Which, in the end, is one of the problems; the Sender sets the importance, not the Reader.
IMHO a simple improvement to email would be no more than twice a day delivery. People would know the corporate email shows up at 6:00 a.m. and 2:00 p.m. Therefore, if that time has passed, you won't get a reply before the next email dump. This removes the pressure on the recipient, who knows he has at least 8 hours before anything has to be done with that email.
A side benefit is that there is only new email twice a day; when you arrive, and mid-afternoon. No more checking it every five minutes, no more boss yelling "did you get my email yet", no little dings/mailbox flags, etc, going off and distracting you from your job. Go a step farther, and let an intelligent agent apply your rules of priority to the message "has the word "superbowl video", so file it under "never"", rather than the sender's, and some of the issues are gone.
For colllabortion between more than 2-3 people, use a Wiki or Notes. Email should be for person-person, ephemeral, communication.
Amusingly, "Fortran Tutorial" returns as many hits as "Ruby on Rails". I don't hear any great clamor around here to drop silly languages such as Ruby, and start writing web-apps in Fortran.
However, PL/I only returns 69,000 hits, which seems to indicate that it's safely dead.
In HS I learned to program using Fortran, on punch-cards. Thankfully, a year later I got a copy of Turbo Pascal, which ran on any PC-compatible (and some times not overly compatible, such as early Epsons) computer. We could do standard class assignments, and code the old graphics-intensive simulations in Scientific American's "computer recreations" column. It was a fun time.
Frankly, one of the classic compiled languages, with some sort of "QuickWin" graphics/windowing library attached (as MS Fortran PowerStation had), would probably be the best teaching language. Simple enough so that the syntax doesn't get in the way of the algorithms, with some sort of basic graphics lib so that graphics can be coded. We got a lot of milage out of the old Turbo-Pascal routines which allowed simple dot-addressed graphics, and presumably so would modern students. The modern equivalent would be some dialect of SmallTalk, as it would introduce OOP into the mix.
An instructor brave enough to say, "just because it's there, doesn't mean you have to use it", could use a sub-set of Java for the same purposes. Alternatively, there's always Matlab/Octave, which offers the advantage that it teaches programming (logic, design, and algorithms), without bogging students down in low-level language-specific syntax. It's best to stay away from the too complicated IDE's, such as MS Visual tools, for the same reason that students being taught composition should be kept away from WYSIWYG word-processors; students will concentrate on the visual presentation, rather than the content.
Bring back Turbo-Pascal 4! Better yet, wrap an IDE and Graphics lib around G95, and introduce Turbo-Fortran! Tools for beginning programmers should be elegant, portable, and capable of creating a result the student would want to run or use, rather than simply hooking them early on one vendor's non-portable development environment.
From WishfullThinkingBusinessTimes: "Well," Mr. Ballmer said, "it was hard to fire myself after all these years, but I called myself into my office, sat myself down, and said, 'Steve, this project is going nowhere fast, and lack of clear leadership at the top is part of the problem. I think it's time you took a break and looked for new opportunities'. I then told myself, I agree, and while I am currently mad at myself for having just fired myself after all of those loyal years to Microsoft, I intend to move on to a company where Developers! Developers! Developers! really matter." "Therefore, I intend to take over Borland, and am giving myself notice that I am going to crush myself in a few years if I don't watch out".
There was the article a year or two back about using microvoltage applied to the temples to counteract (or cause) motion sickness. Seems we should see that first; a flight sim that lets you feel the bank, or a FPS where you feel youself dodging the opponent.
Personally, I'd be more interested in modifying this for Fraud Detection. The robot looks over your data and text, and decides, "Sorry Dave, a leap of faith has occurred here." Presumably, at that point the robot locks you out of your lab.
This could lead to a whole series of literary robots: The Too Many Coincidences in Fiction Detector, The Humanities Thesis Verbiage Reducer, The This Movie Is Going to Suck No Matter Who Acts In/Directs It Detector, and so forth.
Smithsonian Magazine, back in the 80s, ran an article where they compared the staying power of three major american landmarks, if left unmaintained. Unfortunately, I've forgotten the middle one, but the outer two were St. Louis Arch (150 years, tops), and Grand Coulee Dam (50 thousand, or a couple of glaciers, whichever comes second).
So, probably not the bridges (rust thou art, and to rust thou shalt return), but maybe some of the bigger dams, cemented plains of our cities, Yucca Mountain, etc.
Of course, by 100 million years, all that will be left will be the moon landers buried under the dust, and the Voyager and Pioneer probes, drifting through the galaxy.
True, but being as we've lost so many good technologies, such as the Cray-style SHMEM, and others are languishing, i.e. the Origin coupling to make NUMA systems, it's nice to see steps in the right direction. There's always the worry that the future was nothing but business desktops glued together with funky networking and software hacks, and it's good to see that features the HPC community can use directly may become available again. Sometimes you just need the bigger box, and the ability to link Origins together in a modular fashion, from the pair-wise O200 to the full-sized O2K, always seemed to be a neat trick. As I said, whether the first implementation is optimal, bringing that capability to the PC world is a step in the right direction. Having this capability should spur some design effort into implementing the cache coherency and other features of the O2000, and help move the Intel/AMD hardware beyond its business-oriented PowerPoint-running roots.
Of course, I speak as a chemist, for whom the Origin was easy to code for, if somewhat lagging in performance per processor. I would like to see that architecture, modular, and affordably priced, become available again.
So, you take the external interconnects, a large SMP box, and a transfer rate unachievable by anything except channel-bonded Myri/Infiniband/Quadrics, and you've suddenly commoditized (is that a word?) the Origin 2K architecture. Unfortunately, there will be that inevitable gap between "announced" and "benchmarkable", but this should lead to interesting system design.
Computing might just become fun again. Small systems passing information around to form a display wall, or big systems chained together to become huge systems.
It's always looked good on paper, but my one experience in the field has made me radically gun-shy. What are they like to deal with these days? I'd like to be able to run some codes such as CPMD and Siesta distributed parallel.
and lurking there in the background, there's always NWChem....
though I would argue that there was too much time spent discussing GigE, and not enough on the performance and scaling issues seen with the more exotic cards.
Not a technical issue, but a little note about the Infiniband cards reading, "Unlike the alternatives, just try to get information on pricing one of these without leaving all of your contact information for a salesman to use now and in perpetuity." I've been through this recently, and have considered (given the similar performance), purchasing Myrinet because they post their prices out in the open, so that you can make some informed decisions before calling their salescritter.
More technically, some analysis of the stability and maturity of the software stack would be nice. We owned Dolphinics SCI cards once (2001 ish), and while blazingly fast when they worked, on our Opterons the MPI system would mysteriously shut down. They were also very closed and proprietary about their software at that point, so we went round and round over the early 2.4 drivers. Myri, while more expensive, was also more stable.
Finally, to simply geek out for a moment, I saw numbers for the Quadrics cards once. PNNL built their Itanium-2 cluster with multiple quadrics cards per machine to get the bandwidth high enough for their chemistry apps. Light on details, but found at http://www.emsl.pnl.gov/using-emsl/tour/lab.php?fa cility=msc&lab=vr1119/ How to tie together 980 dual-proc Itanium-2 systems.
A casual demonstration as to why not to trust your intuition comes from the early days of flight. Pilots would fly into clouds, and their intuition would tell them they're flying level, but they'd come out tilted. When instruments (the artificial horizon) were added, and they were confronted with the empirical evidence, many of them continued to trust their intuition rather than the damned instrument. Others put their intuition on hold, trusting the instruments even though they didn't feel right.
Guess which of the two sets of pilots had a greater tendency to spiral into the ground. My ex-boss who was a pilot said that when flying and unable to see the ground, "the moment you decide that you're right and the instruments are wrong you're dead". Same principle here; you may not like where the physical evidence takes you, and you may chafe at the limitations it imposes, but it's all you've got to keep you from spinning out of control.
That's because Macs used to have California Scheduling: "hey man, i'm like, not using these cycles, so why don't you go ahead an run for a while..". Now they have New Jersey Sceduling: "YOU! Outta da way! NOW!" Wierdly enough, while the first sounds better in theory, the other works much more effectively in practice.
I remember reading an interview in a business mag about 10 years ago, when the Sparcs were lagging everyone and IBM/HP were gaining market share. His comment was along the lines of, "yeah, everyone's waiting at the end of the runway for us, and not because they think we're going to take off." Then they did take off. I'll admit having thought, "well, they're day is done this time", more than once, and each time they come back. If Scott goes, I hope that attitude of fighting on stays.
And having run both Linux and Solaris in a server environment, here's to figuring out how to push Solarisx86 everywhere. If I ran Opterons, because of the Dev Tools being available, I'd certainly consider running it on a cluster just for the management and profiling tools.
The utilitarian style was part of the Thinkpad's appeal. It reminded you that the money had been spent on the durability, not the flash. I have (though it's been officially retired), an A22m, which still runs, despite being dropped on ice (I went down hard and got a large bruise that lasted weeks. The thinkpad made a sickening crack sound, and didn't even have cosmetic damage). It weighs over 7 lbs, has the high-res screen but only 8m video card, but is amazingly durable and has the nicest keyboard you could ask for. You can type all day on that keyboard without getting wrist-strain. My wife's 600E, despite being older yet, is still bright and clear, with an excellent keyboard, despite being a compact model. Thinkpads just feel good, and have the additional advantage of generally not wearing out prematurely.
Thinkpads were popular because they were built to last, and built to survive the Road Warrior breed of business traveler. The fear with Lenovo is probably that they're pushing into the consumer space, and are building consumer-grade laptops. With IBM, you came to expect stolid efficiency, and Lenovo is going to have to demonstrate a commitment to that as well.
As an aside, I had always hoped that pre-intel-switch, Apple would partner with IBM (ThinkPower? MacPad?), so that you could get an OS-X laptop with IBM build quality. Probably would have come in "His Steveness Sweater Black", just to differentiate it from the regular line-up.
Thankfully the only time I have seen a Token-Ring card was from a Dell screw-up, circa 2001. We needed a computer to control an instrument, which had sensibly been converted to ethernet. The researcher in charge called Dell for a workstation, and said he needed two network cards. Dell convinced him that they couldn't be both of the same type, because you couldn't tell them apart. So, they put in two different cards; a 3Com ethernet, and a Token-Ring.
Just... speechless... when we realized what they'd done. They were also kind of huffy about taking it back. Maybe we accidentally bought their last one, and they thought they could close out that part number for good.
Let's not get too nostalgic. CHRP partly died due to Apple, since they insisted on doing things slightly differently from the standard. We were looking forward to CHRP OS/2 machines, (*sigh*), which quickly materialized and dematerialized almost simultaneously. That would have been a great time for IBM to buy a sufficiently large stake in Apple to make them behave, then offer PPC machines with the MacOS:OS/2:AIX option. Most probably OS/2 would have still faded away, and we would have had cheapish machines (by IBM standards) running MacOS on the desktop, with a clear upgrade path to AIX-based Power machines when you finally needed the Big Iron.
Or we'd still have OS-X, except instead of BSD underpinnings, it would have AIX. One can only imagine what the Aqua version of SMIT would have been like.
Having a spare goat for the troll this morning....
Too keep your linux chops up in case you want a job at a Linux-based organization? Because there is always some app that's not quite happy with the BSDish nature of OS-X, and won't quite build correctly? Because it's a precompiled binary that isn't going to run on OS-X no matter how many black chickens you sacrifice by the dark of the moon?
And, of course, just because you can. It's better than booting Windows on it; at least you don't feel dirty afterwards.
That depends; you're sitting at home and are your own tech support, then you're right. Go wherever you want.
You're on some sort of managed network, such as a business or university system, which is networked intimately to many other systems and has administrators who will have to clean up the machine when you're done, then your browser should prevent you from going to a known malware site. Sorry, I used to have that job, and began thinking, "maybe if I make them all use Lynx and Pine for web and email, this nonsense will stop".
Like most Microsoft innovations, this one is pitched to the home user, to build mindshare, but is really aimed at corporate environments. Don't be surprised if you work at United that if you type http://www.untied.com/, you somehow end up back at United.com. Similarly, small dot-coms might redirect F*d Company's website to wemotivate.com.
That's why you should always homebrew your own. Friends did this, and managed to make a beer that had a decent taste while being consumed, but in about 10 minutes, you would start to fill billious and get a splitting headache. It was the perfect deterrent, Sober to Hangover without ever passing through Drunk. Kind of like Iron City Light, but quicker and with less water retention.
Having grown up United Methodist (communion is done with Welche's Grape Juice), and converted to Episcopalianism (communion is done with Wine), I can authoritatively say that given the choice between Concord Grape, and Fermented _any_ Grape, I'll take the wine. I would still take the wine even if I couldn't get a buzz off it, because at least the yeasts would have done something about all that accursed Sugar in the grape juice before it made it to my glass. Pity that dealcoholized wine is generally pretty vile stuff; with some care, it would be a great drink.
Wine is always the right answer to the question, "what have the Romans ever done for us?"
We've already spent the money on an immensely effective ABM system; it's called MAD. It's the solid belief that if you nuke anything that we value, we're going to make the rubble bounce from one end of your "now a historical question on Jeopardy!" ex-country to the other. It's why we fought only proxy wars after 1945, and didn't, despite the urges to the contrary, use tactical nukes in Korea, Vietnam, Cuba, or any other number of flash-points. This is actually the best incentive certain nominally middle-eastern states have to keep a lid on their stockpiles. It's probable that even if they don't nuke one of our cities, if it happens they're going to get blamed and made an example of.
Basically, we need the old Kissinger/Nixon team at the helm. One of them negotiates with the other power, and says that the President is crazy, and is going to bomb them, then the president acts crazy (and in the case of Nixon and the North Vietnamese), does bomb them, but just enough to get them back to the negotiating table. Same deal here; rail against the imperialist foreign hedonistic infidel running dogs all you want, but don't get too antsy, or bad things happen.
None of the ABM systems we've tested so far has been nearly as successful as good old psychology and enlightened self-interest. We'd be better off spending the money on a couple of underground tests capable of rattling seismometers on the other side of the world and more money for scanning cargo containers as they enter port than on jumbo-jet mounted lasers, missile batteries that have to be told where the missile is coming from, or the rest of the Reagan-era Sci-Fi retreads we are now.
Probably Power5. However, to use desktop processors as the example: I have some benchmarks of a code I've run on machines from a DEC Alpha 21064 through a current G5 2.0.. The timings that matter here are jobs that took 107 seconds on a PII/400 now take 7. The speedup is about constant across the series, so we're looking at a 15x speed-up on my current code versus a PII/400. Since this particular package runs at about the same speed on comparable clock-rate Opterons/G5s, we can infer a probable 15x speedup (30x if they're using PII/266, with the slower bus), for less than twice the power. (250W vs 400W). OTOH, that 400W PS powers two processors, which claim to pull in the >40W range when under load.
In short, as long as you avoid Itanium-2 and certain late P4-Xeons, any modern machine will deliver a pretty acceptable flop/watt rating. This, regrettably, means that while their heart is in the right place, Berkeley's stunt, even if using biodesiel as fuel, would be more efficient if they scrapped the old computers, and used the money to buy a current Mac Mini or two to offer as compute servers. In my last job, even with the institution supplying power as 'overhead', certain machines no longer delivered enough performance to justify them occupying a plug and 9 sq ft of floor space, even if they did work fine. If I had been directly billed for what they consumed versus what they produced, they would have been dumpster-bound years earlier.
This doesn't mean that they shouldn't refurb the old machines, stick an OS and some user-land software on them, and hand them (or sell them cheaply) to potential users for desktop usage, but as a cluster, they're not likely to earn their keep.
Slashdot Mirror
Which, in the end, is one of the problems; the Sender sets the importance, not the Reader.
IMHO a simple improvement to email would be no more than twice a day delivery. People would know the corporate email shows up at 6:00 a.m. and 2:00 p.m. Therefore, if that time has passed, you won't get a reply before the next email dump. This removes the pressure on the recipient, who knows he has at least 8 hours before anything has to be done with that email.
A side benefit is that there is only new email twice a day; when you arrive, and mid-afternoon. No more checking it every five minutes, no more boss yelling "did you get my email yet", no little dings/mailbox flags, etc, going off and distracting you from your job. Go a step farther, and let an intelligent agent apply your rules of priority to the message "has the word "superbowl video", so file it under "never"", rather than the sender's, and some of the issues are gone.
For colllabortion between more than 2-3 people, use a Wiki or Notes. Email should be for person-person, ephemeral, communication.
Amusingly, "Fortran Tutorial" returns as many hits as "Ruby on Rails". I don't hear any great clamor around here to drop silly languages such as Ruby, and start writing web-apps in Fortran.
However, PL/I only returns 69,000 hits, which seems to indicate that it's safely dead.
In HS I learned to program using Fortran, on punch-cards. Thankfully, a year later I got a copy of Turbo Pascal, which ran on any PC-compatible (and some times not overly compatible, such as early Epsons) computer. We could do standard class assignments, and code the old graphics-intensive simulations in Scientific American's "computer recreations" column. It was a fun time.
Frankly, one of the classic compiled languages, with some sort of "QuickWin" graphics/windowing library attached (as MS Fortran PowerStation had), would probably be the best teaching language. Simple enough so that the syntax doesn't get in the way of the algorithms, with some sort of basic graphics lib so that graphics can be coded. We got a lot of milage out of the old Turbo-Pascal routines which allowed simple dot-addressed graphics, and presumably so would modern students. The modern equivalent would be some dialect of SmallTalk, as it would introduce OOP into the mix.
An instructor brave enough to say, "just because it's there, doesn't mean you have to use it", could use a sub-set of Java for the same purposes. Alternatively, there's always Matlab/Octave, which offers the advantage that it teaches programming (logic, design, and algorithms), without bogging students down in low-level language-specific syntax. It's best to stay away from the too complicated IDE's, such as MS Visual tools, for the same reason that students being taught composition should be kept away from WYSIWYG word-processors; students will concentrate on the visual presentation, rather than the content.
Bring back Turbo-Pascal 4! Better yet, wrap an IDE and Graphics lib around G95, and introduce Turbo-Fortran! Tools for beginning programmers should be elegant, portable, and capable of creating a result the student would want to run or use, rather than simply hooking them early on one vendor's non-portable development environment.
From WishfullThinkingBusinessTimes: "Well," Mr. Ballmer said, "it was hard to fire myself after all these years, but I called myself into my office, sat myself down, and said, 'Steve, this project is going nowhere fast, and lack of clear leadership at the top is part of the problem. I think it's time you took a break and looked for new opportunities'. I then told myself, I agree, and while I am currently mad at myself for having just fired myself after all of those loyal years to Microsoft, I intend to move on to a company where Developers! Developers! Developers! really matter." "Therefore, I intend to take over Borland, and am giving myself notice that I am going to crush myself in a few years if I don't watch out".
There was the article a year or two back about using microvoltage applied to the temples to counteract (or cause) motion sickness. Seems we should see that first; a flight sim that lets you feel the bank, or a FPS where you feel youself dodging the opponent.
Personally, I'd be more interested in modifying this for Fraud Detection. The robot looks over your data and text, and decides, "Sorry Dave, a leap of faith has occurred here." Presumably, at that point the robot locks you out of your lab.
This could lead to a whole series of literary robots: The Too Many Coincidences in Fiction Detector, The Humanities Thesis Verbiage Reducer, The This Movie Is Going to Suck No Matter Who Acts In/Directs It Detector, and so forth.
Smithsonian Magazine, back in the 80s, ran an article where they compared the staying power of three major american landmarks, if left unmaintained. Unfortunately, I've forgotten the middle one, but the outer two were St. Louis Arch (150 years, tops), and Grand Coulee Dam (50 thousand, or a couple of glaciers, whichever comes second).
So, probably not the bridges (rust thou art, and to rust thou shalt return), but maybe some of the bigger dams, cemented plains of our cities, Yucca Mountain, etc.
Of course, by 100 million years, all that will be left will be the moon landers buried under the dust, and the Voyager and Pioneer probes, drifting through the galaxy.
True, but being as we've lost so many good technologies, such as the Cray-style SHMEM, and others are languishing, i.e. the Origin coupling to make NUMA systems, it's nice to see steps in the right direction. There's always the worry that the future was nothing but business desktops glued together with funky networking and software hacks, and it's good to see that features the HPC community can use directly may become available again. Sometimes you just need the bigger box, and the ability to link Origins together in a modular fashion, from the pair-wise O200 to the full-sized O2K, always seemed to be a neat trick. As I said, whether the first implementation is optimal, bringing that capability to the PC world is a step in the right direction. Having this capability should spur some design effort into implementing the cache coherency and other features of the O2000, and help move the Intel/AMD hardware beyond its business-oriented PowerPoint-running roots.
Of course, I speak as a chemist, for whom the Origin was easy to code for, if somewhat lagging in performance per processor. I would like to see that architecture, modular, and affordably priced, become available again.
So, you take the external interconnects, a large SMP box, and a transfer rate unachievable by anything except channel-bonded Myri/Infiniband/Quadrics, and you've suddenly commoditized (is that a word?) the Origin 2K architecture. Unfortunately, there will be that inevitable gap between "announced" and "benchmarkable", but this should lead to interesting system design.
Computing might just become fun again. Small systems passing information around to form a display wall, or big systems chained together to become huge systems.
It's always looked good on paper, but my one experience in the field has made me radically gun-shy. What are they like to deal with these days? I'd like to be able to run some codes such as CPMD and Siesta distributed parallel.
and lurking there in the background, there's always NWChem....
though I would argue that there was too much time spent discussing GigE, and not enough on the performance and scaling issues seen with the more exotic cards.
a cility=msc&lab=vr1119/ How to tie together 980 dual-proc Itanium-2 systems.
Not a technical issue, but a little note about the Infiniband cards reading, "Unlike the alternatives, just try to get information on pricing one of these without leaving all of your contact information for a salesman to use now and in perpetuity." I've been through this recently, and have considered (given the similar performance), purchasing Myrinet because they post their prices out in the open, so that you can make some informed decisions before calling their salescritter.
More technically, some analysis of the stability and maturity of the software stack would be nice. We owned Dolphinics SCI cards once (2001 ish), and while blazingly fast when they worked, on our Opterons the MPI system would mysteriously shut down. They were also very closed and proprietary about their software at that point, so we went round and round over the early 2.4 drivers. Myri, while more expensive, was also more stable.
Finally, to simply geek out for a moment, I saw numbers for the Quadrics cards once. PNNL built their Itanium-2 cluster with multiple quadrics cards per machine to get the bandwidth high enough for their chemistry apps. Light on details, but found at http://www.emsl.pnl.gov/using-emsl/tour/lab.php?f
Politely, no.
A casual demonstration as to why not to trust your intuition comes from the early days of flight. Pilots would fly into clouds, and their intuition would tell them they're flying level, but they'd come out tilted. When instruments (the artificial horizon) were added, and they were confronted with the empirical evidence, many of them continued to trust their intuition rather than the damned instrument. Others put their intuition on hold, trusting the instruments even though they didn't feel right.
Guess which of the two sets of pilots had a greater tendency to spiral into the ground. My ex-boss who was a pilot said that when flying and unable to see the ground, "the moment you decide that you're right and the instruments are wrong you're dead". Same principle here; you may not like where the physical evidence takes you, and you may chafe at the limitations it imposes, but it's all you've got to keep you from spinning out of control.
That's because Macs used to have California Scheduling: "hey man, i'm like, not using these cycles, so why don't you go ahead an run for a while..". Now they have New Jersey Sceduling: "YOU! Outta da way! NOW!" Wierdly enough, while the first sounds better in theory, the other works much more effectively in practice.
I remember reading an interview in a business mag about 10 years ago, when the Sparcs were lagging everyone and IBM/HP were gaining market share. His comment was along the lines of, "yeah, everyone's waiting at the end of the runway for us, and not because they think we're going to take off." Then they did take off. I'll admit having thought, "well, they're day is done this time", more than once, and each time they come back. If Scott goes, I hope that attitude of fighting on stays.
And having run both Linux and Solaris in a server environment, here's to figuring out how to push Solarisx86 everywhere. If I ran Opterons, because of the Dev Tools being available, I'd certainly consider running it on a cluster just for the management and profiling tools.
The utilitarian style was part of the Thinkpad's appeal. It reminded you that the money had been spent on the durability, not the flash. I have (though it's been officially retired), an A22m, which still runs, despite being dropped on ice (I went down hard and got a large bruise that lasted weeks. The thinkpad made a sickening crack sound, and didn't even have cosmetic damage). It weighs over 7 lbs, has the high-res screen but only 8m video card, but is amazingly durable and has the nicest keyboard you could ask for. You can type all day on that keyboard without getting wrist-strain. My wife's 600E, despite being older yet, is still bright and clear, with an excellent keyboard, despite being a compact model. Thinkpads just feel good, and have the additional advantage of generally not wearing out prematurely.
Thinkpads were popular because they were built to last, and built to survive the Road Warrior breed of business traveler. The fear with Lenovo is probably that they're pushing into the consumer space, and are building consumer-grade laptops. With IBM, you came to expect stolid efficiency, and Lenovo is going to have to demonstrate a commitment to that as well.
As an aside, I had always hoped that pre-intel-switch, Apple would partner with IBM (ThinkPower? MacPad?), so that you could get an OS-X laptop with IBM build quality. Probably would have come in "His Steveness Sweater Black", just to differentiate it from the regular line-up.
Thankfully the only time I have seen a Token-Ring card was from a Dell screw-up, circa 2001. We needed a computer to control an instrument, which had sensibly been converted to ethernet. The researcher in charge called Dell for a workstation, and said he needed two network cards. Dell convinced him that they couldn't be both of the same type, because you couldn't tell them apart. So, they put in two different cards; a 3Com ethernet, and a Token-Ring.
Just... speechless... when we realized what they'd done. They were also kind of huffy about taking it back. Maybe we accidentally bought their last one, and they thought they could close out that part number for good.
Let's not get too nostalgic. CHRP partly died due to Apple, since they insisted on doing things slightly differently from the standard. We were looking forward to CHRP OS/2 machines, (*sigh*), which quickly materialized and dematerialized almost simultaneously. That would have been a great time for IBM to buy a sufficiently large stake in Apple to make them behave, then offer PPC machines with the MacOS:OS/2:AIX option. Most probably OS/2 would have still faded away, and we would have had cheapish machines (by IBM standards) running MacOS on the desktop, with a clear upgrade path to AIX-based Power machines when you finally needed the Big Iron.
Or we'd still have OS-X, except instead of BSD underpinnings, it would have AIX. One can only imagine what the Aqua version of SMIT would have been like.
Having a spare goat for the troll this morning....
Too keep your linux chops up in case you want a job at a Linux-based organization? Because there is always some app that's not quite happy with the BSDish nature of OS-X, and won't quite build correctly? Because it's a precompiled binary that isn't going to run on OS-X no matter how many black chickens you sacrifice by the dark of the moon?
And, of course, just because you can. It's better than booting Windows on it; at least you don't feel dirty afterwards.
That depends; you're sitting at home and are your own tech support, then you're right. Go wherever you want.
You're on some sort of managed network, such as a business or university system, which is networked intimately to many other systems and has administrators who will have to clean up the machine when you're done, then your browser should prevent you from going to a known malware site. Sorry, I used to have that job, and began thinking, "maybe if I make them all use Lynx and Pine for web and email, this nonsense will stop".
Like most Microsoft innovations, this one is pitched to the home user, to build mindshare, but is really aimed at corporate environments. Don't be surprised if you work at United that if you type http://www.untied.com/, you somehow end up back at United.com. Similarly, small dot-coms might redirect F*d Company's website to wemotivate.com.
That's why you should always homebrew your own. Friends did this, and managed to make a beer that had a decent taste while being consumed, but in about 10 minutes, you would start to fill billious and get a splitting headache. It was the perfect deterrent, Sober to Hangover without ever passing through Drunk. Kind of like Iron City Light, but quicker and with less water retention.
Having grown up United Methodist (communion is done with Welche's Grape Juice), and converted to Episcopalianism (communion is done with Wine), I can authoritatively say that given the choice between Concord Grape, and Fermented _any_ Grape, I'll take the wine. I would still take the wine even if I couldn't get a buzz off it, because at least the yeasts would have done something about all that accursed Sugar in the grape juice before it made it to my glass. Pity that dealcoholized wine is generally pretty vile stuff; with some care, it would be a great drink.
Wine is always the right answer to the question, "what have the Romans ever done for us?"
We've already spent the money on an immensely effective ABM system; it's called MAD. It's the solid belief that if you nuke anything that we value, we're going to make the rubble bounce from one end of your "now a historical question on Jeopardy!" ex-country to the other. It's why we fought only proxy wars after 1945, and didn't, despite the urges to the contrary, use tactical nukes in Korea, Vietnam, Cuba, or any other number of flash-points. This is actually the best incentive certain nominally middle-eastern states have to keep a lid on their stockpiles. It's probable that even if they don't nuke one of our cities, if it happens they're going to get blamed and made an example of. Basically, we need the old Kissinger/Nixon team at the helm. One of them negotiates with the other power, and says that the President is crazy, and is going to bomb them, then the president acts crazy (and in the case of Nixon and the North Vietnamese), does bomb them, but just enough to get them back to the negotiating table. Same deal here; rail against the imperialist foreign hedonistic infidel running dogs all you want, but don't get too antsy, or bad things happen. None of the ABM systems we've tested so far has been nearly as successful as good old psychology and enlightened self-interest. We'd be better off spending the money on a couple of underground tests capable of rattling seismometers on the other side of the world and more money for scanning cargo containers as they enter port than on jumbo-jet mounted lasers, missile batteries that have to be told where the missile is coming from, or the rest of the Reagan-era Sci-Fi retreads we are now.
That's no moon! It's an enormous Federal Boondoggle!
Probably Power5. However, to use desktop processors as the example: I have some benchmarks of a code I've run on machines from a DEC Alpha 21064 through a current G5 2.0.. The timings that matter here are jobs that took 107 seconds on a PII/400 now take 7. The speedup is about constant across the series, so we're looking at a 15x speed-up on my current code versus a PII/400. Since this particular package runs at about the same speed on comparable clock-rate Opterons/G5s, we can infer a probable 15x speedup (30x if they're using PII/266, with the slower bus), for less than twice the power. (250W vs 400W). OTOH, that 400W PS powers two processors, which claim to pull in the >40W range when under load.
In short, as long as you avoid Itanium-2 and certain late P4-Xeons, any modern machine will deliver a pretty acceptable flop/watt rating. This, regrettably, means that while their heart is in the right place, Berkeley's stunt, even if using biodesiel as fuel, would be more efficient if they scrapped the old computers, and used the money to buy a current Mac Mini or two to offer as compute servers. In my last job, even with the institution supplying power as 'overhead', certain machines no longer delivered enough performance to justify them occupying a plug and 9 sq ft of floor space, even if they did work fine. If I had been directly billed for what they consumed versus what they produced, they would have been dumpster-bound years earlier.
This doesn't mean that they shouldn't refurb the old machines, stick an OS and some user-land software on them, and hand them (or sell them cheaply) to potential users for desktop usage, but as a cluster, they're not likely to earn their keep.
Actually, it's a pretty straightforward result. They took old work by Chad Mirkin and Coworkers (http://www.chem.northwestern.edu/~mkngrp/Group%20 Research.htm/), and the use of sugars to solubilize gold for therapeutic uses http://images.google.com/images?q=auranofin&hl=en& btnG=Search+Images/, then changed the receptors. Straightforward applied engineering from known principles and substances. It's only novel if you're a science journalist with space to fill who doesn't follow the chemical literature.