Remember the Iowa and New Jersey had been upgraded with modern ECM and missile systems; they still had most of the old 16" guns, but they were otherwise modern heavy naval platforms. Those date to pre-1945, so they were 40 year old platforms when returned to duty.
When they brought the NJ back the first time, it was partly because of the image it cuts (gunboat diplomacy still has its place, it would seem), partly the 16" guns, which leave a pit 50 some feet across, and partly that the old battleship steel plate is pretty much impervious to Exocets.
On the other hand, smaller, lighter, ships, do tend to get replaced more quickly. I saw the last of the heavy cruisers being scrapped at Philly Naval Yard years ago (before the yard went by the boards as well), and I get the impression from Naval relatives that most of the light ships are less than 30 years old. Nuke carriers and ballistic submarines are probably a different category, but if the basic frame is sound, and there's room to to upgrade the electronics, there is no reason that a well-built and maintained craft shouldn't have a 30-40 year life span.
Now, whether that applies as well to space-shuttles as it does to B52s and 727s is a different question.
When the Buran project successfully flew, the Wall Street Journal Editorial Page (If I have my raving nuts correct) ran an editorial claiming that the Buran was too small to be used for human cargo, but would make a great orbitally delivered, targetable, nuclear bomb delivery system.
Kind of makes you wonder if the Birchers will be out in force this time as well, claiming we need to militarize space before those Euros get us with their 'reusable launch vehicles'.
If you plug "silver antibacterial" into Google Scholar, you'll get enough reading to entertain you for a while. Basically, it's a heavy metal that binds with sulfide groups, inhibits lipid biosynthesis, and may gum up the DNA replication network.
Silver particles on activated carbon are used for water purification, silver nitrate drops in infants' eyes to prevent infection, and silver sulfadiazene for burns.
Seriously enough, the germs can evolve a resistance to silver, just as various protozoa are becoming resistant to antimony and arsenic-based therapeutics. The main mechanism is over-expression of glutathione, which complexes the offending metal ion and removes it from the cell. Metal sequestration machinery is pretty good in most organisms, so resistance to silver should be expected given the number of products relying on silver particles for activity
Re:information is shared smoothly and intelligentl
on
Is This the Holodeck?
·
· Score: 4, Insightful
On the other hand, this is not your Father's (or at least younger uncle's ) Japan. Japanese research, as based upon checking the names in the journals I read, has been becoming more innovative and forward thinking in the last 15 years. They've also shown no fear towards hiring American researchers who have been 'right-sized' by short-sighted native corporations (NEC picked up a lot of researchers from various NJ chemical and materials firms). As RAH once put it, "there is intelligent life in Tokyo."
This probably won't produce a holodeck, but given CAVE technology (works), haptics (works), and some of the newer innovations such as microvoltage stimulation for motion effects, you may actually get something this time, even if it won't be the Trek 23rd century version.
I personally look forward to virtual clothes shopping, at least for other people whom I generally have to follow around and critique during the process.
On the other hand, I read not only RAH and Co, with their optimistic vision of pioneers expanding out until earth was a forgotten myth, but also "A Canticle for Liebowitz", and early William Gibson short stories. Look at the bright side; at least we haven't (1) completely blown ourselves up, or (2) ended up like The Sprawl, with the Soviets and ourselves flat on our backs from a military dust-up, and the world run by Multinational Corporations and the Japanese.
While we don't have friendly and helpful AI's (with libertarian tendencies) like Mike, we aren't populating the astronaut program with people who can memorize log tables to do orbital mechanics in their head. To a certain extent, we decided to study inner space instead, put our money there, and made strides in materials science and computing that were truly out there, even by Sci. Fi. standards.
I'll admit; I had hoped to see NASA (or JSA) revive the Nerva program by now, and see routine cruising to local planets. I had truly hoped that someone would get ambitious and launch an Orion to Alpha Centauri, just to show it could be done. However, in the end, it's not such a bad future, especially given that certain givens in the books we grew up with are probably impossible to highly impractical (transporters and FTL travel).
Most worrisome is that Andre Norton, with her visions of people escaping a society that had agressively turned its back on modernity, or Heinlein's "Revolt in 2100" may be the way we're going instead. Exploration leads to new ideas, and the abandonment of old, comfortable, biases. Better to stay home and hide under the bed.
Of course, they could always have actually fixed the power-consumption issues with the IA-64, attached an x86 emulation unit, and decided to move everyone to 64-bits, ready or not. They'd leave AMD making compatibles with old Intel designs, while Intel trumpets the Bright Shiny Made by Intel(tm) Future.
Regrettably, it's probably Yet Another Damned PIII/IV (YDAPv3/YDAPv4) with underwhelming floating-point performance aimed squarely at the home user and business drone market.
The issue is that most people aren't using 35mm in the studio in the first place, and if you believe Eastman Kodak, a hand-held shot on normal film, under optimal conditions, scanned to a usable 2000x3000 pixel image; 6MPel. Anything higher, and you scan grain, but no greater usable information. I.e. the 35mm frame may have more Data, but not more Information. Tripod-mounted, fine-grain, and you might get 12mp, but you still have the grain effects to deal with. At that point, you can either shrug, and go with the digital work-flow, or work out in the gym, and move up to medium or large format.
I saw a show at the Chicago Botanical Gardens last year of people's home environments from around the world, and the pictures were a mixture of 35mm and (presumably) 6x6cm. The 35mm, blown up to 14x22, was grainy and unsharp. The 6x6 at 20" (approx) square, was grainless with good tonality and sharpness. This is the same at home, where I had a Velvia transparency from 6x6 drum scanned and printed to 16x16; no grain, great detail, wonderful tones, and better than any 35mm I've tried to enlarge to that size, including the much lamented Kodachrome 25. On the other hand, I've also seen prints from a 6mp dslr, and while the detail isn't there when printed 14x22 if you press your nose to the print, the edge sharpness and lack of grain make it a more pleasing image, which appears to be of equal or better quality than 35mm
As much as I respect your friends' opinions, I've seen examples where 6mp DSLRs print as well as scanned 35 up to 8x10, and 8mp prints little larger. There may be more information in the 35mm frame, but it doesn't survive the transition to printed output. Film is a good technology, and there is nothing wrong with saying, "I'm used to it, and I'm not going to change unless forced", but the quality argument on digital is already over. If you want better than 6-12MP DSLR's offer, then it's time to pick up a 6x6, 6x7, or large-format (8x10 transparencies are a delight to look at), and shoot big. Unfortunately, sales of those sizes are dropping even faster than 35mm.
Look, this may just be the whining of some wound-up fashionista out of his element, but there is a kernel of truth in his complaint. While you *can* hack up some home-grown management tool to solve your problem, generally it's not a problem unique to you. What you (that's you as a business, not you the slashdot reader) want is a set of consistent management tools that people you hire have seen before, and that someone else maintains. You provide money, they provide software, and your business machines quietly hum along.
This is where companies such as IBM, Novell, and RedHat come in. They're willing to pay for people to develop consistent tools to make life easier in corporate environments. It's not that the OSS community can't, (look at the web-based integration of Ganglia in Rocks, which allows you to visually scan the state of your entire cluster), but that it often won't. I used SuSE for years, and it's only recently I've seen anyone come up with tools that worked as elegantly as YaST. It doesn't seem to be a priority, and that's a problem for busines.
Hopefully OpenSolaris will begin to catch on, and some cross-fertilization will begin. Consistency is not a flaw, and too much choice is not a virtue.
Actually, well over. When I was a kid 30-odd years ago, I had some time-life books with a picture of flowers in UV, showing how the common buttercup wasn't uniformly yellow, but had dark patches on the petals in UV. I'm surprised that nobody spent the time to photograph some other common things in UV, just to see.
Is this just a case of the difficulty of UV photography (quartz lenses, and expensive filters), or didn't anybody think it was worth the time?
Ubuntu. I installed it on a PIII-933 recently, and the setup was flawless. Found the network card, video card, monitor, and configured everything pretty much correctly (monitor had to be tweaked to higher res) without any fuss on my part. The default theme and eye-candy are very pleasing as well. Has nothing to do with performance, but it does give it a pleasantly polished feel when you see it start running.
Alternately, I'm going to recommend the non-Slashdot option: CentOS, very similar to a certain OS that goes by the initials RHEL. No, it's not sexy, but it's supported by everyone. This may not matter for your apps, but it is the default when someone commercial (Intel's compilers, software from PNNL) thinks "Linux". The current install and hardware detection are also very solid, and if you need more Chrome and home-user apps, they're easy enough to find in pre-compiled RPMs. (or just build them yourself).
About 10 years ago (and if it's a lot better there now, sorry for outdated information), the NY Times had two articles summing up the new Russia.
One was on business practices, with the comment, "to enforce a contract, you often have to take out a contract".
The other was on a clinic doing heroin detox. The basic system was thugs would patrol the streets, find people doing heroin, club them into submission, drag them to the clinic and chain them to a bed, and then let them dry out cold turkey. The Doctor in charge said, "of course this is not the optimal treatment, but here... ".
Looks like our spammer's fate falls under one of those two categories of "solutions". As others have said, it probably wasn't the spam, it was, "just business".
Selective pressure is how we breed various domestic animals. It's simply shorthand for saying, "desirable ones get to breed, and the rest are terminated and/or eaten". Natural examples would be long-term desertification (animals and plants that have genes causing them to be more efficient with water live long enough to breed), or salination (if you have genes which allow you to sequester, transport, or otherwise control salt, then you'll survive long enough to breed and pass on those genes).
Several years ago, as the SU was breaking up, it was reported that a breeder in their fur industry had been selecting for docile foxes. Over a period of about 40 years, he had not only achieved the desired trait, but in the process the animals snouts had shortened, their ears flopped, and their tails had acquired a curve. These other traits, which had been linked, but not actively selected for, are those which help distinguish domestic dogs from wild dogs.
Starting in the 1920s the Heck brothers in Germany had tried the reverse with trying to undomesticate european cattle, in an attempt to bring back the Auroch. The result, the Heck Cattle, do have some resemblence to the Auroch, but this is probably coincidental, as the modern european stock they started from descends from the middle-east, and did not (apparently) interbreed much with the Auroch. They still got a wild-looking bovine with many of the appropriate traits.
So, you can apply positive pressure (anything that looks like an Auroch gets to breed) or negative (anything that looks like a grand piano plus a few rack of billiard balls gets shot), but in the end it is simply that the expressed and perpetuated genome in the population is dependent upon those circumstances related to its being successfully transmitted. It's still evolution, but for once anthropomorphizing the source of the selective pressure is correct; someone made a decision about what kind of environment those animals would live in, but they made no conscious decision on how to respond. Nobody gets up one morning and says, "i'm going to evolve out of this elephantine lifestyle today".
My past experience was that you tend to think in terms of queues; you (physically) queue up for the keypunch, submit your job to a queue, and go find the appropriate print queue the data came back off from. The other experience has always been (Unicos, VM/XA or VM/CMS systems) that the software environment is antiquated to the point where you're encouraged to do as much off-line as possible. Get in, do work, get out.
The computer, therefore, is a far more abstract, remote, and non-interactive object. I remember what an unpleasant change moving from an 11/785 running VMS to a 3090/VF running VM/CMS. The programming tools were arcane, the OS didn't even have subdirectories (it did have minidisks; i.e. your own virtual pile of floppy disks), and the editors definitely underwhelming. On the other hand, like a VAX/VMS system, the queueing system was an integral part of the OS, so it worked smoothly, as opposed to the unix solutions that are bolted on the side as an after thought. Sun Grid Engine and LSF are pretty close to the old VAX queues, but still not quite as well integrated.
This, of course, is not entirely fair, as large VMS systems were used like mainframes, but still had good tools and a friendly user environment. In the end, think of it generally as a tightly-regulated, non-interactive environment. It's the kind of environment for utter reliability, where it's primarily computers talking to other computers.
Actually, *The Market* did not determine the fate of Nuclear Power, an underinformed public, led by demagogues on both sides, couldn't tell Reactors from Bombs and therefore litigated the industry into submission. Before that, the type of reactors to be used was determined in part by military desires (capable of producing plutonium), not by economic factors (thorium, for instance, which is more plentiful but doesn't produce divertable byproducts).
A class like this could be very valuable, if it trained those people likely to end up making decisions (humanities and business majors) in the actual science behind the technology, or the technologists in how to present to the unschooled what they're actually doing.
Remember, DuPont used to boldly proclaim "Better Things for Better Living Through Chemistry". That's still true, but no modern Ad agency would dare say that for fear of reminding the undereducated that the world is made out of Chemicals. Courses that attempt to prevent that sort of dichotomy from occuring with Nanotech, etc, are frankly a good thing, as long as they're not led by the fear-mongering Rifkins of the world.
Not to go pop culture here, (but this is Slashdot), but I'd rather take my chances with the technology and live in the Blade Runner future, than in the unheated, unhygenic, Arthurian Agrarian past.
Having managed a mixed shop once upon a time (RHEL 3.0, FC 2, SuSE 8.1, 1 Debian just because), I can say that I'd like to see a general sanitization of filesystem layout and startup scripts. The whole multiple desktop issue I care less about, as in the end people sort themselves into their own environments (Gnome, KDE, TWM), and tend to support each other. Remembering where things are stored to keep the system running, distribution to distribution, is a pain.
Frankly, for all I care they could all agree to use OpenSMIT for systems administration, and I'd agree as long as all of the majors used it.
Actually, for years the FP performance of the Intel processors lagged, while the integer plowed ahead. That's partly why the Itanium was such a shock; truly excellent floating point performance (pity the integer sucked rocks through a straw). For most codes, which mixed a fair amount of array handling with the FP code, you still got good performance, even though the FP could have been better. The aggregate mix of instructions helped keep the IA32 boxes competitive. In 2000, our Athlon-800/PIII-1GHz were already ahead of the available Sun or SGI boxes, and rapidly catching up with the low-end (i.e. relatively affordable) AIX systems. With the IA-64, Intel caught up on the high end, and some of that technology has (or will) certainly made it to the EM64-T. Translation: I doubt that there will be any serious performance differences between what would have been the G5 of 2007 and the Intel-??? that they'll actually use.
Currently, we have G5s here, because clock for clock they're as fast as the Opterons, but easier to port the codes we use to. (you basically take the AIX code, XLF for MacOS, and hope the case-insensitivity doesn't bite you). For other programs, Gromacs for instance, the P4 is the clear leader, because of time spent on hand-written assembly loops. Current Intels are not bad processors, though they could certainly turn down the power consumption a bit.
In short, while I will be sorry to lose XLF/XLC (as I was sorry to lose DEC Fortran when the Alphas went away), IFort/ICC are capable of generating good code on current Intel processors, so I doubt we'll lose much on that score. If Apple uses EM64T, then I would expect PathScale to make their compiler available as well, and it is head and shoulders above the other alternatives for AMD-64/EM64T. The real question on the servers is whether two years out we're looking at 64-bit, dual-core, EM64T based systems, or the new dual-core, low-power, Itanium for the XServes and high-end PowerMacs.
But what's the margin on a chip for a game console, versus a chip for a mid-range workstation?
One of the problems is that the new IBM, while in business, is not the visionary old IBM. Apple may have been a small market, but a two-tiered Linux approach (buy dual PPC-970s from the Californians, then graduate to Real Power-5 from us) would have helped displace Intel/AMD from many current markets. Given IBM's service focus, they could have pushed it as well as part of an integrated, end-to-end (yack, business speak) solution.
OTOH, maybe the processors for the XBox, et al., are easier to manufacture, or more tolerant of defects, in which case that would make up for the (probably) lower margins. Getting rid of His Steveness is just the cherry on top. (and I'd love to see how that gets written up in the annual report. "jettisoned high-strung customer...")
We will note that the Power Consumption of the Pentium-D is pretty much identical to the first generation Itanium-2. Those chips had a power/heat envelope that convinced us to stop buying them, except for specialized applications, because of the cooling issues.
Not to ask the obligatory Non-Intel question, but how do those numbers compare against G5 (PPC-970) or current IA-64?
Actually, this seems to be halfway to the Right Thing; you get rid of a non-portable extension to Java (good), but encourage people to use a minor (though clean) language for native-mode programming (iffy).
I've always wondered why that piece of NeXT heritage stuck around, as I'd be willing to wager a doughnut (pick your favorite pusher) that there are more Ada-programmers around than Objective-C gurus. (translation: take a deep breath, wave goodbye, and move to C++ or whatever the modern alternative compiled language would be, or at least make sure the library interfaces are correct for both languages).
Actually, I'd be just as happy, since they have the entire GNU compiler collection, is to simply define a uniform calling convention for their libraries from all languages, and stop worrying abou the One True Language. DEC did it back in the 80s with the Common Language Interface, which allowed me to call the same function (with pretty much the same syntax) from Fortran, C, or Pascal. I presume today that would mean using SWIG to generate a common set of wrappers, and remembering that there are Fortran, Ada, and C++ programmers out there somewhere when generating them.
Maybe my memory is fuzzy, but it seemed simpler to call those functions than on Unix systems where I always have to worry about the library interface on a language by language basis (one underscore or two; case-sensitive or not, etc.) I would like that level of transparency back.
It's not popular because previous generations were Hot, Expensive, and Hard to Program. Adoption was slowed mainly by #2, as it was a significant investment for many research groups to even put one on the floor for testing purposes. The Hard to Program meant that you had wonky versions of Linux, or HP-UX with new compilers, as your OS options, which just increased the resistance.
I ran an Itanium-2 cluster (and had briefly an Itanium-1 loaner) and if the compilers were stable, the first generation easily outpaced the UltraSparc-III systems of the era, and the Itanium-2s kept pace or outpaced the Alphas. However, their power-draw was unbelievable, you had to rebuild the GNU toolchain from RedHat stock, then install whatever that week's version of Intel's C/Fortran compiler was, and on any code that was integer bound, they were barely faster than the late Xeons.
Ah, but for floating point, nothing touched them, except possibly the late Power systems that we couldn't afford. A clean 64-bit architecture, and four floats per clock cycle, meant that for a certain class of problems, there were the cost-effective solution; US-III prices, and twice (at least) the performance. They were even pretty easy to port our codes to, as their pure 64-bit environment (8-byte Integers and pointers), meant that generally modifications to run on Alphas and Crays would run on the Itaniums as well. We would have preferred a next generation of Alpha, but the Itaniums work quite well for what they do.
Discouragingly, now for most (but not all) of the simulations I do, AMD-64 or even EM64T systems are cost-effective, and just as fast. With this new price reduction, I may have to look into adding IA-64s back into the mix, though they're going to have trouble competing against IBM's linux-based Power-5 systems, which offer similar prices, and a more well-known and supported architecture.
Having programmed on them, and managed the PowerChallenge and Origins, I'm not really as sad to see them go. The 4D series with NeWS as its desktop in the late 80s/early 90s was beyond nifty, and the first R8000s we saw (with the 64-bit Fortran compilers that didn't require us to talk to DEC reps!) were also very neat. I'm still entertained by Electropaint when I pass an SGI. That demo (plus flight simulator) probably sold more boxes during their heydey than they'll ever admit.
However, by the time we were in the middle 90s, the programming languages were basically bug of the month, the n32 addressing on the R5000 was creating headaches for porting code, and the price/performance was dreadful. They still built some nice machines (thinking of an O200 w/ Craylink that I still miss), but they were getting harder to justify versus IBM Power systems on the high-end, and PCs on the low end.
The killer for us was when we benchmarked a major simulation package on an O2 with SGI 7.1 compilers versus a PII-400 + G77, and the PC walked away from the SGI. We bought our last machines, warned people that they were transitionary and the porting of code to modern architectures should start now.
I'll lift my SGI mug to them in respect, but will not mourn their passing as I did that of DEC.
Have you looked at Thomas Ray's Tierra simulator? I saw results from that in the early 90s when he was at U. Delaware, and he got quite complex behaviour from a few simple rules, and a single parent organism.
One of the interesting byproducts of Tierra was that one of the first organisms to evolve was a more optimzed version (fewer instructions) of his hand made one, which then went on to give rise to parasites, anti-parasites, predators, etc.
It *is* the right size... Any chance someone can convince His Steveness to add a "phaser" functionality to the next iPod? I'd settle for Stun.
On a more serious note, the clean look of the iMacs, Mini, and XServes has been valuable for PR around here. The reaction is much similar to that of non-techies facing an old Refrigerator-sized VAX than a pile of beige PCs.
Thank you. It's been a long time since I looked into VAX emulation. I had considered ordering an OpenVMS kit for our Itania, but just couldn't justify it.
Regrettably, I haven't handled a VMS system since the Crystallographers went to Silicon Graphics, and thence to NT.
Slashdot Mirror
Remember the Iowa and New Jersey had been upgraded with modern ECM and missile systems; they still had most of the old 16" guns, but they were otherwise modern heavy naval platforms. Those date to pre-1945, so they were 40 year old platforms when returned to duty.
When they brought the NJ back the first time, it was partly because of the image it cuts (gunboat diplomacy still has its place, it would seem), partly the 16" guns, which leave a pit 50 some feet across, and partly that the old battleship steel plate is pretty much impervious to Exocets.
On the other hand, smaller, lighter, ships, do tend to get replaced more quickly. I saw the last of the heavy cruisers being scrapped at Philly Naval Yard years ago (before the yard went by the boards as well), and I get the impression from Naval relatives that most of the light ships are less than 30 years old. Nuke carriers and ballistic submarines are probably a different category, but if the basic frame is sound, and there's room to to upgrade the electronics, there is no reason that a well-built and maintained craft shouldn't have a 30-40 year life span.
Now, whether that applies as well to space-shuttles as it does to B52s and 727s is a different question.
When the Buran project successfully flew, the Wall Street Journal Editorial Page (If I have my raving nuts correct) ran an editorial claiming that the Buran was too small to be used for human cargo, but would make a great orbitally delivered, targetable, nuclear bomb delivery system.
Kind of makes you wonder if the Birchers will be out in force this time as well, claiming we need to militarize space before those Euros get us with their 'reusable launch vehicles'.
If you plug "silver antibacterial" into Google Scholar, you'll get enough reading to entertain you for a while. Basically, it's a heavy metal that binds with sulfide groups, inhibits lipid biosynthesis, and may gum up the DNA replication network.
Silver particles on activated carbon are used for water purification, silver nitrate drops in infants' eyes to prevent infection, and silver sulfadiazene for burns.
Seriously enough, the germs can evolve a resistance to silver, just as various protozoa are becoming resistant to antimony and arsenic-based therapeutics. The main mechanism is over-expression of glutathione, which complexes the offending metal ion and removes it from the cell. Metal sequestration machinery is pretty good in most organisms, so resistance to silver should be expected given the number of products relying on silver particles for activity
On the other hand, this is not your Father's (or at least younger uncle's ) Japan. Japanese research, as based upon checking the names in the journals I read, has been becoming more innovative and forward thinking in the last 15 years. They've also shown no fear towards hiring American researchers who have been 'right-sized' by short-sighted native corporations (NEC picked up a lot of researchers from various NJ chemical and materials firms). As RAH once put it, "there is intelligent life in Tokyo."
This probably won't produce a holodeck, but given CAVE technology (works), haptics (works), and some of the newer innovations such as microvoltage stimulation for motion effects, you may actually get something this time, even if it won't be the Trek 23rd century version.
I personally look forward to virtual clothes shopping, at least for other people whom I generally have to follow around and critique during the process.
I would have said Colorado, since that's where you go to get your Truck Warshed.
"Warch, Warch, the 17th of March, good buddy"
On the other hand, I read not only RAH and Co, with their optimistic vision of pioneers expanding out until earth was a forgotten myth, but also "A Canticle for Liebowitz", and early William Gibson short stories. Look at the bright side; at least we haven't (1) completely blown ourselves up, or (2) ended up like The Sprawl, with the Soviets and ourselves flat on our backs from a military dust-up, and the world run by Multinational Corporations and the Japanese.
While we don't have friendly and helpful AI's (with libertarian tendencies) like Mike, we aren't populating the astronaut program with people who can memorize log tables to do orbital mechanics in their head. To a certain extent, we decided to study inner space instead, put our money there, and made strides in materials science and computing that were truly out there, even by Sci. Fi. standards.
I'll admit; I had hoped to see NASA (or JSA) revive the Nerva program by now, and see routine cruising to local planets. I had truly hoped that someone would get ambitious and launch an Orion to Alpha Centauri, just to show it could be done. However, in the end, it's not such a bad future, especially given that certain givens in the books we grew up with are probably impossible to highly impractical (transporters and FTL travel).
Most worrisome is that Andre Norton, with her visions of people escaping a society that had agressively turned its back on modernity, or Heinlein's "Revolt in 2100" may be the way we're going instead. Exploration leads to new ideas, and the abandonment of old, comfortable, biases. Better to stay home and hide under the bed.
Of course, they could always have actually fixed the power-consumption issues with the IA-64, attached an x86 emulation unit, and decided to move everyone to 64-bits, ready or not. They'd leave AMD making compatibles with old Intel designs, while Intel trumpets the Bright Shiny Made by Intel(tm) Future.
Regrettably, it's probably Yet Another Damned PIII/IV (YDAPv3/YDAPv4) with underwhelming floating-point performance aimed squarely at the home user and business drone market.
The issue is that most people aren't using 35mm in the studio in the first place, and if you believe Eastman Kodak, a hand-held shot on normal film, under optimal conditions, scanned to a usable 2000x3000 pixel image; 6MPel. Anything higher, and you scan grain, but no greater usable information. I.e. the 35mm frame may have more Data, but not more Information. Tripod-mounted, fine-grain, and you might get 12mp, but you still have the grain effects to deal with. At that point, you can either shrug, and go with the digital work-flow, or work out in the gym, and move up to medium or large format.
I saw a show at the Chicago Botanical Gardens last year of people's home environments from around the world, and the pictures were a mixture of 35mm and (presumably) 6x6cm. The 35mm, blown up to 14x22, was grainy and unsharp. The 6x6 at 20" (approx) square, was grainless with good tonality and sharpness. This is the same at home, where I had a Velvia transparency from 6x6 drum scanned and printed to 16x16; no grain, great detail, wonderful tones, and better than any 35mm I've tried to enlarge to that size, including the much lamented Kodachrome 25. On the other hand, I've also seen prints from a 6mp dslr, and while the detail isn't there when printed 14x22 if you press your nose to the print, the edge sharpness and lack of grain make it a more pleasing image, which appears to be of equal or better quality than 35mm
As much as I respect your friends' opinions, I've seen examples where 6mp DSLRs print as well as scanned 35 up to 8x10, and 8mp prints little larger. There may be more information in the 35mm frame, but it doesn't survive the transition to printed output. Film is a good technology, and there is nothing wrong with saying, "I'm used to it, and I'm not going to change unless forced", but the quality argument on digital is already over. If you want better than 6-12MP DSLR's offer, then it's time to pick up a 6x6, 6x7, or large-format (8x10 transparencies are a delight to look at), and shoot big. Unfortunately, sales of those sizes are dropping even faster than 35mm.
Look, this may just be the whining of some wound-up fashionista out of his element, but there is a kernel of truth in his complaint. While you *can* hack up some home-grown management tool to solve your problem, generally it's not a problem unique to you. What you (that's you as a business, not you the slashdot reader) want is a set of consistent management tools that people you hire have seen before, and that someone else maintains. You provide money, they provide software, and your business machines quietly hum along.
This is where companies such as IBM, Novell, and RedHat come in. They're willing to pay for people to develop consistent tools to make life easier in corporate environments. It's not that the OSS community can't, (look at the web-based integration of Ganglia in Rocks, which allows you to visually scan the state of your entire cluster), but that it often won't. I used SuSE for years, and it's only recently I've seen anyone come up with tools that worked as elegantly as YaST. It doesn't seem to be a priority, and that's a problem for busines.
Hopefully OpenSolaris will begin to catch on, and some cross-fertilization will begin. Consistency is not a flaw, and too much choice is not a virtue.
Actually, well over. When I was a kid 30-odd years ago, I had some time-life books with a picture of flowers in UV, showing how the common buttercup wasn't uniformly yellow, but had dark patches on the petals in UV. I'm surprised that nobody spent the time to photograph some other common things in UV, just to see.
Is this just a case of the difficulty of UV photography (quartz lenses, and expensive filters), or didn't anybody think it was worth the time?
Ubuntu. I installed it on a PIII-933 recently, and the setup was flawless. Found the network card, video card, monitor, and configured everything pretty much correctly (monitor had to be tweaked to higher res) without any fuss on my part. The default theme and eye-candy are very pleasing as well. Has nothing to do with performance, but it does give it a pleasantly polished feel when you see it start running.
Alternately, I'm going to recommend the non-Slashdot option: CentOS, very similar to a certain OS that goes by the initials RHEL. No, it's not sexy, but it's supported by everyone. This may not matter for your apps, but it is the default when someone commercial (Intel's compilers, software from PNNL) thinks "Linux". The current install and hardware detection are also very solid, and if you need more Chrome and home-user apps, they're easy enough to find in pre-compiled RPMs. (or just build them yourself).
Another day on the froniers of capitalism.
... ".
About 10 years ago (and if it's a lot better there now, sorry for outdated information), the NY Times had two articles summing up the new Russia.
One was on business practices, with the comment, "to enforce a contract, you often have to take out a contract".
The other was on a clinic doing heroin detox. The basic system was thugs would patrol the streets, find people doing heroin, club them into submission, drag them to the clinic and chain them to a bed, and then let them dry out cold turkey. The Doctor in charge said, "of course this is not the optimal treatment, but here
Looks like our spammer's fate falls under one of those two categories of "solutions". As others have said, it probably wasn't the spam, it was, "just business".
Selective pressure is how we breed various domestic animals. It's simply shorthand for saying, "desirable ones get to breed, and the rest are terminated and/or eaten". Natural examples would be long-term desertification (animals and plants that have genes causing them to be more efficient with water live long enough to breed), or salination (if you have genes which allow you to sequester, transport, or otherwise control salt, then you'll survive long enough to breed and pass on those genes).
Several years ago, as the SU was breaking up, it was reported that a breeder in their fur industry had been selecting for docile foxes. Over a period of about 40 years, he had not only achieved the desired trait, but in the process the animals snouts had shortened, their ears flopped, and their tails had acquired a curve. These other traits, which had been linked, but not actively selected for, are those which help distinguish domestic dogs from wild dogs.
Starting in the 1920s the Heck brothers in Germany had tried the reverse with trying to undomesticate european cattle, in an attempt to bring back the Auroch. The result, the Heck Cattle, do have some resemblence to the Auroch, but this is probably coincidental, as the modern european stock they started from descends from the middle-east, and did not (apparently) interbreed much with the Auroch. They still got a wild-looking bovine with many of the appropriate traits.
So, you can apply positive pressure (anything that looks like an Auroch gets to breed) or negative (anything that looks like a grand piano plus a few rack of billiard balls gets shot), but in the end it is simply that the expressed and perpetuated genome in the population is dependent upon those circumstances related to its being successfully transmitted. It's still evolution, but for once anthropomorphizing the source of the selective pressure is correct; someone made a decision about what kind of environment those animals would live in, but they made no conscious decision on how to respond. Nobody gets up one morning and says, "i'm going to evolve out of this elephantine lifestyle today".
My past experience was that you tend to think in terms of queues; you (physically) queue up for the keypunch, submit your job to a queue, and go find the appropriate print queue the data came back off from. The other experience has always been (Unicos, VM/XA or VM/CMS systems) that the software environment is antiquated to the point where you're encouraged to do as much off-line as possible. Get in, do work, get out.
The computer, therefore, is a far more abstract, remote, and non-interactive object. I remember what an unpleasant change moving from an 11/785 running VMS to a 3090/VF running VM/CMS. The programming tools were arcane, the OS didn't even have subdirectories (it did have minidisks; i.e. your own virtual pile of floppy disks), and the editors definitely underwhelming. On the other hand, like a VAX/VMS system, the queueing system was an integral part of the OS, so it worked smoothly, as opposed to the unix solutions that are bolted on the side as an after thought. Sun Grid Engine and LSF are pretty close to the old VAX queues, but still not quite as well integrated.
This, of course, is not entirely fair, as large VMS systems were used like mainframes, but still had good tools and a friendly user environment. In the end, think of it generally as a tightly-regulated, non-interactive environment. It's the kind of environment for utter reliability, where it's primarily computers talking to other computers.
Actually, *The Market* did not determine the fate of Nuclear Power, an underinformed public, led by demagogues on both sides, couldn't tell Reactors from Bombs and therefore litigated the industry into submission. Before that, the type of reactors to be used was determined in part by military desires (capable of producing plutonium), not by economic factors (thorium, for instance, which is more plentiful but doesn't produce divertable byproducts).
A class like this could be very valuable, if it trained those people likely to end up making decisions (humanities and business majors) in the actual science behind the technology, or the technologists in how to present to the unschooled what they're actually doing.
Remember, DuPont used to boldly proclaim "Better Things for Better Living Through Chemistry". That's still true, but no modern Ad agency would dare say that for fear of reminding the undereducated that the world is made out of Chemicals. Courses that attempt to prevent that sort of dichotomy from occuring with Nanotech, etc, are frankly a good thing, as long as they're not led by the fear-mongering Rifkins of the world.
Not to go pop culture here, (but this is Slashdot), but I'd rather take my chances with the technology and live in the Blade Runner future, than in the unheated, unhygenic, Arthurian Agrarian past.
Having managed a mixed shop once upon a time (RHEL 3.0, FC 2, SuSE 8.1, 1 Debian just because), I can say that I'd like to see a general sanitization of filesystem layout and startup scripts. The whole multiple desktop issue I care less about, as in the end people sort themselves into their own environments (Gnome, KDE, TWM), and tend to support each other. Remembering where things are stored to keep the system running, distribution to distribution, is a pain.
Frankly, for all I care they could all agree to use OpenSMIT for systems administration, and I'd agree as long as all of the majors used it.
Actually, for years the FP performance of the Intel processors lagged, while the integer plowed ahead. That's partly why the Itanium was such a shock; truly excellent floating point performance (pity the integer sucked rocks through a straw). For most codes, which mixed a fair amount of array handling with the FP code, you still got good performance, even though the FP could have been better. The aggregate mix of instructions helped keep the IA32 boxes competitive. In 2000, our Athlon-800/PIII-1GHz were already ahead of the available Sun or SGI boxes, and rapidly catching up with the low-end (i.e. relatively affordable) AIX systems. With the IA-64, Intel caught up on the high end, and some of that technology has (or will) certainly made it to the EM64-T. Translation: I doubt that there will be any serious performance differences between what would have been the G5 of 2007 and the Intel-??? that they'll actually use.
Currently, we have G5s here, because clock for clock they're as fast as the Opterons, but easier to port the codes we use to. (you basically take the AIX code, XLF for MacOS, and hope the case-insensitivity doesn't bite you). For other programs, Gromacs for instance, the P4 is the clear leader, because of time spent on hand-written assembly loops. Current Intels are not bad processors, though they could certainly turn down the power consumption a bit.
In short, while I will be sorry to lose XLF/XLC (as I was sorry to lose DEC Fortran when the Alphas went away), IFort/ICC are capable of generating good code on current Intel processors, so I doubt we'll lose much on that score. If Apple uses EM64T, then I would expect PathScale to make their compiler available as well, and it is head and shoulders above the other alternatives for AMD-64/EM64T. The real question on the servers is whether two years out we're looking at 64-bit, dual-core, EM64T based systems, or the new dual-core, low-power, Itanium for the XServes and high-end PowerMacs.
But what's the margin on a chip for a game console, versus a chip for a mid-range workstation?
One of the problems is that the new IBM, while in business, is not the visionary old IBM. Apple may have been a small market, but a two-tiered Linux approach (buy dual PPC-970s from the Californians, then graduate to Real Power-5 from us) would have helped displace Intel/AMD from many current markets. Given IBM's service focus, they could have pushed it as well as part of an integrated, end-to-end (yack, business speak) solution.
OTOH, maybe the processors for the XBox, et al., are easier to manufacture, or more tolerant of defects, in which case that would make up for the (probably) lower margins. Getting rid of His Steveness is just the cherry on top. (and I'd love to see how that gets written up in the annual report. "jettisoned high-strung customer...")
We will note that the Power Consumption of the Pentium-D is pretty much identical to the first generation Itanium-2. Those chips had a power/heat envelope that convinced us to stop buying them, except for specialized applications, because of the cooling issues.
Not to ask the obligatory Non-Intel question, but how do those numbers compare against G5 (PPC-970) or current IA-64?
Actually, this seems to be halfway to the Right Thing; you get rid of a non-portable extension to Java (good), but encourage people to use a minor (though clean) language for native-mode programming (iffy).
I've always wondered why that piece of NeXT heritage stuck around, as I'd be willing to wager a doughnut (pick your favorite pusher) that there are more Ada-programmers around than Objective-C gurus. (translation: take a deep breath, wave goodbye, and move to C++ or whatever the modern alternative compiled language would be, or at least make sure the library interfaces are correct for both languages).
Actually, I'd be just as happy, since they have the entire GNU compiler collection, is to simply define a uniform calling convention for their libraries from all languages, and stop worrying abou the One True Language. DEC did it back in the 80s with the Common Language Interface, which allowed me to call the same function (with pretty much the same syntax) from Fortran, C, or Pascal. I presume today that would mean using SWIG to generate a common set of wrappers, and remembering that there are Fortran, Ada, and C++ programmers out there somewhere when generating them.
Maybe my memory is fuzzy, but it seemed simpler to call those functions than on Unix systems where I always have to worry about the library interface on a language by language basis (one underscore or two; case-sensitive or not, etc.) I would like that level of transparency back.
It's not popular because previous generations were Hot, Expensive, and Hard to Program. Adoption was slowed mainly by #2, as it was a significant investment for many research groups to even put one on the floor for testing purposes. The Hard to Program meant that you had wonky versions of Linux, or HP-UX with new compilers, as your OS options, which just increased the resistance.
I ran an Itanium-2 cluster (and had briefly an Itanium-1 loaner) and if the compilers were stable, the first generation easily outpaced the UltraSparc-III systems of the era, and the Itanium-2s kept pace or outpaced the Alphas. However, their power-draw was unbelievable, you had to rebuild the GNU toolchain from RedHat stock, then install whatever that week's version of Intel's C/Fortran compiler was, and on any code that was integer bound, they were barely faster than the late Xeons.
Ah, but for floating point, nothing touched them, except possibly the late Power systems that we couldn't afford. A clean 64-bit architecture, and four floats per clock cycle, meant that for a certain class of problems, there were the cost-effective solution; US-III prices, and twice (at least) the performance. They were even pretty easy to port our codes to, as their pure 64-bit environment (8-byte Integers and pointers), meant that generally modifications to run on Alphas and Crays would run on the Itaniums as well. We would have preferred a next generation of Alpha, but the Itaniums work quite well for what they do.
Discouragingly, now for most (but not all) of the simulations I do, AMD-64 or even EM64T systems are cost-effective, and just as fast. With this new price reduction, I may have to look into adding IA-64s back into the mix, though they're going to have trouble competing against IBM's linux-based Power-5 systems, which offer similar prices, and a more well-known and supported architecture.
Having programmed on them, and managed the PowerChallenge and Origins, I'm not really as sad to see them go. The 4D series with NeWS as its desktop in the late 80s/early 90s was beyond nifty, and the first R8000s we saw (with the 64-bit Fortran compilers that didn't require us to talk to DEC reps!) were also very neat. I'm still entertained by Electropaint when I pass an SGI. That demo (plus flight simulator) probably sold more boxes during their heydey than they'll ever admit.
However, by the time we were in the middle 90s, the programming languages were basically bug of the month, the n32 addressing on the R5000 was creating headaches for porting code, and the price/performance was dreadful. They still built some nice machines (thinking of an O200 w/ Craylink that I still miss), but they were getting harder to justify versus IBM Power systems on the high-end, and PCs on the low end.
The killer for us was when we benchmarked a major simulation package on an O2 with SGI 7.1 compilers versus a PII-400 + G77, and the PC walked away from the SGI. We bought our last machines, warned people that they were transitionary and the porting of code to modern architectures should start now.
I'll lift my SGI mug to them in respect, but will not mourn their passing as I did that of DEC.
Have you looked at Thomas Ray's Tierra simulator? I saw results from that in the early 90s when he was at U. Delaware, and he got quite complex behaviour from a few simple rules, and a single parent organism.
One of the interesting byproducts of Tierra was that one of the first organisms to evolve was a more optimzed version (fewer instructions) of his hand made one, which then went on to give rise to parasites, anti-parasites, predators, etc.
It *is* the right size... Any chance someone can convince His Steveness to add a "phaser" functionality to the next iPod? I'd settle for Stun. On a more serious note, the clean look of the iMacs, Mini, and XServes has been valuable for PR around here. The reaction is much similar to that of non-techies facing an old Refrigerator-sized VAX than a pile of beige PCs.
Thank you. It's been a long time since I looked into VAX emulation. I had considered ordering an OpenVMS kit for our Itania, but just couldn't justify it.
Regrettably, I haven't handled a VMS system since the Crystallographers went to Silicon Graphics, and thence to NT.