I thought the same thing. The second point in the first sentence of the student's Project Overview demonstrates that the students are trying to talk smart about music without actually understanding music. It's worse than that. They offer up a bunch of mathematically derived mumbo jumbo graphs, but miss the math behind the music: If you start with one reference tone (Do), the musical fifth ('So') is 1.5x the frequency, the octave ('Do') is 2x the frequency, and the musical fourth ('Fa') is 2/1.5 the frequency. There's math behind the entire structure of simple pleasing vibrations.
So it's pretty much some fancy shmancy mathlike pictures of music data by self proclaimed internet math wizards that don't understand music or math. I would fail them based on the first sentence.
They could have have gotten the answer right if they'd wiki'd their initial premise:
http://en.wikipedia.org/wiki/Chord_progression#Three-chord_progressions
Of course then any music pedant notices that wiki has no G note in their G chord. It's almost like they let any bozo edit it.
I've been trying for over 6 months to get USPS to pay on a computer that was evidently stolen by a postal employee. They don't pay. Package was registered, signature required, insured, and they don't contest that they never delivered. Still the beaurocracy doesn't stop. Here's what I've learned: USPS doesn't really track anything; they have little idea whatever happened to a package or where it went when. Postal employees know this and the bad ones will cherry pick expensive packages. When a package goes missing, they don't pay; they require you to wait before filing. By the time you file, they will forevermore tell you they don't have any information on your package & insurance number; their computer system finds the number but all the information is 'archived' and the archive has no online access and requires days to look up, which they say they will do, but they don't. Eventually, they will tell you to resubmit all your paperwork, and you repeat this process. Also, when you insure a package for $1000, you might assume they will pay $1000 when they admit they never delivered, but no. They will require you to furnish proof of this value, and they will evaluate whether your proof warrants the insurance value you bought. Then they will continue to not pay. Good luck escalating, because all escalation procedures will require that they get back in touch with the people that have been ignoring the problem all along.
Without a spectrum analyzer, you are blind to the real problems. With a S.A. you can see the signal that your radio is trying to decode, and you can see what is obscuring your signal. With a little practice, you learn to see the signature of every radio/antenna combo. In the frequency domain, a radio looks like an ubrella centered on the frequency of your channel, and in the best case, you'll a big umbrella for your radio and a weaker umbrella under it that is the access point talking back.
If all the radios around you are tuned to non-overlapping channels (1,6, 11) and you look "down the spectrum through time" you have 3 distinct bands that all the radios can tune and demodulate easily. When people use an intermediate channel like 3, it doesn't get them away from the channel 1 signal so much as it fills in the spectrum that makes the neighbor channels tunable. Intermediate channels just interfere with 2 channeles, it doesn't get you clear of either. Tuning is difficult when the spectrum is filled out without separation, it's like trying to listen to a distant flute concerto while you're in the shower.
The best way I know to really tune your wireless environment: Run Kismet on a linux laptop in one window, and run Wi-spy Chanelyzer in a vmware unity windown next to it. The USB wispy works perfectly in a virtual machine, and you can correlate the Kismet traffic to the radios you see in the spectrum graph. Makes it real obvious what the problems are, and what the best solutions will be, assuming you have good solutions available. In a crowded environment, the unregulated microwave spectrum can be a noisy mess, but until you see what a radio sees, your actions are likely to be little better than random changes.
By the way, you can't really trust your wireless card to give you a good picture of the spectrum. It misses all noise that is not its own protocol, it tunes 1 channel at a time and scans and samples slowly, and even when it recognizes a signal the strength that it reports to the driver may not be calibrated in a meaningful way.
The Alps keyswitches you like were found in many top quality keyboards of that era, and I do agree that these are about the best ever done. Northgate used them, so did the keyboards of SGI, Control Data, Dell, and others. I know this because I used to cruise Silicon Valley thrift shops, and on quite a few occasions I would bring home an excellent used keyboard, only to realize it was exactly the same as all my other favorites.
I also have a bunch of IBM model M's. I like them, but they make quite a commotion when you are cranking code, and they are comparitively high effort, almost like playing a piano. I did some tests and found I had better speed and accuracy with the Alps, which requires no deliberate effort and is less fatiguing for long term use. I still keep model M's on my servers both for nostalgia and because they are just fun to use, plus the keyboard change at a server console reminds me to be a bit more deliberate.
I used to keep an old pre-AT IBM keyboard around just to remind me of what a keyboard could be. I think it was an XT model, smaller than the modern 101's but it weighed about 10 pounds and every keyswitch was a beautiful machine. The action was less jarring than any model M, still klicky but oh-so-fine, felt like precise bearings. No way I know to make this work on any modern keyboard controller and the layout isn't what I would prefer in this world, but this is what keyboard would be if people payed hundreds of dollars for them.
I have an SXGA t22 (made in Mexico by IBM) and a UXGA t43p (designed by IBM, made in China by Lenovo) and to the extent there is any quality difference, the Lenovo wins. Slightly better keyboard, trackpoint, display, case solidity, etc. My t43p is a quantum leap ahead in features and crushes the t22 in CPU, video, memory, disk i/o, battery life etc. I can't help but notice that the t43 scrolls a terminal window faster with the output of a kernel build than the t22 can dump text using 'cat'. (Stupid benchmark I know!) For surfing the web and data entry, there is surprisingly little difference between them and I forget which one I am using.
FWIW, both systems triple boot Windows, Kubuntu, and FreeBSD. I had to do a bunch of minor tweaking to get Linux & BSD working well on the t22 (driver config to keep X from freezing, etc). On the t43 Kubuntu and FreeBSD basically 'work' out of the box but there's a bunch of tweaking the user must do to duplicate the functionality of the generally excellent IBM windows software; things like hibernation and battery-longevity-enhanching behaviour (etc etc...) are left as an exercise for the user. I don't mind too much but I have better things to do with my time and it takes a while to get it all nearly as good as Windows. (And for the defensive, obviously there are things I prefer about Linux & BSD.)
If anyone believes Lenovo is very different from past Thinkpad factories, your experience is different than mine. My T22 has always been excellent. The T43 is a lot more of the same, just a bit better.
I am surprised that all the posts I read miss the point of Skype's P2P. It's not the file transfer utility (that gets a free ride), but it's the NAT traversal itself. If you have good bandwidth and a public IP, you become the connectable server for unconnectable pairs of private IP's.
This is what makes Skype brilliant but also undesirable. Skype scales because the company doesn't have to run servers to compensate for all the broken networks of its clients, and if your network isn't broken, you host Skype's business. This is also the reason why Skype requires encryption and isn't open source; as long as I'm hosting your calls of course I'd be inclined to have me a listen.
> (bug) where after a long period of inactivity the screen won't come back up?
I fixed this for my thinkpad on Kubuntu breezy by disabling ACPI and enabling APM via GRUB commands. Can't be more specific now since I'm at a different system, but I think I found the answers on Ubuntuforums.com, and lots of other good answers are there too.
FWIW I'm mostly a freeBSD user and I regularly look at various linux's to try to get the cutting edge features that appear first there. Kubuntu needed a few more tweaks than some to become solid and functional, but I have found it to be one of the best supported distros and the only one where all my nits were easily fixable. For the first time I have all the features I have ever wanted from linux and no deal breaking nits. I have my Mom running Mepis, but for me Kubuntu is the first 'keeper' linux.
I build quiet PC's with fanless video cards. One of mine has a Matrox G550, another has an nVidia Quadro4 550XGL. I run these at 1600x1200 analog to a Samsung 213t LCD display, and Samsung includes an "auto pattern" program that displays a black/white checkerboard pattern that is optimal for tuning the LCD a/d clock to the card.
The nVidia display for this is dead sharp and visually quiet, indistinguishable from DVI. The Matrox isn't generally bad, but this kind of display shows a lot of scanning flicker, which I surmise is indicative of clock jitter or less crisp D/A's. There's just no comparison of the analog video quality, nVidia is way superior.
Maybe quadro4's (which are intended for engineers and CAD) have better DACs and clocks than their consumer cards, who knows? And you could argue that this isn't a fair test, but IMO it's fair enough; they were both nearly the last generation of mainstream fanless cards from these manufacturers, and the nVidias are cheaper on eBay. Besides video quality, the nVidia smokes the Matrox on 2D and 3D speed, and the Matrox can't even do DVI at 1600x1200.
If you would argue that a better choice would be a Parahelia, I might agree (though it's way more $$) but then you should probably compare to a modern nVidia and Matrox loses badly again on performance and doesn't play nice with Linux.
IMO Matrox hasn't been competitive for years and the reputed superiority of Matrox analog quality is just an outdated myth.
I always had bad experiences with stable machines becoming unstable after installing Creative's drivers, and never liked that you can't seem to just install what seems like a driver but have to screw up your system with what seems more like an application suite / (buggy) driver combo. What's worse is that despite the bloat Creative's stuff never has the features that I actually want in a sound card.
The only salvation for my SB cards has been the 'kx project' drivers:
If you are a musician these drivers have the features you actually want; WDM, ASIO, GSIF - other than the sound floor (on my SBLive) they make the card competative with a mid level music card. No bloat and I've found the driver to be solid, though the UI is rather yucky.
Ack, replying to my own post, my $30 PSU is a Fortron, another good inexpensive choice. Sorry bout that. In retrospect maybe all my sparkles were loud, but fine units all in my experience.
I was surprised to get a HEC bundled in a cheap case and had read the good reviews of these. The unit runs quietly and has the heft and finish of quality so I plunked it in a server and the unit has been burp free for a year and a half so far.
I have a couple of re-badged Sparkle units that came with quality cases (they say "SPI" for sparkle power, I forget if they explicitly say sparkle) and they come with loud high volume fans and near as I can tell no circuits that slow these suckers down. IIRC the power supplies also say something like "noise killer" but I think this is a warning, not a feature. I have replaced the fans in my OEM Sparkles with lower flow units (careful that the exhaust air is still cool under load) and never had any trouble with these units.
I also have a $30 Sparkle from New Egg with a 120mm fan and it's fairly quiet. So not all Sparkle fans are over-spec'd but anyway I can't be too critical of a cheap reliable power supply with excessive ventilation.
FWIW the HEC came in a 6A19 from servercase.com, this is a good case for modding for good airflow and quiet accoustics. And my favorite quiet cases (the 6919) are cheap there too. Plain but solid, good features and superior airflow.
anytime you connect through an ISP they can watch what you do.
Except that I pay my ISP and they have a privacy policy saying they won't sniff. I don't pay google except for the data they can mine from me and sell (umm, all my searches, incoming & outgoing mail, and now all routed traffic) and they explicitly state they will sniff, log, & possibly store this traffic. You really trust Google more than your ISP? Really? Based on their pre-IPO slogan?
When I use a coffee shop, the default route on my firewalled Freebsd laptop is an openvpn tunnel to my home router; tough to sniff or hack. But if I didn't do this, I would place Google last on my list of who I would trust as my router. Chances are very high the patrons and barristas don't have the competence or the interest in sniffing (it's stupid but has never been a big problem). If the shop put up an open AP on a DSL link they likely are in the same boat. If they use a commercial captive portal, the provider has a vested interested in not screwing up the relationship with a paying customer. But Google has a vested interest in evil here and they have stated as much in their policy.
I can only conclude that Google is really the WORST... VPN... ENDPOINT... EVAR! But what did you expect for free?
If you've got something to hide and the FBI has a warrant to tap your ISP's router, you're fairly screwed (and your ISP is not the worst of your problems) but I have high confidence my ISP does not have a cache of past traffic.
I expect that google does cache and mine everything that passes through them as well as everything they have ever found.
If you wanted to know what someone has been up to for the last decade, Google will likely have far better information than their ISP or the FBI.
Us geeks only control the lanuguage amongst ourselves
A geek is a circus freak that bites the heads off of live animals, so clearly you have already accepted other's terminology. Had you said "We the cognoscente..." I would find your grammer and terminology less objectionable.
Anyway, personally I find that BlueSpew has the right ring for this crap.
I cut my teeth programming the Atari ST way back when. Data Becker had a rocking book ("Atari ST Internals") that included a complete *commented* bios listing that they had disassembled and reverse engineered.
If you wanted to see how see how vertical blank interrupts or midi I/O was handled, it was all there. I think everybody that wrote midi sequencers back then reimplimented the interrupt handlers from the DB source code, and it was tight! To this day you can't get tighter midi on your 3 GHz machines than you could get on an 8MHz Atari.
My tools back then were really top notch; I loved Laser C (a big improvement on Megamax which was also excellent) and the DevPak assembler. I still have my original source floppies and documentation. Though my machines went to the thrift store years ago, I would still consider running these tools under emulation to teach a class on how computers really work.
The hardware was so accessible then. I still think the motorola 68k was the clearest machine to program in assembly we've ever had.
Though the parent is intended as a joke, based on my own thoughts from reading the works of Einstein I believe it to be more insightful than the article and most of the posts here that fail to understand wny time travel is romantic BS.
There really is no such thing as time, only the relative differences in the states of energy of systems, where energy is observed as velocity (integration of acceleration as time goes to zero) or relative positions in an arbitrary coordinate system.
To go back in time is to rewind the state of the universe, effectively to reverse the vector of all energy (and fields). You can create a perspective that can change the relative state of a simple system, but this function will not magically rewind the states of arbitrary systems, let alone the gestalt.
Any equation that involves time is actually measuring some other assumed relationship. Time was linear in Newtonian physics because it only solves for simple systems with a constant field, but there's a reason that better theories involve an integration with time going to zero.
I cringe when I read Hawking and he jabbers endlessly about not understanding why time seems to have a vector and then ponders the magic of some wonder field that reverses it all. It's clear that he read Einstein and he could do the math but he missed the big point entirely; time is a big zero, always now, but energy has a vector and the universe is a directed rotor that cannot be simply deconvolved.
ps Thanks for the universe, God, it's mostly working pretty well thought I have a few suggestions.
I'm running FreeBSD 4.11 on my servers and laptop. To this point I've avoided 5.x due to rumours of a performance downgrade for my single processor systems. (Well that & pragmatism, I am slow to change good systems.)
I have no great dissatisfactions with my 4.11 systems, but could make use of cardbus support on the laptop, and if 5.x's threading could improve the performance of samba, mySQL, postgreSQL, maybe java and apache (my primary server loads) that would be a win.
Is 5.4 ready for primetime for satisfied 4.x users? What are the real world performance implications at this time in terms of memory usage, I/O throughput, performance on my favorite server apps and as a GUI workstation? Is it as solid as 4.x has always been?
Though I tend to believe these things about as much as the amateur stories in Hustler, the story seems technically plausable and I have no doubt many people actually are this technically incompetent.
I used to work with a fellow whose hobby was to harass and attempt to discredit the Scientologists. They claimed he had posted their intellectual property on the internet, and when he was deposed he stated that their material was indeed available at 127.0.0.1 .
They quoted this verbatim in their claim against him, so it made it past whoever was responsible for fact checking the claim, and he boasted around the water cooler how he had bested the scientology legal staff. He explained 'localhost' on the stand and the claim was thrown out.
He was rather an odd fellow. I had quite forgotten him until some 10 years later I saw his name appear on slashdot in an appeal for money for his legal defense fund.
I agree that memtest86 is useful but not sufficient and that prime95 is much more throrough. Memtest confirms that patterns that have been set hold their state briefly, which is a good test against gross failures (and I have seen these).
But Prime95 confirms that no bit anywhere in nearly the complete memory space ever spuriously changes. I have seen plenty of memory that passes metest86 that fails prime95.
Based on my experience, Corsair will replace memory that fails prime95. Mushkin will NOT (despite a "lifetime" warranty); they basically told me that memory can't be expected to be 100% perfect all the time and that prime95 was too strenuous; if it passes memtest86 there will be no replacement.
My other modules (from Geil, Samsung, and a few old no-name sticks) have always been perfect. IMO it's unconscionable to sell untested ram given how hard it is to return.
I'm sure I don't know all of Google's products, but to the extent to which I use their services, they are not selling or otherwise distributing GPL derived software.
They do provide me with a variety of free, useful services in order to sell my eyeballs and Google's knowlege of my interests and habits to their customers, which are companies interested in potential consumers.
The GPL has led to a terrific and amazing revolution; our information infrastructure, which is one of this generation's most crucial assets and contributions, is effectively given to the public trust, rather than locked away as a tool of the powerful to exploit the rest. Furthermore, mankind can use the tools of the past to move forward, rather than reinvent the same tools over and again.
The usage of GPL software to provide services strikes me as completely within the intent of the GPL, and unlikely to spur the kinds of abuses that the GPL was intended to discourage. Would anyone benefit if everyone that provided a service derived from GPL'd code were required to pay? Hmmm, would anyone EVER then run a GPL'd webserver, or any other service run on a GPL'd OS? What users of GPL software would be liable, any whose services indirectly contribute to anything providing revenue? Such a GPL would be the death of free software.
I think your argument reflects your feelings on C. The object part of the syntax couldn't be simpler and expresses the messaging feature of the language clearly.
But you can't wish the C out of obj-C, that's one of the main attractions of the language. If you're going to write a codec (say) or an image filter, you use the object features to facilitate reuse of tight data munging code. Sometimes there are good reasons to prefer cstrings over string objects, and processor-native number types over number objects.
Objective-C is mostly about writing reusable C code. If you're big on OO dogma and objects everywhere it's a poor choice. But there's a good place for well structured, late-bound C too. Obj-c enables OO programming, but it deliberately does not impose it.
If you want pure and focused OO, you have your pick of a dozen flavors, but none will be the efficient system programming language the objc is.
Java versus Obj-C is rather a religious argument, so I would not argue that your preference is wrong given your priorities.
But for believers (and those in the know:-) obj-c is a terrific lightweight and very capable tool. The programs can be, in practice, as fast and small as C applications, yet you can still design abstract api's that easily support inheritance and polymorphism. It's not an academic argument; Obj-C has been key to the design of one of the best application development environments for 15 _years_. And I would still consider it one of the strongest choices for such development, starting clean-sheet today.
Obj-C is in the OSX kernel, and it was in the NeXT kernel a decade ago. It's in the Mac display system. You would be crazy to use Java in these circumstances.
I would not argue that Obj-C is one of the better OO languages for applications where true OO is necessary or useful. The biggest issue is that memory management is in your face, for better and for worse. There are plenty of situations where such control is useful or critical, and much as I love the convenience & safety of a garbage collector, a large gc app in my experience is hell on virtual memory, whereas it is possible to have a very light footprint with obj-c.
Your argument on the expense of the message dispatcher doesn't hold water. A message dispatch is 3x as expensive as a C function call; compared to the cost of most algorithms, it's totally in the noise. Even if it were substantial in a heavy loop, you could indirect through the message's function for a cost no higher than normal C. This is trivial though rarely useful. You're arguing for a Java implementation, where the cost will be higher even before the garbage collector takes its toll on the CPU and VM.
Objective-C is a terrific choice as a system programming language, and I would consider it a strong choice for most problems sets you would otherwise choose C or C++ for, and it wallops these for extensibility and reuse. It's a poor choice for problems where you would have a good reason to choose Java or Smalltalk.
Slashdot Mirror
I thought the same thing. The second point in the first sentence of the student's Project Overview demonstrates that the students are trying to talk smart about music without actually understanding music. It's worse than that. They offer up a bunch of mathematically derived mumbo jumbo graphs, but miss the math behind the music: If you start with one reference tone (Do), the musical fifth ('So') is 1.5x the frequency, the octave ('Do') is 2x the frequency, and the musical fourth ('Fa') is 2/1.5 the frequency. There's math behind the entire structure of simple pleasing vibrations. So it's pretty much some fancy shmancy mathlike pictures of music data by self proclaimed internet math wizards that don't understand music or math. I would fail them based on the first sentence. They could have have gotten the answer right if they'd wiki'd their initial premise: http://en.wikipedia.org/wiki/Chord_progression#Three-chord_progressions Of course then any music pedant notices that wiki has no G note in their G chord. It's almost like they let any bozo edit it.
I wish I was joking!
If all the radios around you are tuned to non-overlapping channels (1,6, 11) and you look "down the spectrum through time" you have 3 distinct bands that all the radios can tune and demodulate easily. When people use an intermediate channel like 3, it doesn't get them away from the channel 1 signal so much as it fills in the spectrum that makes the neighbor channels tunable. Intermediate channels just interfere with 2 channeles, it doesn't get you clear of either. Tuning is difficult when the spectrum is filled out without separation, it's like trying to listen to a distant flute concerto while you're in the shower.
The best way I know to really tune your wireless environment: Run Kismet on a linux laptop in one window, and run Wi-spy Chanelyzer in a vmware unity windown next to it. The USB wispy works perfectly in a virtual machine, and you can correlate the Kismet traffic to the radios you see in the spectrum graph. Makes it real obvious what the problems are, and what the best solutions will be, assuming you have good solutions available. In a crowded environment, the unregulated microwave spectrum can be a noisy mess, but until you see what a radio sees, your actions are likely to be little better than random changes.
By the way, you can't really trust your wireless card to give you a good picture of the spectrum. It misses all noise that is not its own protocol, it tunes 1 channel at a time and scans and samples slowly, and even when it recognizes a signal the strength that it reports to the driver may not be calibrated in a meaningful way.
The Alps keyswitches you like were found in many top quality keyboards of that era, and I do agree that these are about the best ever done. Northgate used them, so did the keyboards of SGI, Control Data, Dell, and others. I know this because I used to cruise Silicon Valley thrift shops, and on quite a few occasions I would bring home an excellent used keyboard, only to realize it was exactly the same as all my other favorites.
I also have a bunch of IBM model M's. I like them, but they make quite a commotion when you are cranking code, and they are comparitively high effort, almost like playing a piano. I did some tests and found I had better speed and accuracy with the Alps, which requires no deliberate effort and is less fatiguing for long term use. I still keep model M's on my servers both for nostalgia and because they are just fun to use, plus the keyboard change at a server console reminds me to be a bit more deliberate.
I used to keep an old pre-AT IBM keyboard around just to remind me of what a keyboard could be. I think it was an XT model, smaller than the modern 101's but it weighed about 10 pounds and every keyswitch was a beautiful machine. The action was less jarring than any model M, still klicky but oh-so-fine, felt like precise bearings. No way I know to make this work on any modern keyboard controller and the layout isn't what I would prefer in this world, but this is what keyboard would be if people payed hundreds of dollars for them.
I have an SXGA t22 (made in Mexico by IBM) and a UXGA t43p (designed by IBM, made in China by Lenovo) and to the extent there is any quality difference, the Lenovo wins. Slightly better keyboard, trackpoint, display, case solidity, etc. My t43p is a quantum leap ahead in features and crushes the t22 in CPU, video, memory, disk i/o, battery life etc. I can't help but notice that the t43 scrolls a terminal window faster with the output of a kernel build than the t22 can dump text using 'cat'. (Stupid benchmark I know!) For surfing the web and data entry, there is surprisingly little difference between them and I forget which one I am using.
...) are left as an exercise for the user. I don't mind too much but I have better things to do with my time and it takes a while to get it all nearly as good as Windows. (And for the defensive, obviously there are things I prefer about Linux & BSD.)
FWIW, both systems triple boot Windows, Kubuntu, and FreeBSD. I had to do a bunch of minor tweaking to get Linux & BSD working well on the t22 (driver config to keep X from freezing, etc). On the t43 Kubuntu and FreeBSD basically 'work' out of the box but there's a bunch of tweaking the user must do to duplicate the functionality of the generally excellent IBM windows software; things like hibernation and battery-longevity-enhanching behaviour (etc etc
If anyone believes Lenovo is very different from past Thinkpad factories, your experience is different than mine. My T22 has always been excellent. The T43 is a lot more of the same, just a bit better.
In Bind9 you don't have to return cached data, so though it happens by default you can turn it off ("additional-from-cache"):
view "internal" {
match-clients { internals; guests; };
recursion yes;
zone "." {
type hint;
file "bootstrap/cache";
};
zone "example.com"{
type master;
file "example-int.com";
};
};
view "external" {
match-clients { any; };
recursion no;
additional-from-auth no;
additional-from-cache no;
zone "example.com"{
type master;
file "example-ext.com";
allow-query { any; };
};
};
---------
I believe that should prevent bind from being too useful from the outside.
I am surprised that all the posts I read miss the point of Skype's P2P. It's not the file transfer utility (that gets a free ride), but it's the NAT traversal itself. If you have good bandwidth and a public IP, you become the connectable server for unconnectable pairs of private IP's.
This is what makes Skype brilliant but also undesirable. Skype scales because the company doesn't have to run servers to compensate for all the broken networks of its clients, and if your network isn't broken, you host Skype's business. This is also the reason why Skype requires encryption and isn't open source; as long as I'm hosting your calls of course I'd be inclined to have me a listen.
> (bug) where after a long period of inactivity the screen won't come back up?
I fixed this for my thinkpad on Kubuntu breezy by disabling ACPI and enabling APM via GRUB commands. Can't be more specific now since I'm at a different system, but I think I found the answers on Ubuntuforums.com, and lots of other good answers are there too.
FWIW I'm mostly a freeBSD user and I regularly look at various linux's to try to get the cutting edge features that appear first there. Kubuntu needed a few more tweaks than some to become solid and functional, but I have found it to be one of the best supported distros and the only one where all my nits were easily fixable. For the first time I have all the features I have ever wanted from linux and no deal breaking nits. I have my Mom running Mepis, but for me Kubuntu is the first 'keeper' linux.
I build quiet PC's with fanless video cards. One of mine has a Matrox G550, another has an nVidia Quadro4 550XGL. I run these at 1600x1200 analog to a Samsung 213t LCD display, and Samsung includes an "auto pattern" program that displays a black/white checkerboard pattern that is optimal for tuning the LCD a/d clock to the card.
The nVidia display for this is dead sharp and visually quiet, indistinguishable from DVI. The Matrox isn't generally bad, but this kind of display shows a lot of scanning flicker, which I surmise is indicative of clock jitter or less crisp D/A's. There's just no comparison of the analog video quality, nVidia is way superior.
Maybe quadro4's (which are intended for engineers and CAD) have better DACs and clocks than their consumer cards, who knows? And you could argue that this isn't a fair test, but IMO it's fair enough; they were both nearly the last generation of mainstream fanless cards from these manufacturers, and the nVidias are cheaper on eBay. Besides video quality, the nVidia smokes the Matrox on 2D and 3D speed, and the Matrox can't even do DVI at 1600x1200.
If you would argue that a better choice would be a Parahelia, I might agree (though it's way more $$) but then you should probably compare to a modern nVidia and Matrox loses badly again on performance and doesn't play nice with Linux.
IMO Matrox hasn't been competitive for years and the reputed superiority of Matrox analog quality is just an outdated myth.
I always had bad experiences with stable machines becoming unstable after installing Creative's drivers, and never liked that you can't seem to just install what seems like a driver but have to screw up your system with what seems more like an application suite / (buggy) driver combo. What's worse is that despite the bloat Creative's stuff never has the features that I actually want in a sound card.
The only salvation for my SB cards has been the 'kx project' drivers:
http://kxproject.lugosoft.com/index.php?skip=1
(sorry I don't know to enter a URL here...)
If you are a musician these drivers have the features you actually want; WDM, ASIO, GSIF - other than the sound floor (on my SBLive) they make the card competative with a mid level music card. No bloat and I've found the driver to be solid, though the UI is rather yucky.
Ack, replying to my own post, my $30 PSU is a Fortron, another good inexpensive choice. Sorry bout that. In retrospect maybe all my sparkles were loud, but fine units all in my experience.
I have a couple of re-badged Sparkle units that came with quality cases (they say "SPI" for sparkle power, I forget if they explicitly say sparkle) and they come with loud high volume fans and near as I can tell no circuits that slow these suckers down. IIRC the power supplies also say something like "noise killer" but I think this is a warning, not a feature. I have replaced the fans in my OEM Sparkles with lower flow units (careful that the exhaust air is still cool under load) and never had any trouble with these units.
I also have a $30 Sparkle from New Egg with a 120mm fan and it's fairly quiet. So not all Sparkle fans are over-spec'd but anyway I can't be too critical of a cheap reliable power supply with excessive ventilation.
FWIW the HEC came in a 6A19 from servercase.com, this is a good case for modding for good airflow and quiet accoustics. And my favorite quiet cases (the 6919) are cheap there too. Plain but solid, good features and superior airflow.
When I use a coffee shop, the default route on my firewalled Freebsd laptop is an openvpn tunnel to my home router; tough to sniff or hack. But if I didn't do this, I would place Google last on my list of who I would trust as my router. Chances are very high the patrons and barristas don't have the competence or the interest in sniffing (it's stupid but has never been a big problem). If the shop put up an open AP on a DSL link they likely are in the same boat. If they use a commercial captive portal, the provider has a vested interested in not screwing up the relationship with a paying customer. But Google has a vested interest in evil here and they have stated as much in their policy.
I can only conclude that Google is really the WORST... VPN... ENDPOINT... EVAR! But what did you expect for free?
I expect that google does cache and mine everything that passes through them as well as everything they have ever found.
If you wanted to know what someone has been up to for the last decade, Google will likely have far better information than their ISP or the FBI.
Anyway, personally I find that BlueSpew has the right ring for this crap.
And worse, for every kind of incident that isn't reported, there are thousands just like it.
I cut my teeth programming the Atari ST way back when. Data Becker had a rocking book ("Atari ST Internals") that included a complete *commented* bios listing that they had disassembled and reverse engineered.
If you wanted to see how see how vertical blank interrupts or midi I/O was handled, it was all there. I think everybody that wrote midi sequencers back then reimplimented the interrupt handlers from the DB source code, and it was tight! To this day you can't get tighter midi on your 3 GHz machines than you could get on an 8MHz Atari.
My tools back then were really top notch; I loved Laser C (a big improvement on Megamax which was also excellent) and the DevPak assembler. I still have my original source floppies and documentation. Though my machines went to the thrift store years ago, I would still consider running these tools under emulation to teach a class on how computers really work.
The hardware was so accessible then. I still think the motorola 68k was the clearest machine to program in assembly we've ever had.
Though the parent is intended as a joke, based on my own thoughts from reading the works of Einstein I believe it to be more insightful than the article and most of the posts here that fail to understand wny time travel is romantic BS.
There really is no such thing as time, only the relative differences in the states of energy of systems, where energy is observed as velocity (integration of acceleration as time goes to zero) or relative positions in an arbitrary coordinate system.
To go back in time is to rewind the state of the universe, effectively to reverse the vector of all energy (and fields). You can create a perspective that can change the relative state of a simple system, but this function will not magically rewind the states of arbitrary systems, let alone the gestalt.
Any equation that involves time is actually measuring some other assumed relationship. Time was linear in Newtonian physics because it only solves for simple systems with a constant field, but there's a reason that better theories involve an integration with time going to zero.
I cringe when I read Hawking and he jabbers endlessly about not understanding why time seems to have a vector and then ponders the magic of some wonder field that reverses it all. It's clear that he read Einstein and he could do the math but he missed the big point entirely; time is a big zero, always now, but energy has a vector and the universe is a directed rotor that cannot be simply deconvolved.
ps Thanks for the universe, God, it's mostly working pretty well thought I have a few suggestions.
http://www.google.com/search?hl=en&q=pultonmiu+dri ty+bmbs&btnG=Google+Search
Google, the patriot's friend!
I'm running FreeBSD 4.11 on my servers and laptop. To this point I've avoided 5.x due to rumours of a performance downgrade for my single processor systems. (Well that & pragmatism, I am slow to change good systems.)
I have no great dissatisfactions with my 4.11 systems, but could make use of cardbus support on the laptop, and if 5.x's threading could improve the performance of samba, mySQL, postgreSQL, maybe java and apache (my primary server loads) that would be a win.
Is 5.4 ready for primetime for satisfied 4.x users? What are the real world performance implications at this time in terms of memory usage, I/O throughput, performance on my favorite server apps and as a GUI workstation? Is it as solid as 4.x has always been?
I used to work with a fellow whose hobby was to harass and attempt to discredit the Scientologists. They claimed he had posted their intellectual property on the internet, and when he was deposed he stated that their material was indeed available at 127.0.0.1 .
They quoted this verbatim in their claim against him, so it made it past whoever was responsible for fact checking the claim, and he boasted around the water cooler how he had bested the scientology legal staff. He explained 'localhost' on the stand and the claim was thrown out.
He was rather an odd fellow. I had quite forgotten him until some 10 years later I saw his name appear on slashdot in an appeal for money for his legal defense fund.
But Prime95 confirms that no bit anywhere in nearly the complete memory space ever spuriously changes. I have seen plenty of memory that passes metest86 that fails prime95.
Based on my experience, Corsair will replace memory that fails prime95. Mushkin will NOT (despite a "lifetime" warranty); they basically told me that memory can't be expected to be 100% perfect all the time and that prime95 was too strenuous; if it passes memtest86 there will be no replacement. My other modules (from Geil, Samsung, and a few old no-name sticks) have always been perfect. IMO it's unconscionable to sell untested ram given how hard it is to return.
They do provide me with a variety of free, useful services in order to sell my eyeballs and Google's knowlege of my interests and habits to their customers, which are companies interested in potential consumers.
The GPL has led to a terrific and amazing revolution; our information infrastructure, which is one of this generation's most crucial assets and contributions, is effectively given to the public trust, rather than locked away as a tool of the powerful to exploit the rest. Furthermore, mankind can use the tools of the past to move forward, rather than reinvent the same tools over and again.
The usage of GPL software to provide services strikes me as completely within the intent of the GPL, and unlikely to spur the kinds of abuses that the GPL was intended to discourage. Would anyone benefit if everyone that provided a service derived from GPL'd code were required to pay? Hmmm, would anyone EVER then run a GPL'd webserver, or any other service run on a GPL'd OS? What users of GPL software would be liable, any whose services indirectly contribute to anything providing revenue? Such a GPL would be the death of free software.
But you can't wish the C out of obj-C, that's one of the main attractions of the language. If you're going to write a codec (say) or an image filter, you use the object features to facilitate reuse of tight data munging code. Sometimes there are good reasons to prefer cstrings over string objects, and processor-native number types over number objects.
Objective-C is mostly about writing reusable C code. If you're big on OO dogma and objects everywhere it's a poor choice. But there's a good place for well structured, late-bound C too. Obj-c enables OO programming, but it deliberately does not impose it.
If you want pure and focused OO, you have your pick of a dozen flavors, but none will be the efficient system programming language the objc is.
But for believers (and those in the know :-) obj-c is a terrific lightweight and very capable tool. The programs can be, in practice, as fast and small as C applications, yet you can still design abstract api's that easily support inheritance and polymorphism. It's not an academic argument; Obj-C has been key to the design of one of the best application development environments for 15 _years_. And I would still consider it one of the strongest choices for such development, starting clean-sheet today.
Obj-C is in the OSX kernel, and it was in the NeXT kernel a decade ago. It's in the Mac display system. You would be crazy to use Java in these circumstances.
I would not argue that Obj-C is one of the better OO languages for applications where true OO is necessary or useful. The biggest issue is that memory management is in your face, for better and for worse. There are plenty of situations where such control is useful or critical, and much as I love the convenience & safety of a garbage collector, a large gc app in my experience is hell on virtual memory, whereas it is possible to have a very light footprint with obj-c.
Your argument on the expense of the message dispatcher doesn't hold water. A message dispatch is 3x as expensive as a C function call; compared to the cost of most algorithms, it's totally in the noise. Even if it were substantial in a heavy loop, you could indirect through the message's function for a cost no higher than normal C. This is trivial though rarely useful. You're arguing for a Java implementation, where the cost will be higher even before the garbage collector takes its toll on the CPU and VM.
Objective-C is a terrific choice as a system programming language, and I would consider it a strong choice for most problems sets you would otherwise choose C or C++ for, and it wallops these for extensibility and reuse. It's a poor choice for problems where you would have a good reason to choose Java or Smalltalk.