That would also require keeping the Declaration of Independence out of this.
Though the way the parent complains about stupidity in the EU, it sounds like he's European. That might explain why he missed or ignored the reference to the Declaration of Independence.
Because the trial judge blew it. The lawyers were pushing the lines as hard as they could, and the trial judge didn't stop them. On the appeal, the lawyers basically said, "Hey, it's the judges job to stop us, and he didn't." So the appeals court overturned the trial verdict, because the trial verdict is invalid if the judge doesn't enforce the rules as to what the jury gets to hear. And the appeals court slammed the lawyers, because they knew what the lines were and stomped all over them anyway, just because the judge let them. The appeals court doesn't think that's acceptable, even if the trial judge lets it slide.
All in all: Good for the appeals court!
Note well: IANAL. This is just my opinion, based on reading the ABA summary.
You have to restrict the monopolist before there is (viable) competition to be harmed, because otherwise there is no chance for there to be competition, ever.
Well, IBM would be somewhat handicapped in their consulting business, having to work with a pile of junk like MSOOXML. I mean, it would probably result in more billable hours, but there are some things that even a consultant won't (willingly) do.
Dueling parallel standards is fine, provided that:
- Both are technically good (clear, easy to implement, do what they're supposed to do), and - Neither are encumbered (patents or other royalty requirements).
If that's the case, then, while one standard might be better, two competing standards is actually just fine.
But I absolutely have to agree with your first statement: We don't want it.
Um, last I checked, Microsoft paid no corporate income tax, despite an income of tens of billions of dollars. There was a big furor of this maybe 5 years ago. It was all legal, thanks to the insane tax laws, but it was obviously not a reasonable outcome. And the situation may have changed in the last 5 years (too lazy to look up Microsoft's SEC filings to check).
Now, what you said is still probably true, because Microsoft still has to pay payroll taxes, property taxes, sales taxes, and so on. But they don't (or at least didn't) pay income tax, even with all their billions.
So before you get all snotty, and start calling people "kids", maybe you ought to check your own facts...
You save something in OOXML today. The standard gets re-written in February. Now Microsoft has a problem. Everybody running Office 2007 is saving in a non-standard-conformant format. What to do?
Windows Update to the rescue! So MS pushes out an update that patches Office. Now it saves in the real format, the one that came out of the February meeting...
But now nobody's saved stuff can be read back in.
But hey, that's all just hypothetical. Microsoft wouldn't be that stupid...
These things line up in space-time as follows: Galaxy 1 is on the line 11 billion years ago, galaxy 2 is on the line 6 billion years ago, galaxy 3 is on the line 3 billion years ago, and the Milky Way is on the line right now.
This does not mean that the reverse is true. It does not mean that there is a line that the Milky Way was on 11 billion years ago, and galaxy 3 was on 8 billion years ago, and galaxy 2 was on 5 billion years ago, and galaxy 3 is on now. Why not? Because galaxies move.
Still, even if not technically correct, it was a really awesome thought by the OP...
What I was trying to say is that, when you're an established company, with an existing codebase and talent that knows how to develop software using language X, language Y has to be a big win before it's worth moving. It doesn't matter whether you should have used Y rather than X in the first place. It doesn't even matter that Y is better. Y has to be a lot better.
Yeah. We moved from assembly, too. And COBOL/Fortran. But we don't move often, and the grandparent explains why. Companies only move to a new language when: - It's enough better that it's a Big Win(TM), or - It's got good enough marketing to persuade some pointy-haired bosses that it will solve all their problems.
I'd guess that companies move more often for the second reason, because when you come right down to it, the first doesn't happen very often...
No, not really. The US will permit its citizens to gamble in casinos in the US, and also in casinos overseas. The US will not permit its citizens to gamble online with sites hosted in the US, and also will not permit it with sites hosted overseas.
So where's the difference between domestic and foreign?
Not so fast. If I've got a contract with you that says "this contract is the totality of the agreement between the two parties", and you come back a decade later and say that you remember it differently, your memory may be legally admissible evidence in some contexts, but it's not admissible in this context (unless the contract is ambiguous - but that's something the judge can rule on as a matter of law).
The admissibility of your evidence goes down further (as if it needed to) if you weren't actually involved with drafting the final version of the contract. Earlier drafts don't count.
As I understand the applicable law (and IANAL), the words on the paper are all you are allowed to look at, unless they are ambiguous. Kimball ruled properly.
Regular polymorphism works like this: I've got a base class B that has a polymorphic function f. I've got derived classes B1 and B2. If I've got an object that is a B, I can call B.f(), and I'll either get B1.f() or B2.f(), depending on whether the object is a B1 or a B2.
But suppose f takes a parameter of type P. And suppose that P is a base class, and I want slightly different versions for f depending on whether the parameter is a P1 or a P2. Then I've got four versions of the f: B1.f(P1), B1.f(P2), B2.f(P1), and B2.f(P2). And I've got a problem. If I've got a B and a P, how do I get the right version of f called?
You can't use function overloading to accomplish this (at least not in C++), because that's done at compile time. It doesn't work if you don't know the types until run time.
In C++, you can do something called "double dispatch". It works like this: You don't call B.f(P). Instead, you call P.g(B), where g is a polymorphic function on P. So if P1's version of g gets called, you know that P is a P1; if P2's version is called, P is a P2. Then g calls B.f(this). Since the type of this is known at compile time, the compiler can distinguish between a call to B.f(P1) and B.f(P2). And polymorphism gets you the correct one of B1.f() and B2.f(). The net result is that the right one of the four possible f() functions is called.
You can start to think about trying to do that with three or four parameters, and you can see why what Larry is talking about - letting the language do it - is a lot better than the way you have to do it in C++.
I posted this on another sub-thread, but I'll repeat myself here.
The judge gets to rule on what is admissable evidence. In fact, he has to, because part of the definition of "admissible evidence" is "evidence that the jury is allowed to see". So the jury can't be the one to determine what is admissible - they'd have to see the evidence to determine that. So it has to be the judge, nobody else can do it.
In the SCO case, the judge threw out a bunch of SCO's BS precisely because it was BS. Second-hand information, for instance, is hearsay, not admissible evidence. Kimball, rightly, threw out the stuff that wasn't actual evidence. Then he ruled against SCO, assuming the most positive (for SCO) interpretation of the admissible evidence.
More precisely, the judge gets to rule on whether the evidence is admissible or not. (Obviously - who else is going to rule on that? The jury? The question is whether the jury even gets to hear it.)
Slashdot Mirror
Citation unavailable until November...
That would also require keeping the Declaration of Independence out of this.
Though the way the parent complains about stupidity in the EU, it sounds like he's European. That might explain why he missed or ignored the reference to the Declaration of Independence.
Because the trial judge blew it. The lawyers were pushing the lines as hard as they could, and the trial judge didn't stop them. On the appeal, the lawyers basically said, "Hey, it's the judges job to stop us, and he didn't." So the appeals court overturned the trial verdict, because the trial verdict is invalid if the judge doesn't enforce the rules as to what the jury gets to hear. And the appeals court slammed the lawyers, because they knew what the lines were and stomped all over them anyway, just because the judge let them. The appeals court doesn't think that's acceptable, even if the trial judge lets it slide.
All in all: Good for the appeals court!
Note well: IANAL. This is just my opinion, based on reading the ABA summary.
*cough*ARP poisoning*cough*
No, IIRC the EU asked Microsoft to document (not submit for standardization) MS's existing file formats, not to create a new, badly-documented one.
You have to restrict the monopolist before there is (viable) competition to be harmed, because otherwise there is no chance for there to be competition, ever.
Well, IBM would be somewhat handicapped in their consulting business, having to work with a pile of junk like MSOOXML. I mean, it would probably result in more billable hours, but there are some things that even a consultant won't (willingly) do.
Dueling parallel standards is fine, provided that:
- Both are technically good (clear, easy to implement, do what they're supposed to do), and
- Neither are encumbered (patents or other royalty requirements).
If that's the case, then, while one standard might be better, two competing standards is actually just fine.
But I absolutely have to agree with your first statement: We don't want it.
Now, what you said is still probably true, because Microsoft still has to pay payroll taxes, property taxes, sales taxes, and so on. But they don't (or at least didn't) pay income tax, even with all their billions.
So before you get all snotty, and start calling people "kids", maybe you ought to check your own facts...
You save something in OOXML today. The standard gets re-written in February. Now Microsoft has a problem. Everybody running Office 2007 is saving in a non-standard-conformant format. What to do?
Windows Update to the rescue! So MS pushes out an update that patches Office. Now it saves in the real format, the one that came out of the February meeting...
But now nobody's saved stuff can be read back in.
But hey, that's all just hypothetical. Microsoft wouldn't be that stupid...
Would they?
But even having enough money to go somewhere else doesn't help if there's nowhere else that's free.
Don't. Even though technically wrong, it was still a really cool, mind-blowing thought.
These things line up in space-time as follows: Galaxy 1 is on the line 11 billion years ago, galaxy 2 is on the line 6 billion years ago, galaxy 3 is on the line 3 billion years ago, and the Milky Way is on the line right now.
This does not mean that the reverse is true. It does not mean that there is a line that the Milky Way was on 11 billion years ago, and galaxy 3 was on 8 billion years ago, and galaxy 2 was on 5 billion years ago, and galaxy 3 is on now. Why not? Because galaxies move.
Still, even if not technically correct, it was a really awesome thought by the OP...
What I was trying to say is that, when you're an established company, with an existing codebase and talent that knows how to develop software using language X, language Y has to be a big win before it's worth moving. It doesn't matter whether you should have used Y rather than X in the first place. It doesn't even matter that Y is better. Y has to be a lot better.
Yeah. We moved from assembly, too. And COBOL/Fortran. But we don't move often, and the grandparent explains why. Companies only move to a new language when:
- It's enough better that it's a Big Win(TM), or
- It's got good enough marketing to persuade some pointy-haired bosses that it will solve all their problems.
I'd guess that companies move more often for the second reason, because when you come right down to it, the first doesn't happen very often...
His pointing out that Rails required restarting every 4 minutes was both technical and interesting - shocking, in fact.
I can answer part of that. Vista might be showing consistent growth in market share all year because it started from zero, maybe?
No, not really. The US will permit its citizens to gamble in casinos in the US, and also in casinos overseas. The US will not permit its citizens to gamble online with sites hosted in the US, and also will not permit it with sites hosted overseas.
So where's the difference between domestic and foreign?
Not so fast. If I've got a contract with you that says "this contract is the totality of the agreement between the two parties", and you come back a decade later and say that you remember it differently, your memory may be legally admissible evidence in some contexts, but it's not admissible in this context (unless the contract is ambiguous - but that's something the judge can rule on as a matter of law).
The admissibility of your evidence goes down further (as if it needed to) if you weren't actually involved with drafting the final version of the contract. Earlier drafts don't count.
As I understand the applicable law (and IANAL), the words on the paper are all you are allowed to look at, unless they are ambiguous. Kimball ruled properly.
Here's my understanding. (Your mileage may vary.)
Regular polymorphism works like this: I've got a base class B that has a polymorphic function f. I've got derived classes B1 and B2. If I've got an object that is a B, I can call B.f(), and I'll either get B1.f() or B2.f(), depending on whether the object is a B1 or a B2.
But suppose f takes a parameter of type P. And suppose that P is a base class, and I want slightly different versions for f depending on whether the parameter is a P1 or a P2. Then I've got four versions of the f: B1.f(P1), B1.f(P2), B2.f(P1), and B2.f(P2). And I've got a problem. If I've got a B and a P, how do I get the right version of f called?
You can't use function overloading to accomplish this (at least not in C++), because that's done at compile time. It doesn't work if you don't know the types until run time.
In C++, you can do something called "double dispatch". It works like this: You don't call B.f(P). Instead, you call P.g(B), where g is a polymorphic function on P. So if P1's version of g gets called, you know that P is a P1; if P2's version is called, P is a P2. Then g calls B.f(this). Since the type of this is known at compile time, the compiler can distinguish between a call to B.f(P1) and B.f(P2). And polymorphism gets you the correct one of B1.f() and B2.f(). The net result is that the right one of the four possible f() functions is called.
You can start to think about trying to do that with three or four parameters, and you can see why what Larry is talking about - letting the language do it - is a lot better than the way you have to do it in C++.
I posted this on another sub-thread, but I'll repeat myself here.
The judge gets to rule on what is admissable evidence. In fact, he has to, because part of the definition of "admissible evidence" is "evidence that the jury is allowed to see". So the jury can't be the one to determine what is admissible - they'd have to see the evidence to determine that. So it has to be the judge, nobody else can do it.
In the SCO case, the judge threw out a bunch of SCO's BS precisely because it was BS. Second-hand information, for instance, is hearsay, not admissible evidence. Kimball, rightly, threw out the stuff that wasn't actual evidence. Then he ruled against SCO, assuming the most positive (for SCO) interpretation of the admissible evidence.
More precisely, the judge gets to rule on whether the evidence is admissible or not. (Obviously - who else is going to rule on that? The jury? The question is whether the jury even gets to hear it.)
At a minimum, they told the bankruptcy court that they thought Kimball's decision against them was wrong...
Sure. And Newton's law is an old physics theory.
Not to mention other bookstores. What makes one bookstore own the ISBN number? Can they sue other bookstores for having that number in their store?
This is so insanely bogus, words fail me.