Illinois is like Louisiana. Allegedly one of the Longs said, "I want to be buried in Louisiana, so that I can stay active in politics". When I lived there, I don't think they were able to elect a dead guy, but given that that constituency can only go up, it's only a matter of time. We'll also note that dead aldermen will be just as brave as live aldermen in acting independently of the mayor.
I miss Illinois some days; it was corrupt, but it was competent corruption.
Presumably you don't have to buy a new copy of Matlab for each course, though. If they're the full versions, then you're getting a bargain, as only MatLab has a good Open alternative. (Octave)
Quite seriously, I paid around $150/course for books for upper-level science courses 20 years ago, so I presume you're tongue-in-cheekly pointing out the real issue; buy legit copies of software at academic discount prices, and have to drink MGD instead of Bass for a week. A quick look around campus and chat with my students reveals that undergrads these days can afford nice (not beater) cars, iPods, air travel for spring-break, and a host of other luxuries. Those three packages you listed are an iPod, and will probably last as long. Therefore, I would assert that you can afford legit copies of software as well.
Personally, i think we should just factor the prices of such software into tuition, and hand it out for "free", kind of like a Student Activities or Gym fee. I'm sure there's something, such as EndNote which would benefit the other side of campus sufficiently to have it included as well, so that Humanities gets something back other than tools that frighten them.
It's a more refined version of RAH's suggestion in Moon is a Harsh Mistress and other 50's SciFi; things coming in from orbit move quickly. Heavy things moving quickly have lots of kinetic energy, without requiring fallout-producing enhancements. Rods from God is tungsten rods (high heat resistance, heavy), which in theory should be good ground penetrators, effectively LawnDarts of Zeus. I'd be curious to know if once they've punched through a bunker whether they dissapate enough energy to matter, or if they just make a neat little hole and keep going. Just as efficient for most purposes would be to get a bunch of sofa-sized rocks, attach some minimal guidance system to them, and just nudge them in when necessary. OTOH, somehow throwing rocks from orbit just doesn't get the military mind as excited as exotic materials machined to high tolerances with cost-overruns aplenty.
I don't think they're actually using it, or that it's gotten beyond the projector-ware stage. Conspiracy theories abound, but multiple sources say it's not deployed or likely to be deployed any time soon. Sounds great in theory, but for most purposes there is no reason to use space-based lasers/rods/rocks/transluminal battlestations when an A10 Warthog with a couple of laser-guided JDAMs will do the job just as efficiently with a shorter reload time, more accurate targeting, and more flexibility in timing when targeting can be applied. The system is no good if you need to whack someone at 12:43 and the satellite won't be back in the appropriate orbit until 2-5 hours later. In short, someone at the Pentagon has a supply of good alcohol, a plasma HDTV, and a first-rate Sci-Fi movie collection.
Which sounds great until the first buggy 3d driver hangs a desktop when under load, Dell realizes that it has to get its suppliers to write reliable Linux drivers every time they change a chip, and then deal with the screaming from the more radical corners of the community that driver X is (binary/closed/doesn't work on Debian). (I had a 3d screensaver hang a Radeon9000 hard just a couple of months back. The card is old, I know, and never all that open, but it's the kind of thing that gives Dell PR headaches. Headaches beyond the ones from catching fire, of course) In theory, since they already do this for Windows, what's the big deal? In practice, since the suppliers have razor-thin margins and they'll have to add staff, it will raise Dell's price, and slow their supply chain, further eroding their advantage. Then we get to adjusting the distro to average users (adding MP3 and video support, Crossover so they can run real Office and other apps, removing tools that they'll only hurt themselves with) from whatever stock ships from RH. For the ultimate biege-boxer, it looks like a minor nightmare.
Probably they'd end up offering RHEL, and adding the price of RHEL support into the purchase price. Much as I'd like to see a dent in Microsoft's market share, I can't see offering Linux being a winning proposition for Dell on anything smaller than mid-level servers, where the customers either aren't noobs, or have enough money to hire machine jocks who know what they're doing. I'm probably wrong, being as I backed VMS, OS/2, and Itaniums as viable technologies (and still do the third), but one has to suspect that any sounds of Linux emerging from Dell means that a contract negotiation with Microsoft is in progress.
For security reasons, you should have abandoned NIS long ago, and OpenDirectory works just beautifully with OSX, Linux, and Windows clients. Turn on the ssh daemon, and your ssh -X, etc work just fine, and Macs understand NFS, as well as other file systems just fine. In other words, there is no reason to do anything to your RAID array other than tell the Macs where it is and what protocol to use to connect to it. There are also tools, of course, to enable you to make standardized disk images with configurable parameters and use those for future client installs.
I'm running my entire lab off OS-X, with a compute cluster and file system integrated into distributed desktops (OSX and Linux. We had a windows but I sensibly turned it off when we bought the first IntelMac), and not so much as a hiccup. The main problems you're describing are the classic, "it looks unixy, so I'm going to treat it as if it were a Linux box." No, it's a Mac, descended from NeXTs. Get the Apple docs out (dreadful though they may be), read a little of "The Mac Way", and quit fighting it. I found most of my problems at first arose from trying to treat Macs as if they were just nice-looking RedHat boxes, rather than something different.
Pardon for sounding rude, but it sounds like you've learned one system, and aren't willing to attempt to learn another. Current Macs are one of the easiest machine to integrate into a mixed environment that I've encountered, and this is after over a decade and a half of running various Unices, Linuces, Windows, and VMS systems in mixed environments.
Maybe, and even quite possibly. I still note (and this is said by others), that you can break corporate life-cycles into a variety of stages, where the risk-taking that gets you out of the garage and onto the NYSE is the wrong technique to get you from successful to huge. The problem really comes in when the caution which makes innovative projects run quietly internally until they're ready to be released turns into paralysis which prevents them from getting started at all.
Using IBM as an example, maintaining backwards compatibility with the big iron as machines got smaller during the 70s was a good idea, but intentionally crippling disruptive technologies such as the PC/RT in order to protect doomed product lines that could have been emulated on the newer systems was not.
Again, that's *now*. 1991-1993, BSD is chained to the ground because of fear of AT&T's lawyers, so people wanting a Unix for personal use go Linux instead. By the time that's sorted out, people have too much time/mindshare invested in Linux, have started to become experts, and have begun evangalising it. It's not a bad solution, but many of the reasons that are currently suggested for why Linux took off and BSD didn't as much rely upon legal issues that are now almost 20 years gone. Technically, BSD was much more advanced, but fear of AT&T held it back just long enough.
Kind of like a new yard; if you keep spraying Roundup to keep the broadleaf weeds down, you'll give grass time to start. Later, when you stop spraying, the grass is still dominant because of the interlocking roots that got established while you kept the dandelions at bay. Think of lawyers as intellectual Roundup.
I've been programming, personal and academic out to semi-pro (it was and is a component of my job, but not my primary purpose) for about 30 years, and all I can say is, "did you ever use a Gateway PC versus an IBM back in the 80s/90s?" They were inferior, unless by "not inferior" you mean "could be completely gutted and rebuilt from parts you ordered from the back of a magazine". I've retired more IBM/HP/Sun machines due to relatively decreasing performance than actual failure, as opposed to various companies rhyming with "Hell" who seem to add a "halt and catch fire" instruction to their chipsets.
You'll notice that businesses didn't really get enthusiastic about Open-source until companies like RedHat grew to support it via standard commercial service contracts, and behemoths like IBM and HP bought into it and made improvements. In short, until the revolutionaries have been coopted and become part of the establishment, they remain primarily guerillas in the hills. Those managers aren't entirely stupid; when someone is going to blame you for a billion dollar mistake, you tend to act a little cautiously. Big corporations tend to be a touch conservative, which is how they last long enough to get big.
You're forgetting the BSD copyright issues. About the time Linux was becoming public, the various BSD releases had been held back to AT&T owned code, which meant you had to buy an expensive license to get your own copy. If you were a Uni, it wasn't bad, but for an individual, somewhat daunting. When they finally factored the last of the proprietary code out (during the 386BSD days), people were still worried that it wasn't really gone, and therefore waited for someone else to be the first to be sued. I knew Comp-Sci grad students at that time who thought it was great, if it was really clean, but with Muppet-Labs just up the Jersey Turnpike from us, didn't want to contribute to the project, and be the test case for AT*T's lawyers. It wasn't the BSD license, but the licensing of BSD by AT&T that held it back.
So, some of Linux's success was timing, and BSD being held back by fear of lawyers. Had Linus waiting another year or two to make his public release, you might all be running *BSD on your home machines, and arguing why it hadn't taken over the desktop yet. You'll also notice that the early user environment wrapped around the Linux kernel was heavily BSD flavored, which made an easy transition from the other cheap Unix of the day, SunOS.
Because gnuplot makes ugly output and ploticus is command-line driven? It's like stepping back into the early 80s, or earlier if you had access to 123 in the 80's. I use GNUPlot for band-structure plots, or quick and dirty visualizations, but I've had physical science journals complain about how gnuplot graphs looked, and ask whether they could be replotted using something better. That would be other chemists, btw, not the secretarial staff, that was raising the objection.
Would that there were something better than Excel that was reasonably priced, but there's a reason that people who don't use Excel tend to have something like Deltagraph lying around on their HDs as well.
It'll still work, you'll just have headphones, instead of big electrical jacks. Just hope the robots don't find out that Barry Manilow produces the most energy.
The entertainment aspect is an issue. I'm trying to pick a textbook for next fall for accelerated freshman chem (college), and while I have a couple I like, with reasonable prices, I've been informed that those are too difficult, with too few illustrations, multi-media extras, etc. They've tried one before, and while the faculty like it and use it as a personal supplement (brush up on unfamiliar subjects or get tougher problems from it), the students whine too much, and we just can't have that.
What gets me is that one of the books I've been warned off from is Linus Pauling's "College Chemistry", which is eminently readable, has everything that one expects from freshman chemistry, and was written by a true master of the material. However, apparently freshmen will be put off by its monochrome only illustrations, and paucity of those.
If you're teaching high-school, put a little iron in those kids before you let them out, will you?
If it's an enterprise client, then the cost of the hardware is moot, and you'll run the same commodity hardware for Win2K3 or Linux/Samba anyway. The data is more expensive than the machine in those cases, and you're not (at least if you're sane and like your job), going to cut corners by buying cheap hardware unless you're like Google, buy it by the metric ton, and have the skills to make it completely redundant.
This applies, btw, if you're a one person business, or a fortune 500. Only the size of the hardware changes. (most of us, even in our geekier moments, do not buy E15Ks or p690s for personal/small-business use) Licensing costs are a real issue, but if you're going to make your CIO happy by going with a commercially supported distro (RHEL/SLES), the cost differences aren't going to be as great as you might think.
I ran it 10 years ago. 10 years ago was the cusp of when it was finally breaking out in the technical community, and getting established as the future. You had many choices for desktops: you could run TWM, and pretend you had a 1991 VAXstation, FVWM, and try to convince yourself that you only *really* needed to get a new prescription for your glasses, or early KDE and pretend you were running CDE on something like AIX. LessTIF was trying hard (and generally doing pretty well) to emulate closed Motif, which it seemed everything had been built for, and applications like OpenOffice were still StarOffice, and shipping for OS/2. You could get the kneecapped WordPerfect for Linux, though.
It was getting usable for replacing low-end Unix boxes back then. We replaced several SGI O2 and similar beasties when we discovered that a 16-meg Matrox G400, GNU Fortran (g77), and an early version of Accelerated OpenGL (DRI?) was a fifth the price and twice the performance of the R5K SGI O2. As a desktop OS, though, there were as many Be or OS/2 afficionados running loose as Linux. We also tended to have to purchase a commercial X solution because of some proprietary whacked-out laptop chipset that the XFree team couldn't get docs to or reverse engineer. I remember discussions at work where people would express how they'd rather run Solaris-x86 on their machines than Linux, because Solaris was clean, and Linux had files all over the place, in a mix of SysV, BSD, and downright Odd formats.
I would also add that SMP multiprocessing was new then as well, so while it was showing what was possible, it was hardly a good choice for day to day users unless they were already Unix jocks. It was considered cool to run, but scavenging an old SparcStation was considered an equally viable option.
Don't kid yourself; corporate adoption of Linux for End-users 10 years ago would have been the end of the experiment. Your boss now wouldn't even listen to you when you mentioned Linux at work, and would be probably testing you for drugs at this very moment. It needed time to grow quietly in the dark. After all, 65 million years ago some dinosaur could have paid attention to the early mammals and said, "hmm, those make a nice after-dinner snack", and then where would we be now?
Welcome to the joys of Linux. This is one of the rare times when posting a "So what's the difference between RHEL and Fedora other than $200/yr" to here might have been worthwhile. After you sorted out the responses telling you to compile everything from scratch continually (Gentoo), or use completely free GNU/Linux (Debian), or use completely free GNU/Linux with a tasteful graphical front-end and some thoughts about the end-user (Ubuntu), you would have found that Fedora is a testing version where ideas are tried. RHEL is a much more slowly evolving beast that takes only the pieces from Fedora that are reliable and proven, and then supports them for years. In a way, Fedora is like the old consumer RedHat Linux, except on speed as it tries to keep ahead of the curve with new drivers, interface ideas, cutting-edge software, etc. It's a good choice for personal use, but it doesn't track RHEL closely enough to be used for validation. (I tried for a while to use it as a development system for a RHEL-based cluster, and finally gave up over library dependencies. Fedora rapidly got too far ahead of the cluster libraries, and went unsupported too quickly)
Don't feel bad; it's hard to find a free version of RHEL, though Centos is the current leader. After a while, you understand why vendors support it or SuSE's equivalent. This is one of those cases where sighing and signing the check really would have been the right answer.
Only if it ships with a phaser set to "dematerialize" in the box. Oh how I longed for one of those when I was a sysadmin, and had to deal with end users. Of course, I would have settled for the much lower-tech, biomaterial-based, Louiseville Slugger, most days.
On the other hand, if you're inside a fencing helmet (steel mesh), and someone runs the tip of their blade (steel pointy object) across your mask at high speed, the sparks are extremely visible, at least to you. Onlookers tend to be less impressed.
So, I can see space colonies for (1) national pride and cultural jockeying for position, or (2) a frontier stirs the imaginations back home, or even (3) a place for people who reject the Lockean "social contract" of our society to go. However, why is colonization to ensure our species's survival an issue? What makes our DNA so uniquely valuable that if removed from the civilization we've built (here and there) it deserves to be propagated. If the big one hits, then unless someone thought to digitize and send along our cultural heritage, then we've lost most of what made us noteworthy anyway. If you do, then it's removed from living contexts, and just ashes of a lost world. Memento Mori.
Press a bunch of gold disks with a mixture of Bach and Shakespeare, send them into the void in all directions, and announce, "We are, We were" far more effectively than colonizing Mars ever will.
This post probably influenced by a caffeine headache, but putting money into trying to make a self-sustaining colony on mars just in case an asteroid hits home, versus putting that money into finding the blasted things and learning how to steer them away (and maybe mining them in the process; imagine Pt/Ir cheap enough to use for plumbing), seems like misplaced hubris. Again, if you say we're doing the colony-thing for reasons 1-3, I say go for it, as it's an extension of our civilization. Option three is simply a change to crawl back up from the mud again, with memory of where we fell from. (and just imagine the legends from that process that will be passed down, misinterpreted, and misapplied.)
That's what I wonder about; is there some psychological link between Deistic (somewhat personified deities) religion and intolerant evangalism? The link has been made for monotheism (if my god is real, yours can't be, and since I'm never wrong...), but I don't know beyond that. It is known that the original atheistic buddhisms have mostly (with the exception of Vietnamese or Japanese Zen) given way to ones with at least spirits and demons, if not all-powerful gods outright, so there is some pull towards believing in a personified diety. Probably the same pull that makes us believe in the personal touch of charismatic leaders, but it's a curious question.
Not really my field either. I have a minor from undergrad in it (which allowed me to take great courses like "Atonement"), but we never really examined why a particular belief-system took over.
As for Essjay, we could do that. I've been waiting for him to post to this thread, but no signs yet.
OTOH (just to be a pain), look at the number of believers of Theistic religions (Christianity, Islam, Hinduism, Judaism), the ones that weren't but increasingly are practiced as such (Thai or Tibetian Buddhism), and the explosion in the above in FSU and China as soon as controls were relaxed, then there is seems to be a trend towards both mystical experiences and authoritarian systems with a Deity on top. (at least one, and often many)
This is also seen throughout history, as Deistic and authoritarian (Catholicism) drives out less deistic and non-authoritarian (Gnostics). Substitute Southern Baptist versus Quaker if discussing the US.
Or, we could try a free-market approach: We charge ranchers market-rates for grazing lands and water, we stop propping up corn-states with subsidies so that dirt-cheap corn isn't available for feedlot use as clearance prices, and we treat feedlots as the industrial polluters they are, and regulate them accordingly. While we're at it, label beef openly. You should be able to pick up a package of beef, look at the label, and see "feedlot-raised herford routinely injected with the following antibiotics@concentration@interval just in case". Next to it would be, "grass-fed, open-range, dusted for ticks and certified treated only for illness and not within 120 days of slaughter". End result: a whole whack of inefficient cattlemen go under (at least we won't have to list to the whining of rugged individualists who only need continual tax subsidies but no other gub'mint involvement to stay in business), meat prices rise a bit, successful ranchers will go back to open-range grazing like Argentina does, and possibly some of that fertilizer-gulping corn will be plowed under and prarie grass for grazing planted in its place.
With better animal-husbandry (don't feed a grazing animal expecting a high-fiber diet acidic corn and make it stand in one place all day), you'll get healthier animals, and less need for antibiotics and other promoters to make them grow. And before anyone starts the accusations, I eat meat and look askance at soybean-based alleged food products. These are the same people pushing irradiation of food so that they don't have to slow down slaughterhouses and worry about what bits of cattle-waste end up on or in meat. Sometimes the answer really is, "it's not a machine, and we should not be producing beef as if it was Nikes, so worry about public health first, then about m
Or, in a one-picture nutshell: PC vs. Mac vs. Linux (of course, to be truer to the subject matter, there should be a dozen of the third guy, all slightly differently colored, wrestling, while the PC and Mac guys just look on in horror.)
They should probably hand out a copy of Brooks, "The Mythical Man-Month" to the entire Vista team (managers down to programmers), and tell them to get back to work and try again. It has that air of OS/360 about it. Maybe they can hand out decks of bronzed punch-cards to the team leaders to congratulate them on their runaway success.
My students are playing with it, and their comment is that it's ok, but on less than brand-new hardware, it doesn't run Aero, so it's basically WindowsXP+, with a few annoying hardware incompatibilities. Thankfully we run on PPC Macs, so it won't be making much of an inroad where I have to deal with it for a while yet.
Slashdot Mirror
Illinois is like Louisiana. Allegedly one of the Longs said, "I want to be buried in Louisiana, so that I can stay active in politics". When I lived there, I don't think they were able to elect a dead guy, but given that that constituency can only go up, it's only a matter of time. We'll also note that dead aldermen will be just as brave as live aldermen in acting independently of the mayor.
I miss Illinois some days; it was corrupt, but it was competent corruption.
Presumably you don't have to buy a new copy of Matlab for each course, though. If they're the full versions, then you're getting a bargain, as only MatLab has a good Open alternative. (Octave)
Quite seriously, I paid around $150/course for books for upper-level science courses 20 years ago, so I presume you're tongue-in-cheekly pointing out the real issue; buy legit copies of software at academic discount prices, and have to drink MGD instead of Bass for a week. A quick look around campus and chat with my students reveals that undergrads these days can afford nice (not beater) cars, iPods, air travel for spring-break, and a host of other luxuries. Those three packages you listed are an iPod, and will probably last as long. Therefore, I would assert that you can afford legit copies of software as well.
Personally, i think we should just factor the prices of such software into tuition, and hand it out for "free", kind of like a Student Activities or Gym fee. I'm sure there's something, such as EndNote which would benefit the other side of campus sufficiently to have it included as well, so that Humanities gets something back other than tools that frighten them.
It's a more refined version of RAH's suggestion in Moon is a Harsh Mistress and other 50's SciFi; things coming in from orbit move quickly. Heavy things moving quickly have lots of kinetic energy, without requiring fallout-producing enhancements. Rods from God is tungsten rods (high heat resistance, heavy), which in theory should be good ground penetrators, effectively LawnDarts of Zeus. I'd be curious to know if once they've punched through a bunker whether they dissapate enough energy to matter, or if they just make a neat little hole and keep going. Just as efficient for most purposes would be to get a bunch of sofa-sized rocks, attach some minimal guidance system to them, and just nudge them in when necessary. OTOH, somehow throwing rocks from orbit just doesn't get the military mind as excited as exotic materials machined to high tolerances with cost-overruns aplenty.
I don't think they're actually using it, or that it's gotten beyond the projector-ware stage. Conspiracy theories abound, but multiple sources say it's not deployed or likely to be deployed any time soon. Sounds great in theory, but for most purposes there is no reason to use space-based lasers/rods/rocks/transluminal battlestations when an A10 Warthog with a couple of laser-guided JDAMs will do the job just as efficiently with a shorter reload time, more accurate targeting, and more flexibility in timing when targeting can be applied. The system is no good if you need to whack someone at 12:43 and the satellite won't be back in the appropriate orbit until 2-5 hours later. In short, someone at the Pentagon has a supply of good alcohol, a plasma HDTV, and a first-rate Sci-Fi movie collection.
Which sounds great until the first buggy 3d driver hangs a desktop when under load, Dell realizes that it has to get its suppliers to write reliable Linux drivers every time they change a chip, and then deal with the screaming from the more radical corners of the community that driver X is (binary/closed/doesn't work on Debian). (I had a 3d screensaver hang a Radeon9000 hard just a couple of months back. The card is old, I know, and never all that open, but it's the kind of thing that gives Dell PR headaches. Headaches beyond the ones from catching fire, of course) In theory, since they already do this for Windows, what's the big deal? In practice, since the suppliers have razor-thin margins and they'll have to add staff, it will raise Dell's price, and slow their supply chain, further eroding their advantage. Then we get to adjusting the distro to average users (adding MP3 and video support, Crossover so they can run real Office and other apps, removing tools that they'll only hurt themselves with) from whatever stock ships from RH. For the ultimate biege-boxer, it looks like a minor nightmare.
Probably they'd end up offering RHEL, and adding the price of RHEL support into the purchase price. Much as I'd like to see a dent in Microsoft's market share, I can't see offering Linux being a winning proposition for Dell on anything smaller than mid-level servers, where the customers either aren't noobs, or have enough money to hire machine jocks who know what they're doing. I'm probably wrong, being as I backed VMS, OS/2, and Itaniums as viable technologies (and still do the third), but one has to suspect that any sounds of Linux emerging from Dell means that a contract negotiation with Microsoft is in progress.
For security reasons, you should have abandoned NIS long ago, and OpenDirectory works just beautifully with OSX, Linux, and Windows clients. Turn on the ssh daemon, and your ssh -X, etc work just fine, and Macs understand NFS, as well as other file systems just fine. In other words, there is no reason to do anything to your RAID array other than tell the Macs where it is and what protocol to use to connect to it. There are also tools, of course, to enable you to make standardized disk images with configurable parameters and use those for future client installs.
I'm running my entire lab off OS-X, with a compute cluster and file system integrated into distributed desktops (OSX and Linux. We had a windows but I sensibly turned it off when we bought the first IntelMac), and not so much as a hiccup. The main problems you're describing are the classic, "it looks unixy, so I'm going to treat it as if it were a Linux box." No, it's a Mac, descended from NeXTs. Get the Apple docs out (dreadful though they may be), read a little of "The Mac Way", and quit fighting it. I found most of my problems at first arose from trying to treat Macs as if they were just nice-looking RedHat boxes, rather than something different.
Pardon for sounding rude, but it sounds like you've learned one system, and aren't willing to attempt to learn another. Current Macs are one of the easiest machine to integrate into a mixed environment that I've encountered, and this is after over a decade and a half of running various Unices, Linuces, Windows, and VMS systems in mixed environments.
Maybe, and even quite possibly. I still note (and this is said by others), that you can break corporate life-cycles into a variety of stages, where the risk-taking that gets you out of the garage and onto the NYSE is the wrong technique to get you from successful to huge. The problem really comes in when the caution which makes innovative projects run quietly internally until they're ready to be released turns into paralysis which prevents them from getting started at all.
Using IBM as an example, maintaining backwards compatibility with the big iron as machines got smaller during the 70s was a good idea, but intentionally crippling disruptive technologies such as the PC/RT in order to protect doomed product lines that could have been emulated on the newer systems was not.
Again, that's *now*. 1991-1993, BSD is chained to the ground because of fear of AT&T's lawyers, so people wanting a Unix for personal use go Linux instead. By the time that's sorted out, people have too much time/mindshare invested in Linux, have started to become experts, and have begun evangalising it. It's not a bad solution, but many of the reasons that are currently suggested for why Linux took off and BSD didn't as much rely upon legal issues that are now almost 20 years gone. Technically, BSD was much more advanced, but fear of AT&T held it back just long enough.
Kind of like a new yard; if you keep spraying Roundup to keep the broadleaf weeds down, you'll give grass time to start. Later, when you stop spraying, the grass is still dominant because of the interlocking roots that got established while you kept the dandelions at bay. Think of lawyers as intellectual Roundup.
I've been programming, personal and academic out to semi-pro (it was and is a component of my job, but not my primary purpose) for about 30 years, and all I can say is, "did you ever use a Gateway PC versus an IBM back in the 80s/90s?" They were inferior, unless by "not inferior" you mean "could be completely gutted and rebuilt from parts you ordered from the back of a magazine". I've retired more IBM/HP/Sun machines due to relatively decreasing performance than actual failure, as opposed to various companies rhyming with "Hell" who seem to add a "halt and catch fire" instruction to their chipsets.
You'll notice that businesses didn't really get enthusiastic about Open-source until companies like RedHat grew to support it via standard commercial service contracts, and behemoths like IBM and HP bought into it and made improvements. In short, until the revolutionaries have been coopted and become part of the establishment, they remain primarily guerillas in the hills. Those managers aren't entirely stupid; when someone is going to blame you for a billion dollar mistake, you tend to act a little cautiously. Big corporations tend to be a touch conservative, which is how they last long enough to get big.
You're forgetting the BSD copyright issues. About the time Linux was becoming public, the various BSD releases had been held back to AT&T owned code, which meant you had to buy an expensive license to get your own copy. If you were a Uni, it wasn't bad, but for an individual, somewhat daunting. When they finally factored the last of the proprietary code out (during the 386BSD days), people were still worried that it wasn't really gone, and therefore waited for someone else to be the first to be sued. I knew Comp-Sci grad students at that time who thought it was great, if it was really clean, but with Muppet-Labs just up the Jersey Turnpike from us, didn't want to contribute to the project, and be the test case for AT*T's lawyers. It wasn't the BSD license, but the licensing of BSD by AT&T that held it back.
So, some of Linux's success was timing, and BSD being held back by fear of lawyers. Had Linus waiting another year or two to make his public release, you might all be running *BSD on your home machines, and arguing why it hadn't taken over the desktop yet. You'll also notice that the early user environment wrapped around the Linux kernel was heavily BSD flavored, which made an easy transition from the other cheap Unix of the day, SunOS.
Because gnuplot makes ugly output and ploticus is command-line driven? It's like stepping back into the early 80s, or earlier if you had access to 123 in the 80's. I use GNUPlot for band-structure plots, or quick and dirty visualizations, but I've had physical science journals complain about how gnuplot graphs looked, and ask whether they could be replotted using something better. That would be other chemists, btw, not the secretarial staff, that was raising the objection.
Would that there were something better than Excel that was reasonably priced, but there's a reason that people who don't use Excel tend to have something like Deltagraph lying around on their HDs as well.
It'll still work, you'll just have headphones, instead of big electrical jacks. Just hope the robots don't find out that Barry Manilow produces the most energy.
The entertainment aspect is an issue. I'm trying to pick a textbook for next fall for accelerated freshman chem (college), and while I have a couple I like, with reasonable prices, I've been informed that those are too difficult, with too few illustrations, multi-media extras, etc. They've tried one before, and while the faculty like it and use it as a personal supplement (brush up on unfamiliar subjects or get tougher problems from it), the students whine too much, and we just can't have that.
What gets me is that one of the books I've been warned off from is Linus Pauling's "College Chemistry", which is eminently readable, has everything that one expects from freshman chemistry, and was written by a true master of the material. However, apparently freshmen will be put off by its monochrome only illustrations, and paucity of those.
If you're teaching high-school, put a little iron in those kids before you let them out, will you?
If it's an enterprise client, then the cost of the hardware is moot, and you'll run the same commodity hardware for Win2K3 or Linux/Samba anyway. The data is more expensive than the machine in those cases, and you're not (at least if you're sane and like your job), going to cut corners by buying cheap hardware unless you're like Google, buy it by the metric ton, and have the skills to make it completely redundant.
This applies, btw, if you're a one person business, or a fortune 500. Only the size of the hardware changes. (most of us, even in our geekier moments, do not buy E15Ks or p690s for personal/small-business use) Licensing costs are a real issue, but if you're going to make your CIO happy by going with a commercially supported distro (RHEL/SLES), the cost differences aren't going to be as great as you might think.
I ran it 10 years ago. 10 years ago was the cusp of when it was finally breaking out in the technical community, and getting established as the future. You had many choices for desktops: you could run TWM, and pretend you had a 1991 VAXstation, FVWM, and try to convince yourself that you only *really* needed to get a new prescription for your glasses, or early KDE and pretend you were running CDE on something like AIX. LessTIF was trying hard (and generally doing pretty well) to emulate closed Motif, which it seemed everything had been built for, and applications like OpenOffice were still StarOffice, and shipping for OS/2. You could get the kneecapped WordPerfect for Linux, though.
It was getting usable for replacing low-end Unix boxes back then. We replaced several SGI O2 and similar beasties when we discovered that a 16-meg Matrox G400, GNU Fortran (g77), and an early version of Accelerated OpenGL (DRI?) was a fifth the price and twice the performance of the R5K SGI O2. As a desktop OS, though, there were as many Be or OS/2 afficionados running loose as Linux. We also tended to have to purchase a commercial X solution because of some proprietary whacked-out laptop chipset that the XFree team couldn't get docs to or reverse engineer. I remember discussions at work where people would express how they'd rather run Solaris-x86 on their machines than Linux, because Solaris was clean, and Linux had files all over the place, in a mix of SysV, BSD, and downright Odd formats.
I would also add that SMP multiprocessing was new then as well, so while it was showing what was possible, it was hardly a good choice for day to day users unless they were already Unix jocks. It was considered cool to run, but scavenging an old SparcStation was considered an equally viable option.
Don't kid yourself; corporate adoption of Linux for End-users 10 years ago would have been the end of the experiment. Your boss now wouldn't even listen to you when you mentioned Linux at work, and would be probably testing you for drugs at this very moment. It needed time to grow quietly in the dark. After all, 65 million years ago some dinosaur could have paid attention to the early mammals and said, "hmm, those make a nice after-dinner snack", and then where would we be now?
Welcome to the joys of Linux. This is one of the rare times when posting a "So what's the difference between RHEL and Fedora other than $200/yr" to here might have been worthwhile. After you sorted out the responses telling you to compile everything from scratch continually (Gentoo), or use completely free GNU/Linux (Debian), or use completely free GNU/Linux with a tasteful graphical front-end and some thoughts about the end-user (Ubuntu), you would have found that Fedora is a testing version where ideas are tried. RHEL is a much more slowly evolving beast that takes only the pieces from Fedora that are reliable and proven, and then supports them for years. In a way, Fedora is like the old consumer RedHat Linux, except on speed as it tries to keep ahead of the curve with new drivers, interface ideas, cutting-edge software, etc. It's a good choice for personal use, but it doesn't track RHEL closely enough to be used for validation. (I tried for a while to use it as a development system for a RHEL-based cluster, and finally gave up over library dependencies. Fedora rapidly got too far ahead of the cluster libraries, and went unsupported too quickly)
Don't feel bad; it's hard to find a free version of RHEL, though Centos is the current leader. After a while, you understand why vendors support it or SuSE's equivalent. This is one of those cases where sighing and signing the check really would have been the right answer.
Only if it ships with a phaser set to "dematerialize" in the box. Oh how I longed for one of those when I was a sysadmin, and had to deal with end users. Of course, I would have settled for the much lower-tech, biomaterial-based, Louiseville Slugger, most days.
On the other hand, if you're inside a fencing helmet (steel mesh), and someone runs the tip of their blade (steel pointy object) across your mask at high speed, the sparks are extremely visible, at least to you. Onlookers tend to be less impressed.
So, I can see space colonies for (1) national pride and cultural jockeying for position, or (2) a frontier stirs the imaginations back home, or even (3) a place for people who reject the Lockean "social contract" of our society to go. However, why is colonization to ensure our species's survival an issue? What makes our DNA so uniquely valuable that if removed from the civilization we've built (here and there) it deserves to be propagated. If the big one hits, then unless someone thought to digitize and send along our cultural heritage, then we've lost most of what made us noteworthy anyway. If you do, then it's removed from living contexts, and just ashes of a lost world. Memento Mori.
Press a bunch of gold disks with a mixture of Bach and Shakespeare, send them into the void in all directions, and announce, "We are, We were" far more effectively than colonizing Mars ever will.
This post probably influenced by a caffeine headache, but putting money into trying to make a self-sustaining colony on mars just in case an asteroid hits home, versus putting that money into finding the blasted things and learning how to steer them away (and maybe mining them in the process; imagine Pt/Ir cheap enough to use for plumbing), seems like misplaced hubris. Again, if you say we're doing the colony-thing for reasons 1-3, I say go for it, as it's an extension of our civilization. Option three is simply a change to crawl back up from the mud again, with memory of where we fell from. (and just imagine the legends from that process that will be passed down, misinterpreted, and misapplied.)
Most people have put "30 days of Vista" off until after they finish "Two Years Before the Mast" and 120 Days of Sodom.
That's what I wonder about; is there some psychological link between Deistic (somewhat personified deities) religion and intolerant evangalism? The link has been made for monotheism (if my god is real, yours can't be, and since I'm never wrong...), but I don't know beyond that. It is known that the original atheistic buddhisms have mostly (with the exception of Vietnamese or Japanese Zen) given way to ones with at least spirits and demons, if not all-powerful gods outright, so there is some pull towards believing in a personified diety. Probably the same pull that makes us believe in the personal touch of charismatic leaders, but it's a curious question.
Not really my field either. I have a minor from undergrad in it (which allowed me to take great courses like "Atonement"), but we never really examined why a particular belief-system took over.
As for Essjay, we could do that. I've been waiting for him to post to this thread, but no signs yet.
OTOH (just to be a pain), look at the number of believers of Theistic religions (Christianity, Islam, Hinduism, Judaism), the ones that weren't but increasingly are practiced as such (Thai or Tibetian Buddhism), and the explosion in the above in FSU and China as soon as controls were relaxed, then there is seems to be a trend towards both mystical experiences and authoritarian systems with a Deity on top. (at least one, and often many)
This is also seen throughout history, as Deistic and authoritarian (Catholicism) drives out less deistic and non-authoritarian (Gnostics). Substitute Southern Baptist versus Quaker if discussing the US.
Or, we could try a free-market approach: We charge ranchers market-rates for grazing lands and water, we stop propping up corn-states with subsidies so that dirt-cheap corn isn't available for feedlot use as clearance prices, and we treat feedlots as the industrial polluters they are, and regulate them accordingly. While we're at it, label beef openly. You should be able to pick up a package of beef, look at the label, and see "feedlot-raised herford routinely injected with the following antibiotics@concentration@interval just in case". Next to it would be, "grass-fed, open-range, dusted for ticks and certified treated only for illness and not within 120 days of slaughter". End result: a whole whack of inefficient cattlemen go under (at least we won't have to list to the whining of rugged individualists who only need continual tax subsidies but no other gub'mint involvement to stay in business), meat prices rise a bit, successful ranchers will go back to open-range grazing like Argentina does, and possibly some of that fertilizer-gulping corn will be plowed under and prarie grass for grazing planted in its place.
With better animal-husbandry (don't feed a grazing animal expecting a high-fiber diet acidic corn and make it stand in one place all day), you'll get healthier animals, and less need for antibiotics and other promoters to make them grow. And before anyone starts the accusations, I eat meat and look askance at soybean-based alleged food products. These are the same people pushing irradiation of food so that they don't have to slow down slaughterhouses and worry about what bits of cattle-waste end up on or in meat. Sometimes the answer really is, "it's not a machine, and we should not be producing beef as if it was Nikes, so worry about public health first, then about m
Or, in a one-picture nutshell: PC vs. Mac vs. Linux (of course, to be truer to the subject matter, there should be a dozen of the third guy, all slightly differently colored, wrestling, while the PC and Mac guys just look on in horror.)
They should probably hand out a copy of Brooks, "The Mythical Man-Month" to the entire Vista team (managers down to programmers), and tell them to get back to work and try again. It has that air of OS/360 about it. Maybe they can hand out decks of bronzed punch-cards to the team leaders to congratulate them on their runaway success.
My students are playing with it, and their comment is that it's ok, but on less than brand-new hardware, it doesn't run Aero, so it's basically WindowsXP+, with a few annoying hardware incompatibilities. Thankfully we run on PPC Macs, so it won't be making much of an inroad where I have to deal with it for a while yet.
Because you can watch your data flowing down the Intertubes, which makes it transportation related.