1) in production, or (2) had access to the various network servers. This policy came about after a few 'experienced' users demonstrated their skills. Accidentally changing access privs and ownership of the/Users directory tends to raise sysadmin blood pressure, as does nuking/etc, thinking it was ~/etc, or updating a system library to fix your program, which then breaks production codes that people are actually using.
The problem always seems to be that people who've admined their own, solitary, system, think that experience automatically translates into full privs on a much larger, integrated, environment. This is where I miss VMS, and its fine-grained privs. I'm not sure I'd hand those out either, but at least it's better than all or nothing. The next best solution is giving developers access to a box that you can nuke and reimage back to a standard state, and letting them hack with that.
What will be bad is when the information flows for those two systems get crossed.
You'll go through the booth, a cop will pull you over and demand you get off the freeway, but he'll hand you your new iPod. Probably with C. W. McCall's "4 Wheel Drive" already installed.
Army, with the Navy acquiring whatever portion they don't already do in-house. The missile silos would go to Army, as they're already used to garrison-duty.
The Hog had one other major problem; it's used for ground support, and not to chase other fighter jocks across the sky. The way to promotion and pay is not through aiding the ground-pounders.
Somewhere in here is an argument for giving the bombers and warthogs to the Army, the fighters to the Navy, and dissolving the AirForce. Just think what could be saved just on Airforce-blue socks alone.
Yes, but will that get us closer to Fusion-powered Ramjets?
But seriously, I'd predict better photovoltaics, thermoelectrics, and fuel cells. None of this will power anything bigger than a lawn-mower, but it will look great in the lab.
On the other hand, what won't look better will be designer bioweapons. Not that they'll be released, just the capability will give us one more thing to worry about.
Someone will realize that with Google's increasing suite of information organizing technology, that they can become a privatized CIA/NSA. Watch for new bloggers who actually do data mining, rather than off-their-meds rants.
And to stay on topic, controlled nuclear fusion will be right around the corner.
No, it's something completely Yuck that absolutely has to be done for the good of the organization. Think of it as shorthand for the gold-standard of yuck tasks.
I'll gently disagree. I've had jobs I've enjoyed not because they were fun, per-se, but because the Not-Fun was kept far away from me. Basically, the jobs were interesting, management was supportive or absent, and i was left unencumbered to actually do work and accomplish something. You can skip the foosball tables, hawaiian shirt day, and mandatory bowling outings, and instead fire the incompetent cow-worker and preening, empire-building, managers. Average pressure decreases, more work gets done, and the employees are more contented.
I had a job years later that had more outright "Fun", but it also had some gold-plated broccoli moments, which in the end trumped the former, and caused me to depart. Fun is fine, but if the work in between gives you an acid stomach before you even arrive, whether or not the trigger incident happens that day, then the job is not enjoyable. Of course, the compensations were fun like trying out large simulations on new machines before they were released to users.
So, it's not fun, but the lack of its opposite, that makes a job more enjoyable. I think the motto for workplaces should be "Life, Liberty, and the Pursuit of Nirvana."
It's an interesting question. A few years ago, I was buying $100 Matrox cards for OpenGL applications that we used to use SGI Solid Impact workstations for. A few years later, it was $100 Radeons, when I could get them with 64-128Mb of main memory. The only time we bought something more expensive was when we wanted to drive a stereo-wall, without syncing together two separate machines. (we'd done that, with a pair of Ultra-60s, and didn't want the headache of doing it with Windows).
$100 is a good price point; still good performance, but out long enough for the drivers to be stable and most of the glaring bugs gone. I suspect that the card used to matter more, because a few years back displays started getting better rapidly, while falling in price, so suddenly the average user (i.e. the one buying the $1k - $800 machine) could afford a 1280x1024 monitor. The stock video cards that shipped by default circa 1998 had trouble running one of those without flickering. Now, the performance-limiting component seems to be core memory, and occasionally the harddrive.
Of course, most PC parts are wildly overpowered/priced for the average user, except the one that matters; the Power Supply. That seems to be underspeced, and where people try to save money.
I would think (having installed it) that it's a mixture of good hardware support, easy install, and compactness. One CD to boot, and then it grabs the rest from the net. From a psychology standpoint, that's much better than alternatives such as SuSE (which I used for years and actually prefer), which require you to stand there and swap disks.
Undoubtedly, the power-to-the-people ethos, and an easily pronouncible, yet abstract to most users, name, helps as well. Branding matters; think of Xerox, Kodak, Exxon. I know Ubuntu has a real meaning, but not natively to the english speaking world. It allows people to think of it as a brand, and it's a very comfortable word to pronounce. It has a downright reassuring sound.
The best part is, of course, that they took the pain out of Debian, and brought it to the average, versus power, user.
Uh, dude, you want to give this technology to an perennial bachelor genius acrobat living in a home office/cave, who likes to dress in skin-tight outfits, go out at night, and beat people up?
Slashdot Mirror
Yep. So far those G5s have been comparable (some times a hair faster, sometimes a hair slower) to Opterons of the same clock speed, with slightly less power draw. So, my answer is, I want as little intel-based hardware around the lab as possible, at least until they get the floating-point performance up there on the EM64Ts. I had so hoped that the XServes would have gone to a modified Power5 architecture, rather than right off the map.
Every now and then I get a butterfly in the gut that Intel/Apple is going to spring a low-power Itanium on the high-end users, as an attempt to finally find a market for all that investment. In a way, it would be quite in form. 1984: we use the technologically superior to Intel 8086 Motorola M68K family; 1994: we use the technologically superior to Intel 386/486 IBM/Motorola PowerPC family; 2007 we use the technologically superior to Intel EM64T Intel Itanium-2 family.
Repeat after me: My startup budget went to the cluster of G5s, and it sees ridiculous to put $2100 machines in front of grad students (don't forget they need a screen), when I can buy $1200 -$1800 machines for users (I'm running my University-sanctioned G5 on my desktop), and put any other funds into memory, software, network interfaces, or more compute nodes.
I actually am running 64-bit programs (quantum chemical / quantum materials science simulations), and the advantage of the 64-bit desktop isn't the memory, it's the ability to test the numerical stability of the code on the identical processor/OS combination it's eventually going to be used on in production. It would have been nice to afford PowerMacs right out of the door, but that would have seriously cut into the budget for the cluster, where most of the work gets done.
Given that desks are small and the lab small as well, self-contained 17 to 20" machines for use by developers were a definite feature of the Apple line-up when I started. This isn't a fatal change, just an annoying one. The real annoying one is going to be if the xServes go to EM64T. I'll live, but may at that point move the cluster to either real Opterons, or Power5s for any subsequent nodes.
Not to overly complain (and for the laptop users, this point is moot), but the previous version of the iMac was 64-bit, G5 based. This meant that you could buy an iMac as a development platform, then push the package up to PowerMacs or XServes. The new iMac, while I'm sure quite snappy and all that, can only do that for 32-bit code. I know, you can cross-compile, and test it on the remote system, but that's less convenient than local compiling, testing, and debugging.
In the end, maybe this doesn't matter, but it does seem rather inconvenient. I hope that this trend isn't carried into the xServes (or iServes, or whatever they become).
Just idly curious: I was brought up on old sci-fi like Niven, with the various Bussard Ramjets as the plot device to get us to interstellar space. However, one story had a more reasonable design of a scoop made of some ultra-light mesh, which was magnetized.
So...
Since much of the junk is fairly small, and it should be carrying a charge from the solar wind, could you run a dragnet across some suitable orbit, charged to the same polarity, and electrostatically bump everything below it at a certain range back into the atmosphere? Is this idea (it's late) on the scale of "in a few years with some engineering", "space-elevator time-frame", or "we'll see Bussard ramjets leave for Tau Ceti first?"
I presume the alternative "try to live with it", is going to become untenable soon.
As I remember from running simulations at the time, the top of the game was IBM's Power, with DEC's Alpha close behind, then followed by SGI/MIPS. The MIPS R8000 was the first hard-core contender, but they were already having trouble keeping up with DEC/IBM. Sparc was on anyone's radar because it was cheap (relatively), and all of the software written for the previous 3/XX generation could still run.
We used to be really psyched that the PowerMacs had a version of IBM's workstation chip inside (PPC 601/604 chips were in both PowerMacs and AIX workstations). A lot of people bought them for Mathematica as a result.
Actually, I come from one of those towns. I put up with years of "what are you going to do with all that education", "sitting in an office isn't real work", and "but that's UnGodly!" I've had people from my home church tell me that my major was thought up by "Pot Smoking Wierdos grasping at straws". I had friends who went to college but returned to the area, married locally, and became worried about the unchristian influences being beamed in from the big city.
So, yes, I have some stereotypes about those parts of the country, but I've come by them honestly. I really would like to see my home town (lovely area, good outdoors activities, not unreasonably far from urban areas) start attracting telecommuters, but those attitudes are real, and some adjustments are going to have to be made before there is an exodus to such places.
Not to rain on this particular parade, as I'd love to see certain areas I've lived in remain viable, but one of the issues for knowledge-economy is intellectual openness. How many small towns are going to put up with educated outsiders full of "Ideeers" coming in and changing things? If they have some experience (i.e. upstate NY, which used to have Kodak, Xerox, etc), then it's a return to a more profitable era, but for other regions, it's going to be "you dress funny, eat the wrong foods, don't worship our God often enough and we won't even get started on your foreign car". The school systems are also generally in need of upgrading to attract the type of workers that IT or other high-tech needs, and that starts even more conflicts. In modern societies with functioning educational systems, this idea might work. In many parts of the US, it's probably not worth the trouble.
Look at places such as Binghamton/Owego NY (I'm sure you have your local equivalents); even with a moderate-sized public university present, approximately 3 hours from NYC and Philly, very reasonable property, and a skilled workforce downsized from IBM, you can't attract enough investment to do better than limp along here. No local tech business of any size has been started to replace what's been lost, and the local governments aren't willing to take any meaningful steps to either encourage entrepeneurs or relocation by established businesses. Extrapolate this experience to some former wheat depot in Kansas, and you begin to see the problem.
I would put more money on relocation to the inner-city, gentrification, and reuse of brownfields than I would outsourcing to rural america. A cleaned-up Joiliette or Gary, IN, would be far more attractive than Snakenavel, KS.
That's true, which is why when a local company donated a pile of parts from a failed project to my college, I took the IBM 360 Emulator board. I lost 2 XT slots, but it backfilled my ram to 640K. Too bad the software we had never worked right, as it would have been college-age nerd cool to be the only OS/360 user in a VMS shop. A PC-XT, Hercules (real) MGA adapter, 40M HD, running Wordstar. It may still be the least obtrusive computer I ever owned.
Drat. That will teach me not to look in the source directory before posting. My bad.
You'll have to confirm this with the Folding@Home people, but the versions I saw previously were spitting out Gromacs http://www.gromacs.org/ messages, which is written in C++. Depending on what version they use will depend on whether the Gromacs team included the PPC optimizations.
Incidentally, IBM's XL Fortran simply screams on G5s. I've compiled ab-initio codes (Quantum Chemistry) with it, and get speeds identical to Opterons running Linux, compiled with Pathscale or IFC (in em64T mode). IBM's fortran may be why my cluster gets rolled into PPC Linux, when I can afford memory upgrades, rather than remain under OS-X. I like OS-X for the ease of use and programming, but it's hard to beat XLF in the field.
People are lazy, and the viewfinders of most dSLR are like looking through a toilet-paper tube. In other words, they're awful to focus, especially as the split-image or microprism aids are often missing, as they cause problems for the autofocus.
Oh well, there's enough manual-focus stuff on the used market to keep me set for the rest of my life.
Given evidence from the era of Microsoft hacking with Win 3.11 to make sure that it broke Windows compatibility, OS/2s demise was only partly that IBM couldn't market eternal life in 1993.
We ran it too, used it to multitask DOS programs, run Win3.1 apps more stabily than Win3.1 did, and to run native apps that needed the 32-bit address space. It was great to be able to recompile our VAX apps with Watcom Fortran, run them (and get a speed-boost over the VAX), and still be able to use the computer for other apps. Other research groups had it powering their Mass-Spectrometers, and similar fussy hardware.
More importantly, we never had a problem with frequent crashes. We bought good memory and standard hardware, and made sure that we had 8-16 meg, which seemed to be the sweet spot. It just ran. I didn't leave it behind until NT 4 had a service pack or two behind it, and I'd acquired a PowerIndigo2 with the Cray-derived Fortran compiler at work, pretty much eliminating why I was still running OS/2.
We're still paying for the mistake of not adopting it, as many of the security problems in Windows stem from single-user, insecure, Windows95 getting released and established first, rather than VMS|OS/2 derived NT.
So as long as he's careful to leave before about 4:00 p.m., and avoid the bottlenecks, he should be at Mount Doom a little after rushhour, throw the Ring in, and still be home in time to catch RoTK on cable. Should take no time at all, being as the Dark Lord will be too busy trying to stop pop-ups from obscuring his web-cam.
If it currently works on AS/400, and you really need to upgrade the system, then I'd change one variable (the hardware), and keep running the current ERP. If they later wish to transition to a Windows-only solution, make sure that you get a generous time-table and enough technical support to ensure that it's running smoothly before the old system is turned off.
I'm sure that's being done, but sometimes executives get bit by a buzzword-compliant vendor, and lose sight of what's actually at risk; your entire business. Remember when Hershey shot itself in the foot over a several-month period when their SAP upgrade didn't work as well as it should have.
Note, I have no particular love for AS/400s, but I do believe in being cautious when potentially screwing up my entire environment and calling months of unpleasant work down on my head.
You'll remember that one of the issues that brought Newt distress was that he was conversing freely over a cellphone, and didn't think that anyone was listening. There's also the old Chicago saying, "First guy on the bus gets the best seat"; i.e. when someone's going down, you never know what they've been saving and would be willing to reveal in return for better treatment.
Both LBJ and RMN taped their own conversations, which turned around to bite them later. Never underestimate hubris on the part of the powerful.
Good for you. I don't give myself privs on the system (I have a separate account for root-access), and I'm certainly not giving people who aren't familiar with all of the ins and outs of a production system access. I am most certainly not giving developers, who have a tendency to muck with libraries and paths to solve problems, access, even if its logged. Being able to yell at the specific miscreant later really is poor compensation for having to take a production system down, repair their handiwork, and deal with the rest of the angry users.
/Users directory tends to raise sysadmin blood pressure, as does nuking /etc, thinking it was ~/etc, or updating a system library to fix your program, which then breaks production codes that people are actually using.
I used to deal with this on our production cluster all of the time. I implemented a pretty rigid policy early on, which was no root for users on anything that was (1) in production, or (2) had access to the various network servers. This policy came about after a few 'experienced' users demonstrated their skills. Accidentally changing access privs and ownership of the
The problem always seems to be that people who've admined their own, solitary, system, think that experience automatically translates into full privs on a much larger, integrated, environment. This is where I miss VMS, and its fine-grained privs. I'm not sure I'd hand those out either, but at least it's better than all or nothing. The next best solution is giving developers access to a box that you can nuke and reimage back to a standard state, and letting them hack with that.
What will be bad is when the information flows for those two systems get crossed.
You'll go through the booth, a cop will pull you over and demand you get off the freeway, but he'll hand you your new iPod. Probably with C. W. McCall's "4 Wheel Drive" already installed.
Army, with the Navy acquiring whatever portion they don't already do in-house. The missile silos would go to Army, as they're already used to garrison-duty.
The Hog had one other major problem; it's used for ground support, and not to chase other fighter jocks across the sky. The way to promotion and pay is not through aiding the ground-pounders.
Somewhere in here is an argument for giving the bombers and warthogs to the Army, the fighters to the Navy, and dissolving the AirForce. Just think what could be saved just on Airforce-blue socks alone.
Yes, but will that get us closer to Fusion-powered Ramjets?
But seriously, I'd predict better photovoltaics, thermoelectrics, and fuel cells. None of this will power anything bigger than a lawn-mower, but it will look great in the lab.
On the other hand, what won't look better will be designer bioweapons. Not that they'll be released, just the capability will give us one more thing to worry about.
Someone will realize that with Google's increasing suite of information organizing technology, that they can become a privatized CIA/NSA. Watch for new bloggers who actually do data mining, rather than off-their-meds rants.
And to stay on topic, controlled nuclear fusion will be right around the corner.
No, it's something completely Yuck that absolutely has to be done for the good of the organization. Think of it as shorthand for the gold-standard of yuck tasks.
I'll gently disagree. I've had jobs I've enjoyed not because they were fun, per-se, but because the Not-Fun was kept far away from me. Basically, the jobs were interesting, management was supportive or absent, and i was left unencumbered to actually do work and accomplish something. You can skip the foosball tables, hawaiian shirt day, and mandatory bowling outings, and instead fire the incompetent cow-worker and preening, empire-building, managers. Average pressure decreases, more work gets done, and the employees are more contented.
I had a job years later that had more outright "Fun", but it also had some gold-plated broccoli moments, which in the end trumped the former, and caused me to depart. Fun is fine, but if the work in between gives you an acid stomach before you even arrive, whether or not the trigger incident happens that day, then the job is not enjoyable. Of course, the compensations were fun like trying out large simulations on new machines before they were released to users.
So, it's not fun, but the lack of its opposite, that makes a job more enjoyable. I think the motto for workplaces should be "Life, Liberty, and the Pursuit of Nirvana."
It's an interesting question. A few years ago, I was buying $100 Matrox cards for OpenGL applications that we used to use SGI Solid Impact workstations for. A few years later, it was $100 Radeons, when I could get them with 64-128Mb of main memory. The only time we bought something more expensive was when we wanted to drive a stereo-wall, without syncing together two separate machines. (we'd done that, with a pair of Ultra-60s, and didn't want the headache of doing it with Windows).
$100 is a good price point; still good performance, but out long enough for the drivers to be stable and most of the glaring bugs gone. I suspect that the card used to matter more, because a few years back displays started getting better rapidly, while falling in price, so suddenly the average user (i.e. the one buying the $1k - $800 machine) could afford a 1280x1024 monitor. The stock video cards that shipped by default circa 1998 had trouble running one of those without flickering. Now, the performance-limiting component seems to be core memory, and occasionally the harddrive.
Of course, most PC parts are wildly overpowered/priced for the average user, except the one that matters; the Power Supply. That seems to be underspeced, and where people try to save money.
I would think (having installed it) that it's a mixture of good hardware support, easy install, and compactness. One CD to boot, and then it grabs the rest from the net. From a psychology standpoint, that's much better than alternatives such as SuSE (which I used for years and actually prefer), which require you to stand there and swap disks.
Undoubtedly, the power-to-the-people ethos, and an easily pronouncible, yet abstract to most users, name, helps as well. Branding matters; think of Xerox, Kodak, Exxon. I know Ubuntu has a real meaning, but not natively to the english speaking world. It allows people to think of it as a brand, and it's a very comfortable word to pronounce. It has a downright reassuring sound.
The best part is, of course, that they took the pain out of Debian, and brought it to the average, versus power, user.
Uh, dude, you want to give this technology to an perennial bachelor genius acrobat living in a home office/cave, who likes to dress in skin-tight outfits, go out at night, and beat people up?
That's kind of scary too.