Or the protocol not paying attention to (or having) packet sequence numbers. As a result, dropped frames due to collisions, timeouts, and other issues wouldn't get re-sent. And you can do that with IP, too; just not TCP.
Compare NFS over UDP performance with 32K blocks and 5% packet loss. Carrying floppy disks around the building would be faster....
The phrase you're looking for is "USB KVMP"; keyboard-video-mouse-peripheral switch. The ones I've got from IOGear (see also Aten) have separate settings for KVM and P; so I can leave the printer (say) connected to the Mac while I do something on Linux or Windows.
This sounds like just the P part of one of those, with some magic goo to control it from the computer side, instead of hot-keying on an attached keyboard or whacking on a button.
And there's also USB-over-Ethernet sharing, which looked cute but not worth actually paying for.
Maybe they need to allow us to use maildir format mailboxes (each message goes in a separate file). The downside is disk space and whether the file system can efficiently handle thousands of small files in a directory.
Ignoring the number of inodes consumed by maildirs, and disk inefficiencies due to over-large allocation fragments on large disks (both of which are solvable by suitable filesystem tuning....)
The problem of efficiently handling thousands of files in a directory is already solved.
You simply don't put thousands of files in a directory... use subdirectories.
For maildir, the simplest thing to do would be automatically create a new subdirectory when the current one hits some-arbitrary-number. Actually, not so simple; you've got to count--but you could easily set it up so counting a bit wrong is harmless. Then the application scoops up the subdirectories instead of the main directory.
Index caching could then be done on individual subdirectories, speeding that whole part up.
Another approach would be to use part of the filename (my software always uses the time_t value plus other junk, I think the time_t part is standard); so take the time_t part of the filename and do an integer divide by 2592000 (seconds in 30 days), or some other suitable calculation, and use the result for the folder name.
The only race condition to worry about there would be directory creation, and that's easy enough; if mkdir(2) gives EEXIST, you consider that success.
Yup, those controls pre-date the "modern" (1960s? 1970 at the latest) electric stove control that uses a thermostat and heater in the control to do a very slow pulse-width modulation to control the amount of power used in the burner.
I retro-fitted one that had rotary switch controls (like another poster mentioned) using the controls from the old range in my parent's house when they upgraded.
Those switch controls also usually involve two elements per burner (perhaps bonded into one spiral, perhaps visibly separate). By combining the two elements in different combinations of 120V and 240V, you can get a reasonable range of heat. (For example. both in series across 120V for simmer, both in parallel across 240V for boil.)
Of course, what's important is a large diversity of password-generating methods on a system. Leet-speak phrases, initials of favorite songs, word+number+word, pure random gibberish, plain series of letters, and series of numbers.
As soon as you restrict the set of allowable passwords, you also restrict the search space to crack a password. So, for example, I know that I shouldn't try any passwords shorter than 8 characters comprised only of letters and numerals to crack an Intel customer account password. You need at least 8 characters, and there must be both a numeral and punctuation character there.
There's a balance to be walked between "predictable" passwords and arbitrary silly restrictions like "must have a numeral". As long as some users might have a numeral, an attacker has to try all those combinations... and the ones without, too.
Good cracker software already knows how to l33t-sp33k usernames, gcos fields, and dictionary words.
At least, in a recent, inspired-by-this-article check, "gcos-value as password" was cracked first, and "l33t-sp33k username as password" was second, then it got on with the dictionary attacks and permutations.
By the time you get to linear cracking, you might as well give up, unless you've got more CPU time on your hands than you know what to do with.
And I have a commercially-pressed DVD that failed 2 years after I bought it. So did a friend's. And so did the rental store down the street....
Now, that's only 1 out of 1000....
DVD-R media, though; I've got a bunch of early 2X and 4X discs from back when that was pretty new that were unuseable after 6 months. No drive I could get near (and that's at least a dozen) would register that media as being loaded.
I'm not sure about the current state-of-the-art for no-name media; I've been using Verbatim a lot recently. (Mainly because of some crazy-good sales Staples had a year ago; I've got _lots_ of Verbatim blanks now.)
For el-cheapo CDs, one thing to watch is lack of protection on the recording layer: I've got some from Rat Shack that you can peel the recording surface off with a Post-It note.
The other issue with a bazillion registers is the cost of context-switching on that high-performance machine. You've now got to stuff all those registers on the stack and load up the other processes' set. So you lengthen the timeslice to minimize the cost, at the expense of "responsiveness".
Or you can play tricks with register files, so you don't have to actually save and load... but then you run out of space in the file with too many processes, so you have to do it anyway.
Or there's this really weird "windowed" thing on SPARC chips, which is really nifty... until you try and use it on a multi-tasking operating system, and you've got to swap out all the entries in the entire window-thing, so your input registers don't become some other processes output registers.
Same with cache: lots of cache means more entries to manage. You have to compare every physical address to the cache tags to see if you've got it, so the more tags you have to compare, the slower the operation gets (or the more power you need to get larger fan-out). Tricks like set-associative caches help simplify these tests and keep them manageable.
It doesn't matter how much money you've got, there's always a tradeoff to be made in the design.
And, these days, doing bunches of stuff at the same time is where the best bang for your watt happens to be.
You got it on the cooling. Nothing makes you have to call the service guy out to replace a HDD like heat.
Me and the IBM service guy spent a happy afternoon removing the middle bit from a bunch of foam sound baffles that had been left in by the factory. (7013-5xx series RS/6000s.) There was supposed to be a rectangle of foam around the vents on the bottom of the machines, but the cut-out bit in the middle was left in. Lots of dying disks until someone noticed. (This was way before SMART monitoring, which would have told us the drives were way over-temperature.)
Another team in the same machine room had the tech out every 3-4 months to replace disks in one of their boxes. They put all their machines side-to-side touching, so the vents didn't have any airflow....
For some reason, I like to keep the vent grilles clear. I had to pull the plastic bezel off my home server because, being a cheap nasty case, the bezel wasn't designed to match the fan mounts behind it and my drives were running much too hot. One of these days I'll get a better case, but hey, it's just ugly now, it works great.
In no conforming C compiler should any of that math take place in floating point. You have (int)/(int), so you get an integer divide.
(float)=(int)/(int) would still do integer divide and convert the result to float. Probably with some griping from an ANSI compiler.
It's also entirely up to the compiler and the underlying processor; the operating system is not called for simple math operations. (Except those on processors that don't have integer divide and multiply, like say the original SPARC.)
If you're playing fast-and-loose with the types, you can get some interesting results. Like, if you pass a float to printf() with an integer format. They're not passed the same way in the AIX ABI, so printf() will pull junk out of the register where the appropriate integer would be.
Also, under ANSI aliasing rules, an integer can never be used to store a float and vice-versa, so the compiler won't schedule instructions to ensure that a write to a float is complete before a read from an integer (even at the same memory address) is started. Lots of K&R C code breaks under this rule.
Not to say that a certain old C compilers from IBM wasn't garbage; there's a reason why that C compiler got dropped in 1993.
OTOH, continuing to use data after an illegal operation is pretty stupid, too.
According to ISO C 99, the compiler is well within its rights to get you anything if you divide (or modulo) by zero: "the behavior is undefined." So 15 is conforming. SIGILL is conforming. Deleting all your files and sending a rude message to your manager is conforming.
(And what strange world has 'void main();'? That's not hosted C, where main is always 'int main(int,char**);')
Well, Apple customers have had to put up with that for years. Stand in front of a Mac on display and ask a salesclerk if they sell Macs.... "Uh, what's a Mac?"
Which is one reason why Apple started opening their own stores. The "store in a store" thing didn't work, you still got clueless clerks from the rest of the shop.
So, turnabout being fair play and all, Microsoft needs to open up the Zune store.
Geez. Thanks for reminding me, I had one of those early "never seen that company before or since" Asian CD-based MP3 players.
Total piece of crap. It was OK in the car, where I could set it and let it run off a 12V adapter. But on batteries, it was 70 minutes runtime, tops--enough for a regular audio CD, but not 650 MB of MP3 goodness. (Hey, those 700 MB blanks were _expensive_... remember that?)
UI was terrible. Don't even bother trying to navigate, all you had was a "folder number" and "track number" display. Keys were shoddy, hard to press, would either not respond or multi-click (bad debounce coupled with bad hardware, not a good mix).
Sound quality was decent; at least, with real headphones. The ones it came with were never unwrapped.
When the Rio Volt came out a year or two later, that was an AMAZING leap forward--ID3 tag support, fairly usable navigation, controls that worked, decent battery run-time. I could (and did) spend many an afternoon by the lake with the Rio Volt, a good novel, and a comfy beach chair. Still fairly much a geek item, though. You had to make your CDs in a suitable way, and for us Mac geeks, you had to stick ".mp3" on the end of the file for some inexplicable reason.
So when the 5GB iPod came out, it wasn't compelling. Heck, I had 10GB of songs in a Case Logic folder that held the CDs and the Rio, all in the size of a medium city's phone book folded in half. Uhhh.
Not to say I didn't _want_ an iPod, but that sort of money....
Then I won a 20GB iPod on the Pepsi Canada promotion. (With a No Purchase Necessary PIN code, too.)
What did someone say about accessories? "free" iPod. Not free accessories... smart promotion.
Re:When you write for specific hardware....
on
Leopard Vs. Vista
·
· Score: 1
Do I blame SW for the crappy NIC? I shouldn't, although I cursed them repeatedly while trying to find the problem
I know that feeling. Remember the "fake" electrolytic capacitor problem a couple or three years ago? The ones that worked fine for a bit, but would then burst because the electrolyte formula was a crap knock-off?
So I was swearing a blue streak at Red Hat Linux 6.2 for crashing all the time on my work PC. Tried upgrading to 7.3, which I was using at home and was rock solid. Finally got fed up and called the IT group, since I couldn't figure it out; it looked a _bit_ like maybe the power supply was weak and going south.
IT guy comes up, pops the cover and looks at all the capacitor goo sprayed over the inside of the case, and says, "I'll order you a new motherboard from IBM."
So. Strange crashes aren't always software's fault. (I'd had a similar problem before, guru meditations on the Amiga that were solved by adding a UPS... turns out the power in my apartment was more flaky than I thought.)
Mind you, int the space of 4 days, they got 5 defective boards from IBM, and I suggested (quite loudly) to the IT guy that IBM should send a board from a different model of machine. That one had no sound below 500 Hz on the headphone out, but I fixed that with a spare USB audio adapter....
(So I can also understand why IBM sold PC Company to the Chinese; PC Company was always a little different, QC wise, from the rest of IBM.)
There were several ROM messages; one of my machines (I think it was the B2000, the two-layer system board with FCC class B certification) had "The Amiga... born a champion" and "Still a champion!" Think that was KickStart 1.3, which was generic for the A2000, B2000 and A500.
I'm no fan of EULAs or any software licensing (not even the GPL)
Not to defend EULAs, but you know you don't have to accept the GPL to use GPL'ed software? It even says right in the license that you don't have to accept it. See clause 5. (So now we have the irony of brain-dead installers saying "You must accept the license to install the software" and the license says "You do not have to accept this license".)
If you don't like the GPL, you can just let regular copyright law define your rights. Which means no derivative works without permission, no distribution, and so on.
Most other EULAs add restrictions. And many of them say you should return the software to the place of purchase for a refund if you do not agree--at least, the ones I've read recently. Hmmm. If Windows has a clause like that, there's a good reason for taking it back and getting a refund....
The problem being, of course, once you've read the EULA and found out you should return it and ask for a refund, most retailers won't give you one because the box is open.
Those little plastic things (only in jewel cases) used to break like nobody's business when CDs were fairly new. Fortunately, it'd usually be just one or two, and there wasn't anywhere for the disc to go.
DVD cases are too big for the DVD, so when it comes off the tabs (even if they aren't broken), it will almost certainly slide back and forth and gets scratched. When that happens, I make a copy and see what the read performance and error rate is like; it's amazing how mucked up a DVD can appear be and still read at 4x or 8x....
(That's true for rental-by-mail discs, too, I've only had 1 or 2 out of 100 or 200 that couldn't play for being scratched.)
There are two distinct ways in which the contents of the process image file may cause the execution to fail, distinguished by the setting of errno to either [ENOEXEC] or [EINVAL] (see the ERRORS section). In the cases where the other members of the exec family of functions would fail and set errno to [ENOEXEC], the execlp() and execvp() functions shall execute a command interpreter and the environment of the executed command shall be as if the process invoked the sh utility using execl() as follows:
execl(, arg0, file, arg1,..., (char *)0);
where is an unspecified pathname for the sh utility, file is the process image file, and for execvp(), where arg0, arg1, and so on correspond to the values passed to execvp() in argv[0], argv[1], and so on.
Thing is, unless some app is hideously mis-designed, random saved files won't have execute permission on them, and won't execute. This is where tar files and ZIP files from Windows are so much fun, as they tend to have all-bits-on for some strange reason. (I know the Windows NT series has execute permission for files. I don't know if Windows users and developers are aware of that....)
(In other words, all [conforming] UNIXes do this.)
Uhhhh, you might wish to check the statute before saying you could avoid it by manufacturing media in Canada.
82. (1) Every person who, for the purpose of trade, manufactures a blank audio recording medium in Canada or imports a blank audio recording medium into Canada
Anyway, it's called a levy because it isn't paid to the government for public purposes (wandering by wiktionary for that); it's paid to the Copyright Board for distribution to recording artists.
I tried a Harmony. It had no way of controlling by DVD carousel player without going to the per-device soft menu and through 4 screens to get the disc change buttons. Despite the fact that it claimed to know about my model of player; there were no disc change commands under "Play a DVD".
So I'm still using the Sony RM-AV3000; and I wish I could get a Marantz RC2000 Mk II... my RC2000 Mk I got dropped on the floor just a _few_ too many times....
(It might become moot; I'm thinking of getting rid of the changer and just having a single-disc progressive scan player....)
The few times an ad has come up for something I did want, it turned out that it could only be sold to persons in the United States. No shipping to Canada.
So I just block the lot, now, what's the point in wasting my time?
It works just fine with WebDAV. In fact, it works better with WebDAV than the Web Folders thing does. Add "SVNAutoversioning on" to your Subversion repository config and have fun, just for one example.
Nobody else had the gall to sell a player with no display.
You entirely sure about that? Maybe nobody with a big name, but I've used CD-based MP3 players that didn't even have track number displays, let alone something that could show folder and file names--the Rio Volt was an amazing breakthrough....
And I've got a pocket FM radio with no display. Digital synthesized tuning, and no frequency display. There's no tuning knob, all it will do is seek UP the FM band until it hits another station with sufficient signal strength. When it hits the top of the band, it wraps to the bottom and you start over.
What's the point of an antibacterial if the problem is viruses?
People are stupid and will fall for it.
I actually had a doctor tell me, "It's a viral infection." Then he said, "I'm not going to prescribe an antibiotic for it."
I said, "Of course not, it's viral."
Apparently a lot of patients want antibiotics for everything. I was just wondering if I'd make everyone at work sick if I went in and should, instead, stay home some more. Sadly, he said it was OK to go in to the office. But I got some great cough syrup with Codeine....
Slashdot Mirror
Or the protocol not paying attention to (or having) packet sequence numbers. As a result, dropped frames due to collisions, timeouts, and other issues wouldn't get re-sent. And you can do that with IP, too; just not TCP.
Compare NFS over UDP performance with 32K blocks and 5% packet loss. Carrying floppy disks around the building would be faster....
Hey, if it wasn't for him and his like trying to get GTA3 banned, I would have never bought it. And wouldn't have played it and enjoyed it...
So, Jack, keep talking about what games are going to destroy the world, I need help with my shopping list.
The phrase you're looking for is "USB KVMP"; keyboard-video-mouse-peripheral switch. The ones I've got from IOGear (see also Aten) have separate settings for KVM and P; so I can leave the printer (say) connected to the Mac while I do something on Linux or Windows.
This sounds like just the P part of one of those, with some magic goo to control it from the computer side, instead of hot-keying on an attached keyboard or whacking on a button.
And there's also USB-over-Ethernet sharing, which looked cute but not worth actually paying for.
Ignoring the number of inodes consumed by maildirs, and disk inefficiencies due to over-large allocation fragments on large disks (both of which are solvable by suitable filesystem tuning....)
The problem of efficiently handling thousands of files in a directory is already solved.
You simply don't put thousands of files in a directory... use subdirectories.
For maildir, the simplest thing to do would be automatically create a new subdirectory when the current one hits some-arbitrary-number. Actually, not so simple; you've got to count--but you could easily set it up so counting a bit wrong is harmless. Then the application scoops up the subdirectories instead of the main directory.
Index caching could then be done on individual subdirectories, speeding that whole part up.
Another approach would be to use part of the filename (my software always uses the time_t value plus other junk, I think the time_t part is standard); so take the time_t part of the filename and do an integer divide by 2592000 (seconds in 30 days), or some other suitable calculation, and use the result for the folder name.
The only race condition to worry about there would be directory creation, and that's easy enough; if mkdir(2) gives EEXIST, you consider that success.
Yup, those controls pre-date the "modern" (1960s? 1970 at the latest) electric stove control that uses a thermostat and heater in the control to do a very slow pulse-width modulation to control the amount of power used in the burner.
I retro-fitted one that had rotary switch controls (like another poster mentioned) using the controls from the old range in my parent's house when they upgraded.
Those switch controls also usually involve two elements per burner (perhaps bonded into one spiral, perhaps visibly separate). By combining the two elements in different combinations of 120V and 240V, you can get a reasonable range of heat. (For example. both in series across 120V for simmer, both in parallel across 240V for boil.)
Horrible to work with.
Of course, what's important is a large diversity of password-generating methods on a system. Leet-speak phrases, initials of favorite songs, word+number+word, pure random gibberish, plain series of letters, and series of numbers.
As soon as you restrict the set of allowable passwords, you also restrict the search space to crack a password. So, for example, I know that I shouldn't try any passwords shorter than 8 characters comprised only of letters and numerals to crack an Intel customer account password. You need at least 8 characters, and there must be both a numeral and punctuation character there.
There's a balance to be walked between "predictable" passwords and arbitrary silly restrictions like "must have a numeral". As long as some users might have a numeral, an attacker has to try all those combinations... and the ones without, too.
Good cracker software already knows how to l33t-sp33k usernames, gcos fields, and dictionary words.
At least, in a recent, inspired-by-this-article check, "gcos-value as password" was cracked first, and "l33t-sp33k username as password" was second, then it got on with the dictionary attacks and permutations.
By the time you get to linear cracking, you might as well give up, unless you've got more CPU time on your hands than you know what to do with.
And I have a commercially-pressed DVD that failed 2 years after I bought it. So did a friend's. And so did the rental store down the street....
Now, that's only 1 out of 1000....
DVD-R media, though; I've got a bunch of early 2X and 4X discs from back when that was pretty new that were unuseable after 6 months. No drive I could get near (and that's at least a dozen) would register that media as being loaded.
I'm not sure about the current state-of-the-art for no-name media; I've been using Verbatim a lot recently. (Mainly because of some crazy-good sales Staples had a year ago; I've got _lots_ of Verbatim blanks now.)
For el-cheapo CDs, one thing to watch is lack of protection on the recording layer: I've got some from Rat Shack that you can peel the recording surface off with a Post-It note.
The other issue with a bazillion registers is the cost of context-switching on that high-performance machine. You've now got to stuff all those registers on the stack and load up the other processes' set. So you lengthen the timeslice to minimize the cost, at the expense of "responsiveness".
Or you can play tricks with register files, so you don't have to actually save and load... but then you run out of space in the file with too many processes, so you have to do it anyway.
Or there's this really weird "windowed" thing on SPARC chips, which is really nifty... until you try and use it on a multi-tasking operating system, and you've got to swap out all the entries in the entire window-thing, so your input registers don't become some other processes output registers.
Same with cache: lots of cache means more entries to manage. You have to compare every physical address to the cache tags to see if you've got it, so the more tags you have to compare, the slower the operation gets (or the more power you need to get larger fan-out). Tricks like set-associative caches help simplify these tests and keep them manageable.
It doesn't matter how much money you've got, there's always a tradeoff to be made in the design.
And, these days, doing bunches of stuff at the same time is where the best bang for your watt happens to be.
You got it on the cooling. Nothing makes you have to call the service guy out to replace a HDD like heat.
Me and the IBM service guy spent a happy afternoon removing the middle bit from a bunch of foam sound baffles that had been left in by the factory. (7013-5xx series RS/6000s.) There was supposed to be a rectangle of foam around the vents on the bottom of the machines, but the cut-out bit in the middle was left in. Lots of dying disks until someone noticed. (This was way before SMART monitoring, which would have told us the drives were way over-temperature.)
Another team in the same machine room had the tech out every 3-4 months to replace disks in one of their boxes. They put all their machines side-to-side touching, so the vents didn't have any airflow....
For some reason, I like to keep the vent grilles clear. I had to pull the plastic bezel off my home server because, being a cheap nasty case, the bezel wasn't designed to match the fan mounts behind it and my drives were running much too hot. One of these days I'll get a better case, but hey, it's just ugly now, it works great.
In no conforming C compiler should any of that math take place in floating point. You have (int)/(int), so you get an integer divide.
(float)=(int)/(int) would still do integer divide and convert the result to float. Probably with some griping from an ANSI compiler.
It's also entirely up to the compiler and the underlying processor; the operating system is not called for simple math operations. (Except those on processors that don't have integer divide and multiply, like say the original SPARC.)
If you're playing fast-and-loose with the types, you can get some interesting results. Like, if you pass a float to printf() with an integer format. They're not passed the same way in the AIX ABI, so printf() will pull junk out of the register where the appropriate integer would be.
Also, under ANSI aliasing rules, an integer can never be used to store a float and vice-versa, so the compiler won't schedule instructions to ensure that a write to a float is complete before a read from an integer (even at the same memory address) is started. Lots of K&R C code breaks under this rule.
Not to say that a certain old C compilers from IBM wasn't garbage; there's a reason why that C compiler got dropped in 1993.
OTOH, continuing to use data after an illegal operation is pretty stupid, too.
According to ISO C 99, the compiler is well within its rights to get you anything if you divide (or modulo) by zero: "the behavior is undefined." So 15 is conforming. SIGILL is conforming. Deleting all your files and sending a rude message to your manager is conforming.
(And what strange world has 'void main();'? That's not hosted C, where main is always 'int main(int,char**);')
Ahh, excellent! Then they can count the additional sale from the "Open Box" table.
(I just _hope_ I'm joking....)
Well, Apple customers have had to put up with that for years. Stand in front of a Mac on display and ask a salesclerk if they sell Macs.... "Uh, what's a Mac?"
Which is one reason why Apple started opening their own stores. The "store in a store" thing didn't work, you still got clueless clerks from the rest of the shop.
So, turnabout being fair play and all, Microsoft needs to open up the Zune store.
Geez. Thanks for reminding me, I had one of those early "never seen that company before or since" Asian CD-based MP3 players.
Total piece of crap. It was OK in the car, where I could set it and let it run off a 12V adapter. But on batteries, it was 70 minutes runtime, tops--enough for a regular audio CD, but not 650 MB of MP3 goodness. (Hey, those 700 MB blanks were _expensive_... remember that?)
UI was terrible. Don't even bother trying to navigate, all you had was a "folder number" and "track number" display. Keys were shoddy, hard to press, would either not respond or multi-click (bad debounce coupled with bad hardware, not a good mix).
Sound quality was decent; at least, with real headphones. The ones it came with were never unwrapped.
When the Rio Volt came out a year or two later, that was an AMAZING leap forward--ID3 tag support, fairly usable navigation, controls that worked, decent battery run-time. I could (and did) spend many an afternoon by the lake with the Rio Volt, a good novel, and a comfy beach chair. Still fairly much a geek item, though. You had to make your CDs in a suitable way, and for us Mac geeks, you had to stick ".mp3" on the end of the file for some inexplicable reason.
So when the 5GB iPod came out, it wasn't compelling. Heck, I had 10GB of songs in a Case Logic folder that held the CDs and the Rio, all in the size of a medium city's phone book folded in half. Uhhh.
Not to say I didn't _want_ an iPod, but that sort of money....
Then I won a 20GB iPod on the Pepsi Canada promotion. (With a No Purchase Necessary PIN code, too.)
What did someone say about accessories? "free" iPod. Not free accessories... smart promotion.
I know that feeling. Remember the "fake" electrolytic capacitor problem a couple or three years ago? The ones that worked fine for a bit, but would then burst because the electrolyte formula was a crap knock-off?
So I was swearing a blue streak at Red Hat Linux 6.2 for crashing all the time on my work PC. Tried upgrading to 7.3, which I was using at home and was rock solid. Finally got fed up and called the IT group, since I couldn't figure it out; it looked a _bit_ like maybe the power supply was weak and going south.
IT guy comes up, pops the cover and looks at all the capacitor goo sprayed over the inside of the case, and says, "I'll order you a new motherboard from IBM."
So. Strange crashes aren't always software's fault. (I'd had a similar problem before, guru meditations on the Amiga that were solved by adding a UPS... turns out the power in my apartment was more flaky than I thought.)
Mind you, int the space of 4 days, they got 5 defective boards from IBM, and I suggested (quite loudly) to the IT guy that IBM should send a board from a different model of machine. That one had no sound below 500 Hz on the headphone out, but I fixed that with a spare USB audio adapter....
(So I can also understand why IBM sold PC Company to the Chinese; PC Company was always a little different, QC wise, from the rest of IBM.)
There were several ROM messages; one of my machines (I think it was the B2000, the two-layer system board with FCC class B certification) had "The Amiga... born a champion" and "Still a champion!" Think that was KickStart 1.3, which was generic for the A2000, B2000 and A500.
Not to defend EULAs, but you know you don't have to accept the GPL to use GPL'ed software? It even says right in the license that you don't have to accept it. See clause 5. (So now we have the irony of brain-dead installers saying "You must accept the license to install the software" and the license says "You do not have to accept this license".)
If you don't like the GPL, you can just let regular copyright law define your rights. Which means no derivative works without permission, no distribution, and so on.
Most other EULAs add restrictions. And many of them say you should return the software to the place of purchase for a refund if you do not agree--at least, the ones I've read recently. Hmmm. If Windows has a clause like that, there's a good reason for taking it back and getting a refund....
The problem being, of course, once you've read the EULA and found out you should return it and ask for a refund, most retailers won't give you one because the box is open.
Those little plastic things (only in jewel cases) used to break like nobody's business when CDs were fairly new. Fortunately, it'd usually be just one or two, and there wasn't anywhere for the disc to go.
DVD cases are too big for the DVD, so when it comes off the tabs (even if they aren't broken), it will almost certainly slide back and forth and gets scratched. When that happens, I make a copy and see what the read performance and error rate is like; it's amazing how mucked up a DVD can appear be and still read at 4x or 8x....
(That's true for rental-by-mail discs, too, I've only had 1 or 2 out of 100 or 200 that couldn't play for being scratched.)
The short answer is, "Historical reasons";
And the longer answer is, POSIX says so:
Thing is, unless some app is hideously mis-designed, random saved files won't have execute permission on them, and won't execute. This is where tar files and ZIP files from Windows are so much fun, as they tend to have all-bits-on for some strange reason. (I know the Windows NT series has execute permission for files. I don't know if Windows users and developers are aware of that....)
(In other words, all [conforming] UNIXes do this.)
Uhhhh, you might wish to check the statute before saying you could avoid it by manufacturing media in Canada.
Anyway, it's called a levy because it isn't paid to the government for public purposes (wandering by wiktionary for that); it's paid to the Copyright Board for distribution to recording artists.
I tried a Harmony. It had no way of controlling by DVD carousel player without going to the per-device soft menu and through 4 screens to get the disc change buttons. Despite the fact that it claimed to know about my model of player; there were no disc change commands under "Play a DVD".
So I'm still using the Sony RM-AV3000; and I wish I could get a Marantz RC2000 Mk II... my RC2000 Mk I got dropped on the floor just a _few_ too many times....
(It might become moot; I'm thinking of getting rid of the changer and just having a single-disc progressive scan player....)
The few times an ad has come up for something I did want, it turned out that it could only be sold to persons in the United States. No shipping to Canada.
So I just block the lot, now, what's the point in wasting my time?
It works just fine with WebDAV. In fact, it works better with WebDAV than the Web Folders thing does. Add "SVNAutoversioning on" to your Subversion repository config and have fun, just for one example.
You entirely sure about that? Maybe nobody with a big name, but I've used CD-based MP3 players that didn't even have track number displays, let alone something that could show folder and file names--the Rio Volt was an amazing breakthrough....
And I've got a pocket FM radio with no display. Digital synthesized tuning, and no frequency display. There's no tuning knob, all it will do is seek UP the FM band until it hits another station with sufficient signal strength. When it hits the top of the band, it wraps to the bottom and you start over.
People are stupid and will fall for it.
I actually had a doctor tell me, "It's a viral infection." Then he said, "I'm not going to prescribe an antibiotic for it."
I said, "Of course not, it's viral."
Apparently a lot of patients want antibiotics for everything. I was just wondering if I'd make everyone at work sick if I went in and should, instead, stay home some more. Sadly, he said it was OK to go in to the office. But I got some great cough syrup with Codeine....