Buy this, and you are giving money to Disney. True.
But I seriously doubt the effectiveness of a boycott. If all the anime fans boycott the Mononoke DVD, will that block the DeCSS lawsuits? No, but it will probably convince Disney to import more Power Rangers next time instead of more Miyazaki-san.
Disney isn't a monolith; it's made up of various groups. The part of Disney that decided to import Mononoke and release it on DVD should be rewarded. I'm not happy with any part of Disney that sends money or lawyers to the MPAA to attack DeCSS, but it's hard to punish them by refusing to buy Mononoke on DVD.
If you would like to own this but you feel bad about supporting MPAA even indirectly, may I suggest that you send a contribution to the EFF when you buy a DVD. Feel free even to send a letter to Disney telling them you did this.
Princess Mononoke on DVD with a clean transfer and Japanese language and two English subtitles tracks? Yep, I'm gonna buy that.
It really reminded me of the classic old Warner Brothers cartoons (as I was laughing out loud at absurd things).
It won't corrupt any kids, but it also doesn't beat anyone over the head with a moral. It has a very small bit of music, but it isn't a musical. There are some bad guys, but only one of them is bad and they are both more funny than anything else. If you like goofy "toon" humor, go see it!
One example. The emperor is walking, and he decides to walk straight instead of turning right; a swarm of workmen appears, and with furious speed they build a new entrance to the palace and then fall on their backs gasping. The emperor walks straight through the new door. Total elapsed time: about three seconds... there are a lot of goofy throwaway gags like that in there.
I plan to buy a copy. It's the first Disney movie in a long time I have liked that much.
Corel ran magazine ads, month after month. Full-page color ads. And you know what? The ads sucked.
There was a picture of barbells, where the weights were blue squares (to match the color theme of the Corel box) and a headline that said "Power**3" or something like that. There was this picture of an older guy, with a lip piercing, and a little blue square on the lip ring. And so on.
These are the sort of ads you expect Apple to run: too-clever ads that don't reach out to new customers.
They should have said "Gain the power of Linux without retraining your people." "Stability and reliability are now easy to install and have a familiar face." Focus on the strengths of the product; don't try to make it cool. Business people won't buy Linux based on coolness!
There was a legendary bad ad years ago showing a human brain with a floppy disk stuck into it. That lip-piercing guy ad was almost as bad. Corel might as well have taken their ad budget and burned it.
They didn't go after all the Aqua themes. For example, the
auqaos2
theme is still there.
Did they just miss this one? It has been there a long time ("Updated 4th February 2000"). So, I wonder how Apple's lawyers decide which ones to go after and which ones to ignore.
It's as if whatever MS does, they will succeed. If MS started selling cheese tomorrow, every other cheese maker on the planet is supposed to ceremoniously drown themselves in big vats of milk?
Funny you should say that.
Robert X. Cringely
once observed that whenever Microsoft announces that it is going after a particular market, whoever owns that market at the time gets nervous and starts doing things differently, in order to "get ready" for Microsoft. He further said that it is usually a mistake; if they already own the market they must be doing something right, and by suddenly doing things differently they often screw up a good thing.
The front of the Linux distro box should have a logo on it: GameReady 1 Compatible!
GameReady Level 1, or whatever it winds up actually being called, will be a standard, non-moving target. It doesn't matter much what the standard is: for example, I don't care if the 3D part is OpenGL, or some other API that can wrap around OpenGL. The standard will include everything needed to run cool games: a 3D part, a sound part, a 3D sound part, etc.
This is important because you want people to be able to look at the requirements for Linux Quake4 and say "Hey! My computer is GameReady 1!" You don't want to have a long list of 10 different libraries that are required to run the game.
DirectX was valuable because it helped games run more efficiently, but it was also valuable because it provides a unified standard the game companies can write to. We need something similar.
It's also important to get a number in there, so that someday when cool new stuff is invented, it can be standardized as GameReady Level 2.
I have nothing against Red Hat, but I hope never to see games saying "RedHat Compatible". I'd rather see a more open standard.
P.S. When I say "the front of the box", I also mean "on the download page" or anywhere else you get software. I by no means intend that this apply only to boxed retail sales.
In my experience, a software engineer will spend a lot more time modifying existing code than writing new code. And, if you are looking for experience, looking at other peoples' code will be helpful.
Therefore I suggest you find some existing project and add some kind of new features to it. Don't pick a huge, complicated project like Mozilla; start small.
As a concrete, specific suggestion: the Gnome version of
Freecell
has, IMHO, ugly playing cards. There is no way to change them. On the other hand, the
Same Gnome
game allows the user to load different images for the playing pieces. It might be a good project for you to study Freecell, study the Same Gnome code for loading images, and then modify Freecell to allow the user to load different playing card images.
This is a project that will take a certain amount of studying, but it shouldn't take months of work. To some extent you would be able to cut-and-paste code from Same Gnome, but you would also have to modify it. When it is done you will have something you can show to even your non-techie friends! I'm not trying to tell you what to do, so if you don't like this suggestion do something else.
By the way, if you look at the
TODO
file in the Gnome Freecell sources, you will notice an item that says "Make card bitmaps library." In other words, my suggestion was already thought up by the people working on Freecell. You might want to look at the various projects and read the notes in each, and perhaps you will see something you really want to work on.
By the way, I have to say I like your attitude. I hope you have lots of fun with this project.
Everyone rips on the Lynch movie but I think it was great.
I thought the pacing was all wrong. They spent way too much time setting up, and then had no time to play everything out.
I distinctly remember looking at my watch, and thinking "An hour has gone by! They are only at about page 100 and they have only an hour left for the whole rest of the story!?"
The telephone companies have a point that we should all consider: simple fairness.
If you are in the business of providing voice service, you are subject to a huge maze of rules, regulations, and taxes. People doing voice over IP are not forced to deal with this maze. It's unfair to the telcos.
Now, the solution the telcos are asking for is to force everyone else into the maze with them. This is entirely rational from their point of view. I, and other libertarians, would prefer a different approach: wipe out the maze and leave the telcos alone.
One poster suggested that a small farm town would be unable to get affordable phone service, but I find that very difficult to believe. Is that same small farm town unable to get affordable food? How about affordable computers? What is magic about telephone service that makes it impossible for a free market to deliver it cheaply?
Just maybe in the early days of telephony it was actually necessary that government set up monopoly telcos and regulate them, but it certainly isn't true now. If there is a problem with getting a phone, you can get a cell phone. If company A owns all the wires going into a town, company B can set up a microwave relay, route voice onto an Internet backbone, or maybe make a deal with the power company to put voice data on the power lines. (Don't laugh; I've heard that some places in Europe already do Internet service through power lines.)
If telcos have to pay the Al Gore tax, every voice user should have to pay it. But I say get rid of it instead of spreading it around even more.
I haven't read Starship Troopers, but I suspect the fascist tone is not intended in praise but is a p**s take of the first order.
Heinlein said, in an essay, that Starship Troopers was accused of "glorifying" the military; he said it was more intended to praise the footsoldier, the guy who has to do the fighting, rather than the military as a whole. You can make a case that in ST, the most admirable characters are the lowest in rank.
ST is about responsability. It isn't any kind of satire.
These days, we expect a CPU chip to have floating-point math (FPU for short) built in, but in the early days the FPU was an extra chip. And only people who really crunched numbers bothered to pay for it. So, the 286 had a 287 FPU; the 386 had a 387 FPU; and so on.
The 486 had its FPU built in. Then Intel got the brilliant idea to sell 486 chips whose FPU was defective, as a special "486SX" chip that would cost less. People who wanted to save money and were willing to live without an FPU bought the 486SX. So far, I'm not too upset. But the 487SX made me upset.
If you upgrade a computer with a 486SX, how do you do it? Most people would say you should remove the 486SX and plug in a 486DX. After all, the two chips were pin-compatible (remember that the original 486SX was a way to sell defective 486DX chips). But Intel had a better idea. They put a second CPU socket on the motherboard, the "487SX" socket. You were supposed to leave the 486SX chip plugged in to the main socket, and then buy a 487SX chip and plug that into the second socket. The 487SX was nothing more than a 486DX, and it would take over and run your computer as the sole CPU chip! (If I recall correctly, it had a special extra pin so it would not plug into a 486 socket, but I won't swear to it.)
So, Intel wanted you to have a perfectly good 486SX chip locked up in one socket, asleep doing nothing; and buy a special 487SX chip, when you could have just bought an ordinary 486DX chip for less money.
As you can imagine, people didn't do what Intel wanted. Most 486 systems were built with a ZIF socket for the CPU, making it easy to remove and replace the CPU. No one who was cheap enough to buy a 486SX in the first place was willing to throw money away just to make Intel happy.
Neither is right or wrong or more or less efficient.
I read an article about this. Most of the world reads from left to right, but numbers are read from right to left; you have to count backwards from the right to know how big the first digit is. The whole endian-ness issue would probably never have arisen if our numbers were written left-to-right.
Big-endian numbers are nice when a human wants to read hex dumps; the numbers appear in the hex dump the way they appear written out by hand.
Little-endian numbers are nice when you want to read a number from a pointer: if you want the low 8 bits, you don't actually care whether you were passed an 8-bit, 16-bit, or 32-bit integer. In a big-endian architecture, you need to know what size integer the pointer is pointing to, to shift over correctly.
I think the 286 had some brain damage about going in and out of protected mode
Oh yeah. The 286 had the ability to switch to protect mode... and no ability to leave protect mode! The chip would boot up in real mode, but once you switched to protect mode there was no return. I'm still amazed Intel never thought of this or added it before it shipped.
This might not have mattered so much if 286 protect mode had decent backwards compatability, but it did not. In protect mode, a 286 cannot run real-mode programs (i.e., programs written to run on an 8088 or 8086) at all; compare with 386 protect mode, which lets you run real-mode code inside a protected virtual machine, making it possible to run multiple real-mode programs safely.
At the time, everyone was running real-mode DOS programs, so 286 OS/2 really needed to be able to run them. Microsoft did the only thing possible: they switched the 286 in and out of protect mode. Native OS/2 programs ran in protect mode, and real-mode DOS programs had to run in real mode. But Intel had provided no way to switch out of protect mode! Microsoft came up with an amazing hack: they programmed the keyboard controller to force a reset on the CPU, as a way to force the 286 back into real mode. They did this many times a second, to allow DOS and OS/2 programs to multitask together. Ugly, but thanks to Intel they had no choice.
Since real mode DOS programs were not running in protect mode, it was very easy for them to crash the whole computer, and there was nothing OS/2 could do to prevent this. This is why people called the DOS program "compatability box" the "Chernobyl box". When Windows 3.0 shipped with 386 protect-mode support, people could safely run their old DOS programs, and Windows 3.0 sales took off.
286 protect mode, even if you didn't want to leave it, was all messed up. Segments were much bigger, but still present; you had to use them, and unfortunately changing the segment registers was very slow. Multitasking operating systems took a big performance hit by having to reprogram the segment selectors every time they needed to switch processes. The 386 and onward have segments, but no one wants to use them; instead of using segments to protect running programs from stepping on each other, modern operating systems use the memory manager chip.
In short, the 286 was so bad that MS would have been better off not trying to use it for OS/2. Intel messed up big-time.
The name of the chip by AMD was the DX4 (not 486 DX4).
The original poster was right: Intel shipped a chip called 486DX4 that was internally a 3x clock, not a 4x clock. The AMD chip was called the "AM486". Intel sued AMD, saying they couldn't use the "486" but lost; Intel is willing to try to own numbers, but the courts haven't cooperated so far.
Now, if you saw the crap that Microsoft was making before they stopped off at Apple one day (can we say DOS Shell and Windows 1.0?) you would understand how completely ripped Apple must of felt.
Both MS and Apple saw the Xerox PARC stuff and wanted to make their own versions. MS knew what Apple was doing, and for a while tried to be different just to be different; for example, Windows 1.x didn't have overlapping windows, but rather tiled windows. I don't know what MS was thinking -- trying to keep Apple happy? Trying to differentiate their product?
But don't forget that Apple really, really wanted MS to support the struggling new Mac platform. MS, well aware that Windows could compete with the Mac someday when they got it right, demanded that Apple sign an agreement: if MS ported applications like Word and Multiplan to the Mac, Apple would not sue MS for features in MS Windows. Apple signed the agreement, and MacWord and MacMultiplan helped sell the Mac platform among businesses. (Then, a few years later, Apple sued MS over Windows anyway, and eventually the entire lawsuit was thrown out because of the agreement.)
So, given the above history, you would understand how completely betrayed MS must have felt when Apple sued them after agreeing not to sue them.
Actually, I was thinking fuzzy when I wrote all that. Plex86 isn't so much an emulation system, but rather a system for making virtual machines that the x86 runs directly, without messing up the parent operating system that is also being run by the x86. In short, Plex86 is to let you run Win95 in protect mode, inside Linux running in protect mode, and make sure neither one steps on the other.
I should think that emulation and Plex86 have many problems in common. Just think of "virtualizing" the controllers rather than "emulating" them, and what I wrote probably makes sense.
plex86 is a simulation of an x86 system, so why wouldn't "proprietary microsoft stuff" work?
If Plex86 were a perfect emulated x86 system, then sure everything would work. But perfect is difficult.
The first thing they need to get right is full and correct emulation of every x86 instruction. The easiest part would be the older 8086 instructions; those were no doubt perfect long ago. But the modern chips have lots of interesting protect-mode instructions, and I don't know if all of those are 100% perfect yet. For example, is it easy to emulate the x86 memory manager features perfectly (page faults, address remapping, etc. etc.)? I'm not an expert, but I think there are a few land mines there.
So maybe they are done with the x86 instruction set; maybe it's 100% perfect. They are still not done. No computer is just an x86 chip, sitting there alone. There are interrupt controller chips. There are disk controllers, DMA controllers, etc. etc. As an old, outdated example: OS/2 for the 286 switched in and out of protect mode by reprogramming the keyboard controller chip, so you won't be able to run 286 OS/2 on an emulated system unless the keyboard controller is emulated perfectly.
I wonder if you could get a team of Windows driver gurus together to make a set of "emulation drivers" that would let Windows 98 run happily on top of Plex86 without perfectly emulating all the low-level system chips. For example, instead of perfectly emulating an IDE controller, make a Windows disk driver that simply passes disk I/O requests to the Linux kernel. Might not be worth the effort... the work might be better done on emulating the controller chips.
Wouldn't it be more accurate to describe the rendering portion as a "client"
I agree that the X nomenclature is confusing, but it does make sense of a sort. A server, by definition, serves something; what the X server serves is graphic windows.
If you use an old, cheap X terminal to work with a spiffy fast UNIX box, the UNIX box serves up data and the X terminal serves up graphic windows.
But I agree with you: I wish they had called it the X "client".
If Netscape killed Mozilla funding, that would be a very serious blow which Mozilla might not survive.
No. Mozilla has reached critical mass... all anyone could do to it now is delay it. I'm running Mozilla nightly builds both under Windows and under Linux, and it is stable and fast for me. In comparison with the amount of work it took to get it this close to done, the amount remaining is small.
What's more, the remaining work is exactly the kind of work that open-source projects do best: taking something that mostly works, fixing a few bugs, and adding a few features. Loss of Netscape resources would be a blow, but it is far too late for it to be a fatal one.
steveha
Two words: paper trail
on
eLection '04
·
· Score: 1
I am opposed to voting by Internet. Not just "no"; my reaction is "Hell, no!"
Paper ballots give you a paper trail. I want it to be as hard as possible to corrupt the voting process, and I think paper ballots are best if you have to do a recount.
That's not to say that I am a fan of those little hole-punch ballots. Washington state has gone to a new system I like better: your ballot is a simple printed piece of heavy paper (or maybe it's lightweight cardboard), and there are ovals next to the various names. To vote for Bush, you fill in the oval next to Bush's name. When you are done, you feed the ballot into a machine that tallies it instantly; if you filled in two bubbles, it beeps and kicks the ballot back out, and you get a fresh ballot and fill it out again. When the polls close, they take the machines, hook them up to a phone line, and the machines report the results.
I like this system a lot. Easy to vote, you get a second chance if you screw up, the results are reported quickly... and there is a paper trail.
P.S. I've used this system in two elections now. In the first election, they loaned out Sharpie permanent markers and that was the quickest I have ever voted in my life. This year they gave us ordinary black ballpoint pens, and it takes much longer to fill in the bubble with one of those. Next election I'm going to bring my own Sharpie.
Bill Gates does not deny reality the way that essay portrays him. He makes Andy Grove look like a pollyanna.
When I worked at Microsoft, I never met the man personally, but I did attend mass meetings where he would give a lecture on the state of the company. All the Microsoft management were consistent in the way they discussed the competition -- it was always "don't feel we are safe, don't count them out." I remember when it was first becoming clear that Borland was no longer any sort of threat to Microsoft, someone (I think it was Steve Ballmer) said "But Borland is the 'comeback kid'. We still need to watch them."
BillG has his blind spots -- he didn't see the Internet coming, nor did he see Linux coming. But I guarantee you that he considers Linux a threat and he is watching it. The ridicule in the essay rings totally false.
The surprise settlement, which appears to be lower than Caldera's earlier demands, defused a number of potential antitrust time bombs for the software giant. Although many legal analysts thought Microsoft had strong arguments in its favor, a ruling against it could have given fuel to recently filed class-action suits.
In other words, they settled out of court, for much less than Caldera was asking, to get it over with quickly and without admitting guilt. Had they gone through the courts, paying lawyers and getting more publicity they wouldn't like, they might possibly have lost, which would have been a disaster. (And they settled cheap: Caldera was going for $400 million, plus $1.6 billion in punitive damages, for a total of $2 billion; the settlement was about $150 million, or 92.5 percent less than Caldera was going for.)
And I quote from the Planet IT article, on what Caldera charged Microsoft with:
The suit alleged that Microsoft wrongfully stifled competition by pre-announcing new software products, notably MS-DOS, before it was ready, in a deliberate attempt to discourage users of DR-DOS, which was later purchased by Caldera from Novell. The suit, which alleged numerous violations of antitrust laws, was filed in federal court in Salt Lake City.
And I still never saw "test for DR-DOS and break" in the code.
As with most Microsoft software packages, the version for Macintosh systems is an almost completely separate (and much cleaner) codebase.
Could you give more details? That's now how it worked when I was there. When I wrote code in Word 97, I had to compile my code under Windows and under the Macintosh to make sure it worked on both platforms.
Here is the history: Mac Word came before Windows Word. WinWord, I'm sure, borrowed a lot of code form MacWord but they were separate code bases and separate development teams. Over time, WinWord gained hundreds of features MacWord didn't have, and the Mac market started shrinking. In order to reduce development expenses, and bring MacWord to 100% feature parity, a porting library (not unlike WINE!) was used to port the WinWord code base to the Mac. Excel, on the other hand, was ahead of Word in cross-platform code: long ago they cleaned up their code and were able to compile the same code base for Windows, Mac, and OS/2 (once upon a time when MS cared about OS/2).
I know that Internet Explorer has a separate code base and development team, but I don't know of any Microsoft applications where this is true.
I haven't worked there for four years. It is possible that they forked the code base since then, but the reasons they un-forked are stronger than ever so I simply do not believe it.
Slashdot Mirror
But I seriously doubt the effectiveness of a boycott. If all the anime fans boycott the Mononoke DVD, will that block the DeCSS lawsuits? No, but it will probably convince Disney to import more Power Rangers next time instead of more Miyazaki-san.
Disney isn't a monolith; it's made up of various groups. The part of Disney that decided to import Mononoke and release it on DVD should be rewarded. I'm not happy with any part of Disney that sends money or lawyers to the MPAA to attack DeCSS, but it's hard to punish them by refusing to buy Mononoke on DVD.
If you would like to own this but you feel bad about supporting MPAA even indirectly, may I suggest that you send a contribution to the EFF when you buy a DVD. Feel free even to send a letter to Disney telling them you did this.
Princess Mononoke on DVD with a clean transfer and Japanese language and two English subtitles tracks? Yep, I'm gonna buy that.
steveha
It won't corrupt any kids, but it also doesn't beat anyone over the head with a moral. It has a very small bit of music, but it isn't a musical. There are some bad guys, but only one of them is bad and they are both more funny than anything else. If you like goofy "toon" humor, go see it!
One example. The emperor is walking, and he decides to walk straight instead of turning right; a swarm of workmen appears, and with furious speed they build a new entrance to the palace and then fall on their backs gasping. The emperor walks straight through the new door. Total elapsed time: about three seconds... there are a lot of goofy throwaway gags like that in there.
I plan to buy a copy. It's the first Disney movie in a long time I have liked that much.
steveha
There was a picture of barbells, where the weights were blue squares (to match the color theme of the Corel box) and a headline that said "Power**3" or something like that. There was this picture of an older guy, with a lip piercing, and a little blue square on the lip ring. And so on.
These are the sort of ads you expect Apple to run: too-clever ads that don't reach out to new customers.
They should have said "Gain the power of Linux without retraining your people." "Stability and reliability are now easy to install and have a familiar face." Focus on the strengths of the product; don't try to make it cool. Business people won't buy Linux based on coolness!
There was a legendary bad ad years ago showing a human brain with a floppy disk stuck into it. That lip-piercing guy ad was almost as bad. Corel might as well have taken their ad budget and burned it.
steveha
Did they just miss this one? It has been there a long time ("Updated 4th February 2000"). So, I wonder how Apple's lawyers decide which ones to go after and which ones to ignore.
steveha
However, makers of $1000 PCs are still willing to pay it. I doubt this will change within six months.
steveha
Funny you should say that. Robert X. Cringely once observed that whenever Microsoft announces that it is going after a particular market, whoever owns that market at the time gets nervous and starts doing things differently, in order to "get ready" for Microsoft. He further said that it is usually a mistake; if they already own the market they must be doing something right, and by suddenly doing things differently they often screw up a good thing.
steveha
GameReady Level 1, or whatever it winds up actually being called, will be a standard, non-moving target. It doesn't matter much what the standard is: for example, I don't care if the 3D part is OpenGL, or some other API that can wrap around OpenGL. The standard will include everything needed to run cool games: a 3D part, a sound part, a 3D sound part, etc.
This is important because you want people to be able to look at the requirements for Linux Quake4 and say "Hey! My computer is GameReady 1!" You don't want to have a long list of 10 different libraries that are required to run the game.
DirectX was valuable because it helped games run more efficiently, but it was also valuable because it provides a unified standard the game companies can write to. We need something similar.
It's also important to get a number in there, so that someday when cool new stuff is invented, it can be standardized as GameReady Level 2.
I have nothing against Red Hat, but I hope never to see games saying "RedHat Compatible". I'd rather see a more open standard.
P.S. When I say "the front of the box", I also mean "on the download page" or anywhere else you get software. I by no means intend that this apply only to boxed retail sales.
steveha
Therefore I suggest you find some existing project and add some kind of new features to it. Don't pick a huge, complicated project like Mozilla; start small.
As a concrete, specific suggestion: the Gnome version of Freecell has, IMHO, ugly playing cards. There is no way to change them. On the other hand, the Same Gnome game allows the user to load different images for the playing pieces. It might be a good project for you to study Freecell, study the Same Gnome code for loading images, and then modify Freecell to allow the user to load different playing card images.
This is a project that will take a certain amount of studying, but it shouldn't take months of work. To some extent you would be able to cut-and-paste code from Same Gnome, but you would also have to modify it. When it is done you will have something you can show to even your non-techie friends! I'm not trying to tell you what to do, so if you don't like this suggestion do something else.
By the way, if you look at the TODO file in the Gnome Freecell sources, you will notice an item that says "Make card bitmaps library." In other words, my suggestion was already thought up by the people working on Freecell. You might want to look at the various projects and read the notes in each, and perhaps you will see something you really want to work on.
By the way, I have to say I like your attitude. I hope you have lots of fun with this project.
steveha
I thought the pacing was all wrong. They spent way too much time setting up, and then had no time to play everything out.
I distinctly remember looking at my watch, and thinking "An hour has gone by! They are only at about page 100 and they have only an hour left for the whole rest of the story!?"
steveha
If you are in the business of providing voice service, you are subject to a huge maze of rules, regulations, and taxes. People doing voice over IP are not forced to deal with this maze. It's unfair to the telcos.
Now, the solution the telcos are asking for is to force everyone else into the maze with them. This is entirely rational from their point of view. I, and other libertarians, would prefer a different approach: wipe out the maze and leave the telcos alone.
One poster suggested that a small farm town would be unable to get affordable phone service, but I find that very difficult to believe. Is that same small farm town unable to get affordable food? How about affordable computers? What is magic about telephone service that makes it impossible for a free market to deliver it cheaply?
Just maybe in the early days of telephony it was actually necessary that government set up monopoly telcos and regulate them, but it certainly isn't true now. If there is a problem with getting a phone, you can get a cell phone. If company A owns all the wires going into a town, company B can set up a microwave relay, route voice onto an Internet backbone, or maybe make a deal with the power company to put voice data on the power lines. (Don't laugh; I've heard that some places in Europe already do Internet service through power lines.)
If telcos have to pay the Al Gore tax, every voice user should have to pay it. But I say get rid of it instead of spreading it around even more.
steveha
Heinlein said, in an essay, that Starship Troopers was accused of "glorifying" the military; he said it was more intended to praise the footsoldier, the guy who has to do the fighting, rather than the military as a whole. You can make a case that in ST, the most admirable characters are the lowest in rank.
ST is about responsability. It isn't any kind of satire.
steveha
These days, we expect a CPU chip to have floating-point math (FPU for short) built in, but in the early days the FPU was an extra chip. And only people who really crunched numbers bothered to pay for it. So, the 286 had a 287 FPU; the 386 had a 387 FPU; and so on.
The 486 had its FPU built in. Then Intel got the brilliant idea to sell 486 chips whose FPU was defective, as a special "486SX" chip that would cost less. People who wanted to save money and were willing to live without an FPU bought the 486SX. So far, I'm not too upset. But the 487SX made me upset.
If you upgrade a computer with a 486SX, how do you do it? Most people would say you should remove the 486SX and plug in a 486DX. After all, the two chips were pin-compatible (remember that the original 486SX was a way to sell defective 486DX chips). But Intel had a better idea. They put a second CPU socket on the motherboard, the "487SX" socket. You were supposed to leave the 486SX chip plugged in to the main socket, and then buy a 487SX chip and plug that into the second socket. The 487SX was nothing more than a 486DX, and it would take over and run your computer as the sole CPU chip! (If I recall correctly, it had a special extra pin so it would not plug into a 486 socket, but I won't swear to it.)
So, Intel wanted you to have a perfectly good 486SX chip locked up in one socket, asleep doing nothing; and buy a special 487SX chip, when you could have just bought an ordinary 486DX chip for less money.
As you can imagine, people didn't do what Intel wanted. Most 486 systems were built with a ZIF socket for the CPU, making it easy to remove and replace the CPU. No one who was cheap enough to buy a 486SX in the first place was willing to throw money away just to make Intel happy.
steveha
I read an article about this. Most of the world reads from left to right, but numbers are read from right to left; you have to count backwards from the right to know how big the first digit is. The whole endian-ness issue would probably never have arisen if our numbers were written left-to-right.
Big-endian numbers are nice when a human wants to read hex dumps; the numbers appear in the hex dump the way they appear written out by hand.
Little-endian numbers are nice when you want to read a number from a pointer: if you want the low 8 bits, you don't actually care whether you were passed an 8-bit, 16-bit, or 32-bit integer. In a big-endian architecture, you need to know what size integer the pointer is pointing to, to shift over correctly.
Either one gets the job done.
steveha
Oh yeah. The 286 had the ability to switch to protect mode... and no ability to leave protect mode! The chip would boot up in real mode, but once you switched to protect mode there was no return. I'm still amazed Intel never thought of this or added it before it shipped.
This might not have mattered so much if 286 protect mode had decent backwards compatability, but it did not. In protect mode, a 286 cannot run real-mode programs (i.e., programs written to run on an 8088 or 8086) at all; compare with 386 protect mode, which lets you run real-mode code inside a protected virtual machine, making it possible to run multiple real-mode programs safely.
At the time, everyone was running real-mode DOS programs, so 286 OS/2 really needed to be able to run them. Microsoft did the only thing possible: they switched the 286 in and out of protect mode. Native OS/2 programs ran in protect mode, and real-mode DOS programs had to run in real mode. But Intel had provided no way to switch out of protect mode! Microsoft came up with an amazing hack: they programmed the keyboard controller to force a reset on the CPU, as a way to force the 286 back into real mode. They did this many times a second, to allow DOS and OS/2 programs to multitask together. Ugly, but thanks to Intel they had no choice.
Since real mode DOS programs were not running in protect mode, it was very easy for them to crash the whole computer, and there was nothing OS/2 could do to prevent this. This is why people called the DOS program "compatability box" the "Chernobyl box". When Windows 3.0 shipped with 386 protect-mode support, people could safely run their old DOS programs, and Windows 3.0 sales took off.
286 protect mode, even if you didn't want to leave it, was all messed up. Segments were much bigger, but still present; you had to use them, and unfortunately changing the segment registers was very slow. Multitasking operating systems took a big performance hit by having to reprogram the segment selectors every time they needed to switch processes. The 386 and onward have segments, but no one wants to use them; instead of using segments to protect running programs from stepping on each other, modern operating systems use the memory manager chip.
In short, the 286 was so bad that MS would have been better off not trying to use it for OS/2. Intel messed up big-time.
steveha
The original poster was right: Intel shipped a chip called 486DX4 that was internally a 3x clock, not a 4x clock. The AMD chip was called the "AM486". Intel sued AMD, saying they couldn't use the "486" but lost; Intel is willing to try to own numbers, but the courts haven't cooperated so far.
steveha
Both MS and Apple saw the Xerox PARC stuff and wanted to make their own versions. MS knew what Apple was doing, and for a while tried to be different just to be different; for example, Windows 1.x didn't have overlapping windows, but rather tiled windows. I don't know what MS was thinking -- trying to keep Apple happy? Trying to differentiate their product?
But don't forget that Apple really, really wanted MS to support the struggling new Mac platform. MS, well aware that Windows could compete with the Mac someday when they got it right, demanded that Apple sign an agreement: if MS ported applications like Word and Multiplan to the Mac, Apple would not sue MS for features in MS Windows. Apple signed the agreement, and MacWord and MacMultiplan helped sell the Mac platform among businesses. (Then, a few years later, Apple sued MS over Windows anyway, and eventually the entire lawsuit was thrown out because of the agreement.)
So, given the above history, you would understand how completely betrayed MS must have felt when Apple sued them after agreeing not to sue them.
steveha
I should think that emulation and Plex86 have many problems in common. Just think of "virtualizing" the controllers rather than "emulating" them, and what I wrote probably makes sense.
steveha
If Plex86 were a perfect emulated x86 system, then sure everything would work. But perfect is difficult.
The first thing they need to get right is full and correct emulation of every x86 instruction. The easiest part would be the older 8086 instructions; those were no doubt perfect long ago. But the modern chips have lots of interesting protect-mode instructions, and I don't know if all of those are 100% perfect yet. For example, is it easy to emulate the x86 memory manager features perfectly (page faults, address remapping, etc. etc.)? I'm not an expert, but I think there are a few land mines there.
So maybe they are done with the x86 instruction set; maybe it's 100% perfect. They are still not done. No computer is just an x86 chip, sitting there alone. There are interrupt controller chips. There are disk controllers, DMA controllers, etc. etc. As an old, outdated example: OS/2 for the 286 switched in and out of protect mode by reprogramming the keyboard controller chip, so you won't be able to run 286 OS/2 on an emulated system unless the keyboard controller is emulated perfectly.
I wonder if you could get a team of Windows driver gurus together to make a set of "emulation drivers" that would let Windows 98 run happily on top of Plex86 without perfectly emulating all the low-level system chips. For example, instead of perfectly emulating an IDE controller, make a Windows disk driver that simply passes disk I/O requests to the Linux kernel. Might not be worth the effort... the work might be better done on emulating the controller chips.
steveha
I agree that the X nomenclature is confusing, but it does make sense of a sort. A server, by definition, serves something; what the X server serves is graphic windows.
If you use an old, cheap X terminal to work with a spiffy fast UNIX box, the UNIX box serves up data and the X terminal serves up graphic windows.
But I agree with you: I wish they had called it the X "client".
steveha
First on my wish list for Linux: anti-aliased fonts.
Who is working on this? Is anyone?
steveha
No. Mozilla has reached critical mass... all anyone could do to it now is delay it. I'm running Mozilla nightly builds both under Windows and under Linux, and it is stable and fast for me. In comparison with the amount of work it took to get it this close to done, the amount remaining is small.
What's more, the remaining work is exactly the kind of work that open-source projects do best: taking something that mostly works, fixing a few bugs, and adding a few features. Loss of Netscape resources would be a blow, but it is far too late for it to be a fatal one.
steveha
Paper ballots give you a paper trail. I want it to be as hard as possible to corrupt the voting process, and I think paper ballots are best if you have to do a recount.
That's not to say that I am a fan of those little hole-punch ballots. Washington state has gone to a new system I like better: your ballot is a simple printed piece of heavy paper (or maybe it's lightweight cardboard), and there are ovals next to the various names. To vote for Bush, you fill in the oval next to Bush's name. When you are done, you feed the ballot into a machine that tallies it instantly; if you filled in two bubbles, it beeps and kicks the ballot back out, and you get a fresh ballot and fill it out again. When the polls close, they take the machines, hook them up to a phone line, and the machines report the results.
I like this system a lot. Easy to vote, you get a second chance if you screw up, the results are reported quickly... and there is a paper trail.
P.S. I've used this system in two elections now. In the first election, they loaned out Sharpie permanent markers and that was the quickest I have ever voted in my life. This year they gave us ordinary black ballpoint pens, and it takes much longer to fill in the bubble with one of those. Next election I'm going to bring my own Sharpie.
steveha
When I worked at Microsoft, I never met the man personally, but I did attend mass meetings where he would give a lecture on the state of the company. All the Microsoft management were consistent in the way they discussed the competition -- it was always "don't feel we are safe, don't count them out." I remember when it was first becoming clear that Borland was no longer any sort of threat to Microsoft, someone (I think it was Steve Ballmer) said "But Borland is the 'comeback kid'. We still need to watch them."
BillG has his blind spots -- he didn't see the Internet coming, nor did he see Linux coming. But I guarantee you that he considers Linux a threat and he is watching it. The ridicule in the essay rings totally false.
steveha
I quote from cnet.com news:
In other words, they settled out of court, for much less than Caldera was asking, to get it over with quickly and without admitting guilt. Had they gone through the courts, paying lawyers and getting more publicity they wouldn't like, they might possibly have lost, which would have been a disaster. (And they settled cheap: Caldera was going for $400 million, plus $1.6 billion in punitive damages, for a total of $2 billion; the settlement was about $150 million, or 92.5 percent less than Caldera was going for.)
And I quote from the Planet IT article, on what Caldera charged Microsoft with:
And I still never saw "test for DR-DOS and break" in the code.
steveha
Could you give more details? That's now how it worked when I was there. When I wrote code in Word 97, I had to compile my code under Windows and under the Macintosh to make sure it worked on both platforms.
Here is the history: Mac Word came before Windows Word. WinWord, I'm sure, borrowed a lot of code form MacWord but they were separate code bases and separate development teams. Over time, WinWord gained hundreds of features MacWord didn't have, and the Mac market started shrinking. In order to reduce development expenses, and bring MacWord to 100% feature parity, a porting library (not unlike WINE!) was used to port the WinWord code base to the Mac. Excel, on the other hand, was ahead of Word in cross-platform code: long ago they cleaned up their code and were able to compile the same code base for Windows, Mac, and OS/2 (once upon a time when MS cared about OS/2).
I know that Internet Explorer has a separate code base and development team, but I don't know of any Microsoft applications where this is true.
I haven't worked there for four years. It is possible that they forked the code base since then, but the reasons they un-forked are stronger than ever so I simply do not believe it.
steveha