To us, CompSci is useful only insofar as it makes programs run faster, smaller, more reliably, etc. Sure, and engineering is only useful insofar as it makes your car go faster. Leave the physics and the math and the actual foundations of the discipline to the "purists" with their "incredible focus", but make sure Lincoln Tech graduates enough mechanics who can figure out how to tap an intake manifold so you can hook up a nitrous bottle...
The big challenge facing Computer Science right now is to figure out how best to handle parallel processing and concurrency. The megahertz wars are over, you're not going to see CPUs get much faster, they're just going to cram more of them into one chip. Problem is, how do you break your code down so it runs on multiple chips? And what can you do to current code that'll allow it to take advantage of multiple cores/CPUs without rewriting it? Some kind of static analysis, maybe re-target it to a virtual machine that can parallelize it during execution, maybe some other abstraction layer that makes an 8-core machine look like a single core that runs 8 times faster? No amount of programming is going to solve this, you're going to have to go back to square one and re-think how software should be designed, what the languages will look like, what facilities an OS should provide — not the kind of stuff some "go faster" programmer is going to sort out over the weekend.
If you really want to see how CS improves programming, check out Jon Bentley's Programming Pearls books. And if you want to see how engineering can make your car go faster, look at Fred Puhn's How to Make Your Car Handle or Carrol Smith's Tune to Win (or any of his other racing engineering books).
"Programming is to Computer Science as telescopes are to Astronomy" (paraphrased from someone who said it better than I). If you actually take any computer science courses, you find they're more about things like measuring algorithm efficiency and finding ways around NP-hard problems (and how to tell if a problem is merely NP-hard or if it's really NP-complete). Sure, there's programming involved, but once you get past the sophomore-level courses you do stuff like design virtual memory systems or network protocols, and then maybe write a simulator to verify your theories or prove that you can meet your performance goals.
Maybe things have changed since the last CS class I took, but I definitely recall spending more time on the final drawing boxes and arrows than scribbling down code...
record player with USB? doesn't that defeat the purpose of analog sound quality? Transferring any analog source to a computer defeats the purpose of analog sound quality. Even if you get a "audiophile-quality" sound card and plug your turntable directly into that, it gets digitized at some point. You don't do it because you love the way it sounds, you do it so you can have access to it via your computer. You can use lossless compression and a high sampling rate to capture as much fidelity as possible, but don't forget that the inner tracks on a vinyl album are lower fidelity than the outer (due to CAV vs. CLV like a CD), so it becomes less of an issue for some tracks.
I've seen cookies on machines that are years old. Up to three years in some cases. Yeah, our "corporate policy" where I work is to set cookie expiry to (time_t)-1. Even if there's no data in the cookie, it'll never die. On the sites I work on, I always just use session cookies (sorry, Corporate, must have forgotten to set that field...:P), because that's all I need.
I wonder if it'd be worthwhile to have your browser set a default maximum expiration date? Any cookie set would have its expiration date set to MIN(requested-date,browser-max-exp-date). Set it for a day, a week, maybe have a keyword "session" to force all cookies to be session cookies. Hey, if a site wants to stick some data on my machine, I think I should have some say as to how long it sticks around! Hmmm, might be time to take a look at the Firefox cookie code, that shouldn't be too hard to set up...
Yeah, I still miss my Commodore SuperPET. I saw one when they first came out, and had to have one, as it was the only machine I'd ever seen that could run APL using the actual APL character set. It also had 1MB floppies (back when the PC had just switched from 160 to 180K). It was also the first dual-processor machine I ever saw (had both a 6800 and a 6809, I believe).
Twenty or so years later, I picked one up for $50. Hauled it home, fumbled with the not-quite-GPIB cables, finally got everything hooked up, and booted it up. Played with APL for a couple of nights, sketched out some code for a shared-memory multiprocessing scheme (6809 could pass tasks off to the 6800, communicating via a bank-switched block of memory). But then I got tired of all the limitations. While it had impressive performance for a machine of its day (its disk I/O was at least twice as fast as my 200MHz PPro's floppy), its day was long past, and when you got right down to it, I just wasn't that nostalgic for a machine with a firmware monitor. I was when I was debugging VB code, but not when I actually had a spare minute to go play. So I finally sold it when it was time to move (got my $50 back!), but I still wish I had kept it. Now, if I could find an original TRS-80 (my first computer), I'd definitely pick that up...
most bindings in vi are arbitrary Wow, I may have to re-calibrate my bogometer after that one. Arbitrary? ^F goes (F)orward a page, ^B goes (B)ackward a page, ^D goes (D)own half a page, ^U goes (U)p half a page. 'w' goes forward a (w)ord, 'W' goes back a (w)ord, likewise 'n' goes to the (n)ext match of the current search expression, 'N' goes to the (N)ext in the opposite direction. h,j,k and l were chosen because ^H is backspace (left), ^J is a newline (down), ^K is a vertical tab (up) and ^L is &mdash well, OK, ^L is a form feed, so that doesn't really follow. However, if you'd ever used an ADM-3A (a popular serial terminal BITD), those keys had arrows printed on them that showed those directions. I'm not sure why they were there originally, but I'm guessing there was some prior art there that vi followed. Likewise with the / (search) and : (enter command mode) characters: those were brought in from the original Unix 'ed' editor (and so were most of the commands). About the only 'arbitrary' command I can think of is the '%' key — I've never made a connection between it and parentheses.
Now if only I could figure out how to make vim map CAPS LOCK to CONTROL when it starts up, I'd have it made...
the second paragraph makes a good point or two Yeah, and a bad point or several. "plague of security breaches and spread of worms"? I don't think the internet has had its security breached, or suffers from worms. Unprotected systems may suffer from this, but blaming the internet for it is like blaming the streets for drive-by shootings. And "fragile routing protocols"? IP is the canonical example of a robust routing protocol. If an intermediate node drops off the net, IP will find a new route. It may not always be the best route, but we're discussing fragility, not efficiency. A protocol that routes your packets from New York to Miami via Winnipeg when some backhoe operator takes out the bulk of the fiber between Philadelphia and Washington DC doesn't sounds too "fragile" to me. The fact that it may continue to do this even after the fiber has been restored is unfortunate, but hardly a sign that it is "fragile".
As to the rest of the paragraph, it's just as misguided. When was the last time you weren't able to connect to the internet due to "equipment failures" other than your own CPE? Or the last time you couldn't get to a site because there was no route to it? Personally (and I use the internet every day, and have for the last 7-8 years, just like almost everyone else on this site), I haven't seen it. The only time I get "Cannot connect to site" is when a page tries to access doubleclick, which I have routed to 127.0.0.1 in/etc/hosts. And the only time I couldn't get on due to equipment failure, all I had to do was power-cycle my DSL modem. Oh, and since I implemented a cacheing DNS server, my response time is quick enough that I don't notice if it's variable or "unpredictable" — whether a site responds in 0.1 or 0.6 seconds, it looks the same to me.
This article sounds like propaganda from the Committee for a More Profitable Internet.
I wonder if there is no midrange or treble in the song at all. There is, you just can't hear it because it gets filtered down before the whole thing is compressed to about 50dB of dynamic range. Almost a waste of plastic to put them on CDs, cassettes would sound about the same (and wouldn't have a skipping problem, which I've noticed on some cars with [presumably] CD players co-located in the trunk along with the four woofers and two "800 watt" (16 clean) amps...)
What makes you think a compiler will be able to do it better than a human? Guess you missed the mention of "ultra-wide execution units" in the summary. Think Itanium, think of scheduling in terms of "bundles" of simultaneous instructions, ponder how to group the instructions in your multiple threads of execution so that if one thread branches, the code for the other threads is in the same bundle as the code the one thread branches to. I'm sure these chips will run fast, because they won't have to worry about coordinating "in flight" instructions, or keeping register scoreboards — that's all on the compiler writer's plate now. And good luck working it out by hand. I wouldn't bother writing an inner loop in assembler if it was more than a couple hundred lines today; on one of these chips, I wouldn't bother if it was more than twenty or thirty. Snag a copy of the Itanium architecture and instruction set manuals and try scheduling a few instructions for yourself.
I'm almost wondering if we'll finally see VMs starting to edge our statically-compiled code on these chips. Something that could cons up a dynamic instruction stream based on the current task mix might be able to edge out some frozen execution unit that had its code generated based on the idea that it was the only job on the box. If it's too difficult to write a good compiler, then compiled code loses its edge and interpreted code looks better. Of course, the interpreter still has to be compiled, so the performance gap may remain, but something with a sufficiently small run-time (Scheme or Smalltalk, maybe, forget JVMs and the CLR) might see an advantage. How big is the Erlang run time, I wonder...
You are aware that you're sitting in front of the most multi-function device ever conceived of by Man, right? Which is fine, because I'm (as you say) sitting. When I'm leaving a meeting and need to call someone to tell them that I'm leaving early/late, I really don't want to have to negotiate a bunch of menus to pull up my address book to tap somebody's number to make the call. Or if somebody calls me to get somebody else's contact info, I could do without telling them "Hold on...", then searching for the info, then reciting some of it, pull the phone away from my ear to look at the screen, memorize the next line, stick the phone up to my ear, recite the line, pull the phone away...
Not to mention I'd rather have a skinny little phone I can drop in a pocket that's just a phone instead of some bulky yeah-but-I-can-play-Quake-on-it wonder gadget that requires a holster. I wouldn't mind having a slick little tablet that has my address book, calendar, project notes, and e-mail viewer (and Quake), but that's the kind of thing I'd just as soon leave in a folio or something, since I almost never need that information when I'm walking down the street, so if it's not immediately handy it really doesn't matter. My phone is almost exactly the opposite -- if I can't grab it and make or answer a call within a couple of seconds, I really question the utility of owning it.
That's why I still carry my own cell phone, even though the company provides me with a BlackBerry. Due to corporate security policies, all BBs must be password protected. That means if I want to make a call, I first have to enter at least eight characters, at least one of which must be a capital letter and at least one of which must be a number. Try doing that on one of their funky 20-key QWERTY keypads while hustling down a busy hallway...
The "It's not what you know it's who you know" slogan is really getting old. Uh huh. But "working for the man" is hip and fresh? And then there's the term "old fogies"...
Seriously, I've gotten lots of work because I've been recommended by either clients or coworkers. I've also gotten lots of work because I had the right skills (or at least buzzwords) listed on my resume. Neither one is exclusive of the other, and you need both. If you've got an in-demand skill, then you'll probably find it fairly easy to move from position to position. However, if demand for that skill wanes (or the supply comes up sharply), then you're going to have to have some other way to distinguish yourself. That's where having a recognized name comes in. It's also good to have if you like to keep up with current technologies. Somebody starting a new SOA project may not have any idea how to evaluate someone's SOA knowledge level, but if they see your resume and remember you did good work on the $XYZ project, then they're much more likely to call you rather than somebody else who has XML, XSLT and Web Services listed as part of their skills.
And don't forget, it's about soft skills too. Anybody can say "I work weekends if the project gets behind", but if a manager remembers getting a status report from you at 3AM on Monday, that's proof. Of course, they'll also remember if they came in on Saturday and found you deathmatching or using the project's copy of DreamWeaver to redo your band's web site, so it cuts both ways...
Investing in teachers will have more of a benefit I agree, but it's harder to do something like raise salaries or benefits, since that requires perpetual funding. While I understand you said "bonuses", which wouldn't require additional funding, with that scheme you need someone to establish bonus performance criteria, then someone to collect and monitor the performance, and someone to issue the bonuses (after providing reports to all concerned parties), and someone to provide oversight — that's probably enough work for a small-to-medium sized administrative consulting company, which will either cost more money (nowhere near $38M, granted), or decrease the bonus pool substantially.
I did not RTFA, but I hope that it uses "iPod" as a euphemism for DAP. I see no compelling reason to pay the premium for an iPod when a variety of equivalent (for this purpose) players are available for less money. Unless, of course, Apple is offering some kind of educational discount, but it would have to be pretty hefty to make the iPod a compelling choice.
I was going to say I do the same thing, but then it occurred to me that the last time I upgraded, I switched from NetBSD to Solaris, so I didn't really copy anything over. And while I've attempted to keep everything in $HOME, I set up a caching DNS server, so I'd need to save that; I had to make some tweaks to the Apache startup script, so those need to be changed; all the final tweaks to my web sites that I did under htdocs that never got copied back; plus my PostgrSQL databases; plus all the jar files in the Tomcat common directory; thank goodness I was at least able to keep all my Ruby stuff in one place!
Hmmm, it's been a couple of months since I did a backup at home...I think I'll do one tonight....
many windows apps behave badly without the click to focus behavior. Really? I've used this with XP ever since it came out, and I don't think I've run across any apps that have a problem with it. Granted, most of the stuff I use is "name brand" (the MS staples, Lotus Notes, etc.), but I do use NetBeans and Vim quite a bit, as well as Apache/Tomcat, and of course FireFox. None of these apps have ever given me any trouble focusing.
I don't doubt your experience, I'm just curious what kind of apps you use that have problems with auto-raise.
the problem is that it's advertised as unlimited I'd be surprised if this didn't change. Back when "dialers" had "idle timers", unlimited connectivity was a draw. Now, people think "unlimited" means bandwidth, and there's a fixed amount of that available[0]. I think sooner rather than later we'll start seeing little disclaimers like "Unlimited refers to connection availablility, actual usage may not exceed NNN/month". This will also be explicitly spelled out in contracts. All "home-grade" services will have some kind of cap, and if you want more than that you'll have to pay for a business-class account (just like you do to get a static IP address most places).
I would not be surprised to find that McNealy is still backing this Um, why? Sun certifies several of their servers and workstations as being Linux-compatible. Plus, now that Solaris is free-as-in-beer (and with cheaper support contracts than RHEL), it isn't like Linux is really costing them any deals. The only way I could see this helping Sun would be by eliminating some systems that Linux supports that Solaris doesn't, which isn't a very big market segment (maybe some supercomputers and the odd non-x86 server). I think Sun stands to gain more by partnering with Intel & IBM to port Solaris to their processors (which they are doing) than by somehow eliminating or encumbering Linux.
Unless, of course, they can take down AIX with the Linux decision...
Even when you are 100% sure that you are in the right, you still have to pay $250 an hour to your lawyer Just so. Not to mention, once she's a legally "involved party" in the case, SCO may be able to get the judge to issue a gag or restraining order to prevent her from posting any more commentary. Of course, in that case [sorry], maybe PJ's long-lost distant relative "TJ" could step in to keep the community informed (while having no actual contact with PJ, of course)...
The impication [sic] is that some atheists are religious and some are not. Sorry, it has to be one or the other. No it doesn't. Some Christians are religious, and some are not. They'll profess to believe in God and Jesus and the Bible, but they don't go to church, they don't pray, they don't follow the Golden Rule or display any other manifestations of Christendom.
A religious atheist would be one who doesn't just not believe in God, but who believes there isn't a God. If God suddenly appeared before a "non-religious" atheist and somehow proved beyond a shadow of a doubt that he was real, the atheist would do a quick religious 180 and shout hosanna. A "religious" atheist would stay in complete denial, probably shouting him/herself hoarse denying miracle after miracle.
if you don't have pointers, it's really unintuitive to have arrays starting at 0 Not really, I typically find that when calculating an offset into a 1-based array, I have to add one to whatever offset I come up with to compensate for the base. Using zero-based arrays tends to result in more natural solutions.
There was that stuff they had back in the 70s, can't remember the name any more. It was a form of liquid Teflon, I remember they showed a rat with a weight tied to one leg being dropped in a beaker full of the stuff (after it had been oxygenated) and apparently it stayed in there for several minutes and suffered no ill effects afterwards.
One problem with the synthetic blood is that it wasn't as efficient at oxygen transfer as real blood, and ISTR there were some issues with the body's ability to excrete it (either it didn't do it well, so the stuff stayed in your bloodstream until it was removed, or the body found it extremely easy to filter out, so it didn't stay around long).
From a purely practical standpoint, I find the position of "My proven methods are not working, but I'm not even going to TRY you're unproven methods because I'M SURE they won't work either" to be a particularly weak one Not at all. The actual consideration would be more akin to "Every minute that slips by reduces the possibility of finding something. What's the next most reasonable method to use?". I don't believe that the WoC examples I've heard of translate well to this situation, as they all were based on people's guesses of things they all had observed.
If you tell 200 people what you had for lunch, using WoC methods could probably provide you an accurate estimate of how many calories you consumed. If you told those same people that you went out for lunch and asked them to guess where you went, I doubt their "communal wisdom" would pinpoint (or even reasonably bracket) your chosen eatery.
A self-deforming input device that could form itself into buttons or whatever would be a neat solution to reconfiguring your input device. Too bad I have no idea of how that could be accomplished. Apparently with sound and pressure on a flexible touchscreen, if you believe this. Not something I'd want in my car, but then I bought it because I love to drive, not because I wanted a portable thump dispenser....
Frankly, I haven't had a radio in my car for several years now, because all of the nicer ones had horrible interfaces. When a good song comes on, I want to CRANK IT, not press and hold a button for five seconds while the volume comes up. Gimmie a knob with no more than three hundred degrees of rotation, and with a little heft. Oh, and with some kind of reference line on it, so I can tell what the volume is set at before I start the car (so I don't blast my passengers or neighbors), Setting the volume entails selecting a point along a continuum, which is much better accomplished with an analog control. Save the pushbuttons for presets and other discrete-choice functions. I don't miss analog tuning knobs (since you rarely want to tune to a random frequency, station selection boils down to discrete selection too), but give me a volume control I can tweak, twiddle or crank.
I see most of the major manufacturers are going back to analog volume controls, I may just have to check them out. Something tells me, though, that they're probably going to all be weightless little infinitely-spinning knobs with detents. Oh well, it's a start.
If they'd installed Solaris correctly, they'd have had the same out-of-box results Well, it's not their fault the Solaris installer works correctly. Maybe if it had a defective one like Fedora...
(From TFA, describing the Fedora testing):
Despite the previous configuration prompts, the chosen servers [FTP, Mail, NFS, SSH, Samba, HTTPS, telnet and HTTP] were still not enabled. Fedora's so security-conscious, it won't start services that might get compromised! Next release, they may improve security by simply not shipping any network drivers...;)
Did anyone else find it odd that they went out of their way to load the entire Solaris distribution, but cherry-picked the Fedora options they loaded? No, I guess I'm not either... Anyone who claims that all the world's computers run either Windows, Linux or OS X isn't someone who's terribly clued in to begin with.
compute-intensive tasks like emulation can be reliably executed with Java without sucking? Actually, I would think Java would do pretty well at this. Figure the bulk of the emulator is going to be decoding and executing instructions, and the most commonly-executed instructions are going to be executed millions of times, so the hotspot compiler will tag those code paths and optimize them early on. Plus, memory is probably going to be implemented as a huge linear array, allocated at startup, so there's not going to be a lot of object allocation/freeing, so the GC probably won't be a factor. I/O will probably be a bottleneck, but that's true with any emulator.
If the implementors are at all clever, there really shouldn't be too much performance loss.
Slashdot Mirror
The big challenge facing Computer Science right now is to figure out how best to handle parallel processing and concurrency. The megahertz wars are over, you're not going to see CPUs get much faster, they're just going to cram more of them into one chip. Problem is, how do you break your code down so it runs on multiple chips? And what can you do to current code that'll allow it to take advantage of multiple cores/CPUs without rewriting it? Some kind of static analysis, maybe re-target it to a virtual machine that can parallelize it during execution, maybe some other abstraction layer that makes an 8-core machine look like a single core that runs 8 times faster? No amount of programming is going to solve this, you're going to have to go back to square one and re-think how software should be designed, what the languages will look like, what facilities an OS should provide — not the kind of stuff some "go faster" programmer is going to sort out over the weekend.
If you really want to see how CS improves programming, check out Jon Bentley's Programming Pearls books. And if you want to see how engineering can make your car go faster, look at Fred Puhn's How to Make Your Car Handle or Carrol Smith's Tune to Win (or any of his other racing engineering books).
"Programming is to Computer Science as telescopes are to Astronomy" (paraphrased from someone who said it better than I). If you actually take any computer science courses, you find they're more about things like measuring algorithm efficiency and finding ways around NP-hard problems (and how to tell if a problem is merely NP-hard or if it's really NP-complete). Sure, there's programming involved, but once you get past the sophomore-level courses you do stuff like design virtual memory systems or network protocols, and then maybe write a simulator to verify your theories or prove that you can meet your performance goals.
Maybe things have changed since the last CS class I took, but I definitely recall spending more time on the final drawing boxes and arrows than scribbling down code...
I wonder if it'd be worthwhile to have your browser set a default maximum expiration date? Any cookie set would have its expiration date set to MIN(requested-date,browser-max-exp-date). Set it for a day, a week, maybe have a keyword "session" to force all cookies to be session cookies. Hey, if a site wants to stick some data on my machine, I think I should have some say as to how long it sticks around! Hmmm, might be time to take a look at the Firefox cookie code, that shouldn't be too hard to set up...
Yeah, I still miss my Commodore SuperPET. I saw one when they first came out, and had to have one, as it was the only machine I'd ever seen that could run APL using the actual APL character set. It also had 1MB floppies (back when the PC had just switched from 160 to 180K). It was also the first dual-processor machine I ever saw (had both a 6800 and a 6809, I believe).
Twenty or so years later, I picked one up for $50. Hauled it home, fumbled with the not-quite-GPIB cables, finally got everything hooked up, and booted it up. Played with APL for a couple of nights, sketched out some code for a shared-memory multiprocessing scheme (6809 could pass tasks off to the 6800, communicating via a bank-switched block of memory). But then I got tired of all the limitations. While it had impressive performance for a machine of its day (its disk I/O was at least twice as fast as my 200MHz PPro's floppy), its day was long past, and when you got right down to it, I just wasn't that nostalgic for a machine with a firmware monitor. I was when I was debugging VB code, but not when I actually had a spare minute to go play. So I finally sold it when it was time to move (got my $50 back!), but I still wish I had kept it. Now, if I could find an original TRS-80 (my first computer), I'd definitely pick that up...
Now if only I could figure out how to make vim map CAPS LOCK to CONTROL when it starts up, I'd have it made...
As to the rest of the paragraph, it's just as misguided. When was the last time you weren't able to connect to the internet due to "equipment failures" other than your own CPE? Or the last time you couldn't get to a site because there was no route to it? Personally (and I use the internet every day, and have for the last 7-8 years, just like almost everyone else on this site), I haven't seen it. The only time I get "Cannot connect to site" is when a page tries to access doubleclick, which I have routed to 127.0.0.1 in
This article sounds like propaganda from the Committee for a More Profitable Internet.
I'm almost wondering if we'll finally see VMs starting to edge our statically-compiled code on these chips. Something that could cons up a dynamic instruction stream based on the current task mix might be able to edge out some frozen execution unit that had its code generated based on the idea that it was the only job on the box. If it's too difficult to write a good compiler, then compiled code loses its edge and interpreted code looks better. Of course, the interpreter still has to be compiled, so the performance gap may remain, but something with a sufficiently small run-time (Scheme or Smalltalk, maybe, forget JVMs and the CLR) might see an advantage. How big is the Erlang run time, I wonder...
Not to mention I'd rather have a skinny little phone I can drop in a pocket that's just a phone instead of some bulky yeah-but-I-can-play-Quake-on-it wonder gadget that requires a holster. I wouldn't mind having a slick little tablet that has my address book, calendar, project notes, and e-mail viewer (and Quake), but that's the kind of thing I'd just as soon leave in a folio or something, since I almost never need that information when I'm walking down the street, so if it's not immediately handy it really doesn't matter. My phone is almost exactly the opposite -- if I can't grab it and make or answer a call within a couple of seconds, I really question the utility of owning it.
That's why I still carry my own cell phone, even though the company provides me with a BlackBerry. Due to corporate security policies, all BBs must be password protected. That means if I want to make a call, I first have to enter at least eight characters, at least one of which must be a capital letter and at least one of which must be a number. Try doing that on one of their funky 20-key QWERTY keypads while hustling down a busy hallway...
Seriously, I've gotten lots of work because I've been recommended by either clients or coworkers. I've also gotten lots of work because I had the right skills (or at least buzzwords) listed on my resume. Neither one is exclusive of the other, and you need both. If you've got an in-demand skill, then you'll probably find it fairly easy to move from position to position. However, if demand for that skill wanes (or the supply comes up sharply), then you're going to have to have some other way to distinguish yourself. That's where having a recognized name comes in. It's also good to have if you like to keep up with current technologies. Somebody starting a new SOA project may not have any idea how to evaluate someone's SOA knowledge level, but if they see your resume and remember you did good work on the $XYZ project, then they're much more likely to call you rather than somebody else who has XML, XSLT and Web Services listed as part of their skills.
And don't forget, it's about soft skills too. Anybody can say "I work weekends if the project gets behind", but if a manager remembers getting a status report from you at 3AM on Monday, that's proof. Of course, they'll also remember if they came in on Saturday and found you deathmatching or using the project's copy of DreamWeaver to redo your band's web site, so it cuts both ways...
I did not RTFA, but I hope that it uses "iPod" as a euphemism for DAP. I see no compelling reason to pay the premium for an iPod when a variety of equivalent (for this purpose) players are available for less money. Unless, of course, Apple is offering some kind of educational discount, but it would have to be pretty hefty to make the iPod a compelling choice.
I was going to say I do the same thing, but then it occurred to me that the last time I upgraded, I switched from NetBSD to Solaris, so I didn't really copy anything over. And while I've attempted to keep everything in $HOME, I set up a caching DNS server, so I'd need to save that; I had to make some tweaks to the Apache startup script, so those need to be changed; all the final tweaks to my web sites that I did under htdocs that never got copied back; plus my PostgrSQL databases; plus all the jar files in the Tomcat common directory; thank goodness I was at least able to keep all my Ruby stuff in one place!
Hmmm, it's been a couple of months since I did a backup at home...I think I'll do one tonight....
I don't doubt your experience, I'm just curious what kind of apps you use that have problems with auto-raise.
Unless, of course, they can take down AIX with the Linux decision...
A religious atheist would be one who doesn't just not believe in God, but who believes there isn't a God. If God suddenly appeared before a "non-religious" atheist and somehow proved beyond a shadow of a doubt that he was real, the atheist would do a quick religious 180 and shout hosanna. A "religious" atheist would stay in complete denial, probably shouting him/herself hoarse denying miracle after miracle.
There was that stuff they had back in the 70s, can't remember the name any more. It was a form of liquid Teflon, I remember they showed a rat with a weight tied to one leg being dropped in a beaker full of the stuff (after it had been oxygenated) and apparently it stayed in there for several minutes and suffered no ill effects afterwards.
One problem with the synthetic blood is that it wasn't as efficient at oxygen transfer as real blood, and ISTR there were some issues with the body's ability to excrete it (either it didn't do it well, so the stuff stayed in your bloodstream until it was removed, or the body found it extremely easy to filter out, so it didn't stay around long).
So will this converted blood be referred to as being in Antigen-Neutral Distribution Format?
If you tell 200 people what you had for lunch, using WoC methods could probably provide you an accurate estimate of how many calories you consumed. If you told those same people that you went out for lunch and asked them to guess where you went, I doubt their "communal wisdom" would pinpoint (or even reasonably bracket) your chosen eatery.
Frankly, I haven't had a radio in my car for several years now, because all of the nicer ones had horrible interfaces. When a good song comes on, I want to CRANK IT, not press and hold a button for five seconds while the volume comes up. Gimmie a knob with no more than three hundred degrees of rotation, and with a little heft. Oh, and with some kind of reference line on it, so I can tell what the volume is set at before I start the car (so I don't blast my passengers or neighbors), Setting the volume entails selecting a point along a continuum, which is much better accomplished with an analog control. Save the pushbuttons for presets and other discrete-choice functions. I don't miss analog tuning knobs (since you rarely want to tune to a random frequency, station selection boils down to discrete selection too), but give me a volume control I can tweak, twiddle or crank.
I see most of the major manufacturers are going back to analog volume controls, I may just have to check them out. Something tells me, though, that they're probably going to all be weightless little infinitely-spinning knobs with detents. Oh well, it's a start.
(From TFA, describing the Fedora testing): Despite the previous configuration prompts, the chosen servers [FTP, Mail, NFS, SSH, Samba, HTTPS, telnet and HTTP] were still not enabled. Fedora's so security-conscious, it won't start services that might get compromised! Next release, they may improve security by simply not shipping any network drivers...
Did anyone else find it odd that they went out of their way to load the entire Solaris distribution, but cherry-picked the Fedora options they loaded? No, I guess I'm not either... Anyone who claims that all the world's computers run either Windows, Linux or OS X isn't someone who's terribly clued in to begin with.
If the implementors are at all clever, there really shouldn't be too much performance loss.