> Nevermind that hard disk manufacturers for a long time defined a GB as 1000 * 1000 * 1024 bytes.
The capaxity of the first hard disk was specified in powers of 10, not 2. Every subsequent disk from that manufacturer (IBM) has also been measured in powers of 10, not 2. Just when was IBM supposed to redefine a kilobyte, megabyte and gigabyte of disk storage to accomodate computers which could not support base 10?
Lots of luck. Looks like it didn't make the ballot in 2K2.:-). Remember that Lani Guinier
was denied a federal appointment for being a bit too innovative wrt electoral fairness.
Not that the Democrats haven't pulled equally partisan shenanigans.
Shortly after the first moon landing (1969), then-Vice President Spiro Agnew announced that the United States would put a man on Mars by the year 2000. Lets take care of that first, before we attend to the present Command-in-Thief's fancy.
C permits an implementation to trap on overflow. Whether a particular implementation does so is a quality of implementation issue. Of course, some architectures make it trapping easy, others do not. MIPS is quite easy: just emit the right form of the instruction. So is S/370: enable fixed-point overflow. x86 is harder: every overflow check costs you a 1-byte instruction.
I think PowerPC costs an instruction also, and that is a 4-byte instruction. But you don't need to check as often as with x86 because PowerPC keeps a "summary overflow" bit.
I don't know of gcc supports any of these checking mechanisms.
> the first hard drive I ever owned was 10MB - 10240KB
Hard drives go back farther than that. IBM has always quoted hard drive size in base 10. Mainframe could format the tracks to any blocksize they chose (up to actual track size - does 3625 byte ring a bell to anyone?). Most chose a multiple of 80, which was the number of positions on a punched card.
Not having looked at the code, it seems to me that make would only handle half the problem: booting. Shutdown is the other half; the dependencies would be reversed. For example:
For boot you would tell make this:
sshd: network rpcd: network
But for shutdown you need to tell it this:
network: sshd rpcd
Ideally one set of input data should take care of both cases.
Congratulation to all who made this possible. I don't imagine it was easy.
Re:The problem is: that's not the problem
on
Does C# Measure Up?
·
· Score: 1
> SLR is faster because there's no condition code to set.
No, that's not it. SLR does set the condition code. What you are trying to remember is fixed-point overflow, which is possible with SR but not with SLR.
Whether SR really takes longer is arguable nowadays; IBM hasn't published instruction timings for S/370 (except 43xx) for many years. I think the non-DAT 370's may have been the last.
Re:The problem is: that's not the problem
on
Does C# Measure Up?
·
· Score: 1
> SLR R1,R1 is 60ns faster than SL R1,R1
It's the storage reference - unaligned, no less - which kills the SL R1,R1 performance.:-)
Due to inventory taxes the publishing industry has trouble keeping books in print. E-books were supposed to solve that. "Out-of-print" would be a thing of the past.
Now search Amazon for dead-tree books with "barnaby" in the title: there are 148. Now search Amazon e-books: exactly one.
So much for selection. This does not supersede the price or format arguments, but it is another reason why I wouldn't bother to shop for an e-book.
I'll accept that a G3/G4 can run little-endian mode, but do you know that Virtual PC does that? Seems to me that where endian mattered (loads/stores) it would be simpler to stay in Big-endian mode and use its "load/store byte-reversed" opcodes to swap the bytes.
A fast emulator wouldn't "spin" every opcode but would at least identify and convert basic blocks of instructions to the native instruction set for execution.
The parent post implies that G5 does not have an endian mode bit. If true that is a theoretical shame (but a practial don't-care).
A joke? Yes. But a darned insightful one. (Wish I had mod points today.) Parody (and humor in general) is only funny if it contains an element of truth. And the parallel between SCO and RIAA is dead on.
You might consider DVD Studio Pro, which supports a number of external DVD-R drives. It is not free. In fact it is $500. But it provides a bunch of features that iDVD does not, including 16:9 format and higher (greater than 90 minute per DVD) compression. If you want any of its features, or think you might, you may end up buying it later anyway.
Be warned that the current version, 1.5, is not particularly bug-free. Version 2.0, available "in August" might be better.
My favorite change to C would be to make "variables" const by default. This implies that any declaration which did not include an initializer would have to be tagged as "var".
int foo = g();/* OK */ int bar;/* useless */ var int baz;/* OK, we can assign to this */
Remember, that's just an average. Your experience may be different. It could decay all at once just after you swallow. You can't be too careful.
Re:In all non-decimal systems..
on
Eleventy What?
·
· Score: 1
Yeah. And I had a teachers who insisted that "pretty" be pronounced to rhyme with "petty" (not "pity"), "err" rhymes with "slur" (not "air"), "nucleolus" is stressed like "alveolus" and "gigahertz" starts with a soft g. All correct I suppose, but rarely heard nowadays.
I say "thirty-two hex" and even "thirty-two hundred hex" and "charlie thousand hex" on occasion. And my world hasn't collapseD43mjodu4trfk#*(%^&#)$)*(
> Current evolution should be taught as and called "The Law of Evolution."
I'll stick with "Theory," on the grounds that one meaning is "a systematic statement of principles involved." That is broader than "Law," and better describes the standings of both evolution and relativity.
As for the "it's only a theory" people, I doubt that calling evolution a "law" will change their minds.
I worked for IBM when the IBM PC came out (August 12, 1981) and don't remember the company being "deeply mired in red ink" at that time. One reference gives 1990 as the year of IBM's first quarterly loss.
On the one hand... VMWare supports specific OSes, not just any x86 OS. So one wonders how complete their virtualization is.
On the other hand... it should be obvious that software emulation of x86 was not the bulk of VMWare's work. VMWare's major enterpreneurial contribution was the demonstration that it was possible to - pretty much - virtualize a privileged mode x86. I don't know how much device virtualization was in Bochs, but I'll bet VMWare has added a lot of work (not though so much, but effort). Device virtualization is one of those jobs which is never finished.
On the third hand... why spend so much time virtualizing an architecture which wasn't designed for it? That's like seducing an undesirable person just for the challenge. Though in this case it's probably pretty lucrative...
So what other imperfect things should we do away with? Roads with median strips? Safety locks on guns? Metal detectors?
I would argue that strict typing reduces one class of bugs so that I can concentrate on other less tractable classes. Who gave Guido the idea that strong-typing programmers are satisifed with clean compiles. I will be satisfied with clean compiles when the compiler can detect *all* bugs. Until then we (some of us, at least) need to work on improving languages and language tools toward that goal.
> Does anyone know of a way to get iMovie to do 16:9
There are at least 3 ways, all mentioned in the apple.com "support discussions" for iMovie and iDVD. (1) Apply the Virtix letterbox filter. You still get a 4x3 format, but the image is squeezed back to 16:9 with bars added top and bottom. (2) Export the movie (still in "squeezed" aka "anamorphic" form) and modify the resulting Quicktime file to have 16:9 attributes. (3) Export and burn to a DVD, read in the DVD and change the attribute byte to 16:9, and burn again.
2 and 3 give you a "real" 16:9 image; 1 gives you a 16:9 image letterboxed in a 4:3 format.
Slashdot Mirror
> Nevermind that hard disk manufacturers for a long time defined a GB as 1000 * 1000 * 1024 bytes.
The capaxity of the first hard disk was specified in powers of 10, not 2. Every subsequent disk from that manufacturer (IBM) has also been measured in powers of 10, not 2. Just when was IBM supposed to redefine a kilobyte, megabyte and gigabyte of disk storage to accomodate computers which could not support base 10?
Lots of luck. Looks like it didn't make the ballot in 2K2. :-). Remember that Lani Guinier
was denied a federal appointment for being a bit too innovative wrt electoral fairness.
Not that the Democrats haven't pulled equally partisan shenanigans.
>One [Saturn V] is sideways, partially disassembled, exposed to the elements, and "restored," at the Johnson Space Center in Houston
There is one similarly displayed, but indoors, at the Kennedy Space Center.
Shortly after the first moon landing (1969), then-Vice President Spiro Agnew announced that the United States would put a man on Mars by the year 2000. Lets take care of that first, before we attend to the present Command-in-Thief's fancy.
C permits an implementation to trap on overflow. Whether a particular implementation does so is a quality of implementation issue. Of course, some architectures make it trapping easy, others do not. MIPS is quite easy: just emit the right form of the instruction. So is S/370: enable fixed-point overflow. x86 is harder: every overflow check costs you a 1-byte instruction.
I think PowerPC costs an instruction also, and that is a 4-byte instruction. But you don't need to check as often as with x86 because PowerPC keeps a "summary overflow" bit.
I don't know of gcc supports any of these checking mechanisms.
> the first hard drive I ever owned was 10MB - 10240KB
Hard drives go back farther than that. IBM has always quoted hard drive size in base 10. Mainframe could format the tracks to any blocksize they chose (up to actual track size - does 3625 byte ring a bell to anyone?). Most chose a multiple of 80, which was the number of positions on a punched card.
Not having looked at the code, it seems to me that make would only handle half the problem: booting. Shutdown is the other half; the dependencies would be reversed. For example:
For boot you would tell make this:
sshd: network
rpcd: network
But for shutdown you need to tell it this:
network: sshd rpcd
Ideally one set of input data should take care of both cases.
> I sold a Vaio R505 that would whine up and down loudly
...
Well sure. No one likes whiners. But if Apple could license the Harley-Davidson sound
Congratulation to all who made this possible. I don't imagine it was easy.
> SLR is faster because there's no condition code to set.
No, that's not it. SLR does set the condition code. What you are trying to remember is fixed-point overflow, which is possible with SR but not with SLR.
Whether SR really takes longer is arguable nowadays; IBM hasn't published instruction timings for S/370 (except 43xx) for many years. I think the non-DAT 370's may have been the last.
> SLR R1,R1 is 60ns faster than SL R1,R1
:-)
It's the storage reference - unaligned, no less - which kills the SL R1,R1 performance.
Also, "iprmoetnt" is misspelled. SO there's something to be said for correct spelling even when in its absence. :-)
Due to inventory taxes the publishing industry has trouble keeping books in print. E-books were supposed to solve that. "Out-of-print" would be a thing of the past.
Now search Amazon for dead-tree books with "barnaby" in the title: there are 148. Now search Amazon e-books: exactly one.
So much for selection. This does not supersede the price or format arguments, but it is another reason why I wouldn't bother to shop for an e-book.
"Correct" is like "optimized." The unwashed masses think it is an absolute, but we CS elite know otherwise.
I'll accept that a G3/G4 can run little-endian mode, but do you know that Virtual PC does that? Seems to me that where endian mattered (loads/stores) it would be simpler to stay in Big-endian mode and use its "load/store byte-reversed" opcodes to swap the bytes.
A fast emulator wouldn't "spin" every opcode but would at least identify and convert basic blocks of instructions to the native instruction set for execution.
The parent post implies that G5 does not have an endian mode bit. If true that is a theoretical shame (but a practial don't-care).
A joke? Yes. But a darned insightful one. (Wish I had mod points today.) Parody (and humor in general) is only funny if it contains an element of truth. And the parallel between SCO and RIAA is dead on.
Be warned that the current version, 1.5, is not particularly bug-free. Version 2.0, available "in August" might be better.
My favorite change to C would be to make "variables" const by default. This implies that any declaration which did not include an initializer would have to be tagged as "var".
/* OK */ /* useless */ /* OK, we can assign to this */
int foo = g();
int bar;
var int baz;
> At one event per 36 hours
Remember, that's just an average. Your experience may be different. It could decay all at once just after you swallow. You can't be too careful.
Yeah. And I had a teachers who insisted that "pretty" be pronounced to rhyme with "petty" (not "pity"), "err" rhymes with "slur" (not "air"), "nucleolus" is stressed like "alveolus" and "gigahertz" starts with a soft g. All correct I suppose, but rarely heard nowadays.
I say "thirty-two hex" and even "thirty-two hundred hex" and "charlie thousand hex" on occasion. And my world hasn't collapseD43mjodu4trfk#*(%^&#)$)*(
> Current evolution should be taught as and called "The Law of Evolution."
I'll stick with "Theory," on the grounds that one meaning is "a systematic statement of principles involved." That is broader than "Law," and better describes the standings of both evolution and relativity.
As for the "it's only a theory" people, I doubt that calling evolution a "law" will change their minds.
I worked for IBM when the IBM PC came out (August 12, 1981) and don't remember the company being "deeply mired in red ink" at that time. One reference gives 1990 as the year of IBM's first quarterly loss.
On the one hand ... VMWare supports specific OSes, not just any x86 OS. So one wonders how complete their virtualization is.
... it should be obvious that software emulation of x86 was not the bulk of VMWare's work. VMWare's major enterpreneurial contribution was the demonstration that it was possible to - pretty much - virtualize a privileged mode x86. I don't know how much device virtualization was in Bochs, but I'll bet VMWare has added a lot of work (not though so much, but effort). Device virtualization is one of those jobs which is never finished.
... why spend so much time virtualizing an architecture which wasn't designed for it? That's like seducing an undesirable person just for the challenge. Though in this case it's probably pretty lucrative ...
On the other hand
On the third hand
So what other imperfect things should we do away with? Roads with median strips? Safety locks on guns? Metal detectors?
I would argue that strict typing reduces one class of bugs so that I can concentrate on other less tractable classes. Who gave Guido the idea that strong-typing programmers are satisifed with clean compiles. I will be satisfied with clean compiles when the compiler can detect *all* bugs. Until then we (some of us, at least) need to work on improving languages and language tools toward that goal.
> Does anyone know of a way to get iMovie to do 16:9
There are at least 3 ways, all mentioned in the apple.com "support discussions" for iMovie and iDVD. (1) Apply the Virtix letterbox filter. You still get a 4x3 format, but the image is squeezed back to 16:9 with bars added top and bottom. (2) Export the movie (still in "squeezed" aka "anamorphic" form) and modify the resulting Quicktime file to have 16:9 attributes. (3) Export and burn to a DVD, read in the DVD and change the attribute byte to 16:9, and burn again.
2 and 3 give you a "real" 16:9 image; 1 gives you a 16:9 image letterboxed in a 4:3 format.