Doesn't really make me suspicious. For one, the Linux number may be inflated because of the way Linux treats threads as procs and some of the tools used to gauge memory usage don't know how to tell the difference
So, you go around claiming that switching from Linux to Solaris reduces memory usage from 200MB to 50MB, and then you admit that you can't even read the VM stats properly. That's pretty dishonest.
They do use Solaris and I believe Java for parts related to their billing for AdWords or something like that.
What difference does it make? That's not where Google spends their money. You claimed that OpenSolaris brings big performance gains, gains that would save Google hundreds of millions of dollars a year if they were true, yet Google sticks with Linux for their compute platform. Since Google isn't stupid, that's just another indication that performance claims for Solaris are bogus and that it does not perform significantly better than Linux.
NFS is widely used,
You know, if you don't understand what a piece of shit (pre-v4) NFS really is, you really have no basis on which to evaluate Sun's engineering competency. Sun marketing and engineering were touting the many advantages of the NFS design. In practice, NFS managed to be unsafe, slow, complex, hard to manage, and unreliable all at the same time. Sun's own implementation corrupted data for years. NFS probably has corrupted and destroyed more data than any other piece of software in existence. The fact that Sun managed to establish it as a de-facto standard has damaged UNIX more than just about anything else.
As for the kernel, linux developers concentrate on what matters to them and their users and sun developers on what they feel is important for theirs.
Those criteria are important to everybody. Unfortunately, Solaris is not demonstrably superior in those areas, it merely has a lot of code and ideas that claim to help without anybody actually having demonstrated that they do (in fact, software like ZFS is probably harmful).
You say that you buy Sun because you have confidence in their software developers. I don't, based on 20 years experience with the company and their software.
To make it clear: my objection is to people claiming that criticism of Win32 is in some sense "invalid" because I "should" be using.NET.
Well, and your objection is bogus. The fact that something gets added as a Win32 API means that it is accessible from C/C++, whereas if it got added as a.NET API, it wouldn't be. But the fact that something gets added as a Win32 API doesn't say anything about the quality of its binding to C#. In fact, Win32 APIs bound to C# generally end up with managed interfaces.
AFAIK, Apple has never pretended that Cocoa is complete; that you use e.g. BSD sockets is promoted as a virtue, not a flaw.
But in the case of Apple, it is a flaw; in contrast to C#/.NET, where even the Win32 access can be managed, any BSD or other C interface used from Objective-C is intrinsically unsafe and unmanaged. Apple continues sinking into the tarpit of unmanaged code. Objective-C simply doesn't separate high level and low level code enough.
If I "should" be using.NET then MS has got to stop adding features to Win32 and make.NET a complete replacement. Until that happens, criticism of Win32 is fair game.
Microsoft seems to be doing the right thing: they add Win32 APIs for things that need to be callable from both C/C++ and C#. Nothing would be gained by making those APIs.NET-only. And the fact that there are parts of the Win32 API that truly suck badly doesn't need to bother you, because those parts already have better.NET equivalents; you never need to use them.
Of course, I think you should be using neither.NET or Cocoa. But what I said also applies on Linux: C#+Gtk# and Python+PyGtk are much better combos than Objective-C+Cocoa.
For me, the reason I like Solaris is that I've found I get less memory usage and better performance out of Apache/Tomcat. One webapp that was running well over 200MB on RedHat now runs in less then 50MB of memory on the Solaris setup under load.
And that doesn't make you the least bit suspicious? Are you naive enough to think that the use of Solaris gives you a fourfold reduction in memory usage in your web server because... it has a better kernel? Don't you think Google would switch over all their machines to OpenSolaris in a jiffy if they could shave off a few percent in memory and CPU usage, let alone a factor of 4?
I'm sorry, but the most plausible conclusion is that you simply can't read your VM stats or have some configuration difference in your web server.
I don't like Solaris because it has DTrace, ZFS, Crossbow, Zones, etc. I like it because it's developed by people that can come up with those technologies [...] Anyway, we're not going to agree on this but as an aside, I never understood the animosity Linux proponents had towards Solaris.
Probably because many of them were long-term Sun users, like myself. Until the mid-1990's, Sun workstations gave good cost/performance ratios and were fairly inexpensive, and that's why people bought them.
But Sun software always sucked: NFS, SunView, NeWS, the SunOS kernel, and Java were badly designed and had buggy implementations, all the while Sun marketing touted them as the best thing since sliced bread and Sun engineers went around writing academic papers left and right. Please, no more crappy Sun software.
Architecturally, the Linux kernel isn't going to win any prizes. But at least the Linux developers concentrate on stuff that matters, as opposed to Sun engineers, who are trying to set themselves yet another monument in software.
Nope. You err because you take "price" as being built upon "money", and worse, "money" as being only that is issued by a central bank. Neither of these two equalities is true.
I did no such thing. I just pointed out that the definition of free market only refers to the ability to set prices freely, nothing else. The government could require you to do all your contracts by wearing lederhosen and yodeling in Bavarian and it would still be a free market.
In both case, no government gets involved.
Even if your example were valid, so what? I'm not saying that the definition of "free markets" excludes the ability of having transactions without government interference, I'm saying that it doesn't require absence of government interference, except in a narrow and specific sense.
But your example is wrong anyway. Barter in the US has a monetary value and is taxable. Furthermore, the fact that people don't just hold you up at gunpoint and take your goods, and that the goods they barter with you generally don't kill you, is a result of government restrictions.
Taking from your list, I'd say that I believe a government must exist for enforcing contracts and protecting property, but not for the other aspects.
Sure, governments don't need to do any of the other things, and personally I'd also prefer if they stopped doing some of them. But when they do engage in them, it's still a free market according to the standard definition. You can argue against these other governmental activities, but you can't argue against them with free market arguments.
I dunno, increasing the percentage of solved violent crimes by 3% might well be considered "working" by many, in particular if they don't perceive any downsides.
A free market economy is one where scarcities are resolved through changes in relative prices rather than through regulation.
"Free market" refers only to the fact that people set prices freely. Free markets wouldn't even exist without extensive government interference: governments enforce contracts, governments create the infrastructure where economic activity takes place, governments issue money, and governments protect property. And governments create a free market in intellectual works through copyright (whether that's a good idea is another debate, unrelated to free market economics).
What you want is libertarianism or anarcho-capitalism, neither of which even necessarily results in a free market economy. You just like to call your position a "free market" position because that sounds so much less offensive than the actual name for what you espouse.
And I say that because I am a free market evangelist. For us, government must have no say on how the market works,
Just because you fancy yourself a "free market evangelist" doesn't mean you are one.
A proper free market evangelist, on the other hand, always places freedom in front, and would contend that even if central planning was in some sense "better" (whatever that means), it would still be a source of less freedom, thus something he'd be opposed to.
You're confusing free market economics with libertarianism. You're an economically naive libertarian, not an advocate of free markets.
All that expensive furniture and designer layout is going to be money that's not going to be available for growth. Yes, that affects the value of your stock options if you happen to work for a startup, possibly big time. Fancy offices are for big companies whose stock isn't going anywhere anymore.
Spend bucks on things that make people productive: fast machines and big screens. Spend bucks on things that let people have fun. For anything else, go to the surplus store and buy functional and sturdy.
That's because they factor the decrease in hardware cost over the lifetime of the product into their profit calculations. In different words, they had to move to the cheaper hardware for the mass market product.
The GL, having smaller sales volume, ended up being more expensive. It's still a good deal for what it is.
Sorry, but this isn't some evil master plan, it's simple economics and product life cycle.
RAW mode is very useful on small, noisy sensors. That's because RAW software running on a desktop can do a much better job doing the conversion than the in-camera converter.
Second, strict copyright enforcement is neither republican nor democrat, liberal nor conservative. It's an artificial control of the market, and as such it's bad according to the free market evangelists.
Quite wrong. To free market evangelists, copyright is intellectual property and strict copyright enforcement protects private property. In different words, strict copyright enforcement is what makes a free market in intellectual property possible. That is quite consistent with their beliefs. Strict copyright enforcement does not interfere with the market itself (e.g., Disney still can screw consumers any way they want).
The debate is about whether we want to have intellectual property in the first place.
You have to trade that off against the costs: licensing, porting, incompatibilities, potential bugs. For most people, the tradeoff never made sense.
And it's the same with Solaris, DTrace and ZFS: no to small benefit for most people, some design problems, big switching costs, and big long-term risks.
Look, if the argument is going to be "Don't use Win32, use.NET" then.NET has to do all the things that Win32 can do. So which is it?
Many people write C# programs without ever calling Win32 directly. And Win32 API calls can be safely wrapped up in C# classes. In contrast, in Objective-C, you have no choice at all but to call C code. In fact, even Cocoa and Objective-C APIs are unmanaged and unsafe.
I don't like Win32 or.NET, but to claim that Cocoa and Objective-C are better is ridiculous. Cocoa and Objective-C are obsolete. C# and.NET may be bloated, but at least they are on the right path.
Compare this to the current state of.NET, where developers have to constantly mix in win32 calls to do anything but the most basic applications
Are you joking? Objective-C is unmanaged code that mixes C and Smalltalk code everywhere. Even if.NET programmers occasionally call out to C (and they rarely do), that's negligible compared to the unsafe mess that Cocoa and Objective-C programs are.
Apple IS abandoning Carbon.
No, they are merely not making a 64 bit version of it. Furthermore, whatever Apple calls it, there will continue to be huge numbers of C-based APIs. And even if there weren't, except for the method call syntax, Cocoa APIs are just as bad as C based APIs.
But you can't. MS has added new features to Win32 (and VC++/MFC), and MS is going to continue to do so in Windows Seven.
So? Apple keeps adding new features to its C-based APIs as well.
But having to drop into Win32 to call new features that have only just been added?
How's that different from Cocoa? Cocoa programmers constantly have to call plain C APIs and Apple keeps adding them. At least.NET allows unsafe C code to be wrapped safely once. In Cocoa, there is simply no getting away from unsafe C code..NET is not my favorite environment, but it is still technically far ahead of Cocoa and Objective-C.
Apart from the just plain wrong statistical reasoning, the experiment was done under uncontrolled conditions over the Internet. The sample table in the article actually had lines separating the columns and rows. Geez, with that, it's not surprising that the author finds no differences!
Zebra striping may or may not help significantly, but this paper won't tell you either way.
Failure to demonstrate a statistically significant improvement in accuracy is not the same as absence of such an improvement. For example, you can fail to demonstrate a significant improvement simply by making the sample size small enough or making the measurements noisier. In fact, I can pretty much guarantee that if you make the sample large enough, you will find a statistically significant difference; whether there is a statistically significant difference simply isn't the right question to ask. The right question to ask is that, once you have enough samples to demonstrate a statistically significant difference and measure it, whether it is practically important.
In different words, if you can't demonstrate a statistically significant result, you don't have a negative result, but you basically hae no result at all and should consider your paper unpublishable.
Altogether, the paper only shows one thing conclusively: that its author and the OzCHI reviewers are statistically illiterate.
Yes, and they essentially copied it from standard Sun marketing language.
Were you even around to use sun workstations "when they were actually popular"?
Yes. But if you actually think that Sun software is any good, you apparently weren't. Sun succeeded because they were the first to deliver affordable 32bit workstations, and they succeeded despite their operating systems, not because of them.
As for using GCC, that was usually because Solaris didn't come with the Sun compiler tools. You had to buy those seperately. And at the time, and still to this day, they tend to compile code that performs better than GCC.
In fact I will have to wait an see what an MS Office looks like under it before I am entirely convinced that it is prime time yet.
Neither you nor I are going to be writing MS Office and that's simply not a good criterion. The measure of a toolkit is whether it makes implementing the day-to-day simple GUIs and software people need to implement simple. Note that MS Office is not written in Cocoa either.
I also think that Objective-C is a better OO C than c++.
That's true, but a pretty low standard. In the end, Objective-C is still a C-derived language with all the problems that that entails. Objective-C was a poor man's Smalltalk, a compromise that made sense 20 years ago, but doesn't make sense anymore.
By the sounds of it.NET is your.
No, not really. If I had to choose, I'd prefer writing.NET over Cocoa (I tried both). Most of my stuff, I do in Python and Gtk+, which is far more pleasant than either.NET or Cocoa/ObjC programming.
Since Apple is stuck with Cocoa's API syntax, their best bet would be to buy themselves a good Smalltalk implementation and move mainstream development over to that, with the option of dropping into Objective-C when necessary.
And how is this different from Carbon/Cocoa? Carbon/Cocoa libraries also don't check for the validity of memory passed to them, also have weird undocumented edge cases, etc.
In contrast to Carbon/Cocoa,.NET actually fixes this. Wouldn't it be nice if Apple also had a managed runtime and libraries...
Again, I totally agree. The.NET Framework class libraries and the C# language are the definition of "The Right Way" to organize and structure code. They are technically sophisticated, architecturally beautiful, and make use of all of the best ideas at the culmination of nearly three (3) decades of object oriented programming theory and software engineering.
That's not necessarily a good thing, however: much of OOP and software engineering is ivory tower ideas that often don't work out in the real world. Look at the decades of failed S/E ideas about how to manage project and ensure correctness.
I think, on balance,.NET is still better than Cocoa. But all that OOP theory and software engineering that has gone into.NET has made it a worse platform than it could be, and I can see why some people prefer Cocoa.
If you can't even read his argument without getting mixed up, then your future is in trouble. (OTOH you have a bright future as a Slashdot poster. Not reading TFA is henceforth passé; reading it and getting it entirely wrong is the new standard.)
No, he is saying that the argument is confused because it mixes up Win32 and.NET. That's a problem with TFA.
If you want to compare Windows and OS X as modern platforms, you need to compare the modern APIs, that is,.NET against Cocoa, nothing else.
Slashdot Mirror
Doesn't really make me suspicious. For one, the Linux number may be inflated because of the way Linux treats threads as procs and some of the tools used to gauge memory usage don't know how to tell the difference
So, you go around claiming that switching from Linux to Solaris reduces memory usage from 200MB to 50MB, and then you admit that you can't even read the VM stats properly. That's pretty dishonest.
They do use Solaris and I believe Java for parts related to their billing for AdWords or something like that.
What difference does it make? That's not where Google spends their money. You claimed that OpenSolaris brings big performance gains, gains that would save Google hundreds of millions of dollars a year if they were true, yet Google sticks with Linux for their compute platform. Since Google isn't stupid, that's just another indication that performance claims for Solaris are bogus and that it does not perform significantly better than Linux.
NFS is widely used,
You know, if you don't understand what a piece of shit (pre-v4) NFS really is, you really have no basis on which to evaluate Sun's engineering competency. Sun marketing and engineering were touting the many advantages of the NFS design. In practice, NFS managed to be unsafe, slow, complex, hard to manage, and unreliable all at the same time. Sun's own implementation corrupted data for years. NFS probably has corrupted and destroyed more data than any other piece of software in existence. The fact that Sun managed to establish it as a de-facto standard has damaged UNIX more than just about anything else.
As for the kernel, linux developers concentrate on what matters to them and their users and sun developers on what they feel is important for theirs.
Those criteria are important to everybody. Unfortunately, Solaris is not demonstrably superior in those areas, it merely has a lot of code and ideas that claim to help without anybody actually having demonstrated that they do (in fact, software like ZFS is probably harmful).
You say that you buy Sun because you have confidence in their software developers. I don't, based on 20 years experience with the company and their software.
To make it clear: my objection is to people claiming that criticism of Win32 is in some sense "invalid" because I "should" be using .NET.
.NET API, it wouldn't be. But the fact that something gets added as a Win32 API doesn't say anything about the quality of its binding to C#. In fact, Win32 APIs bound to C# generally end up with managed interfaces.
.NET then MS has got to stop adding features to Win32 and make .NET a complete replacement. Until that happens, criticism of Win32 is fair game.
.NET-only. And the fact that there are parts of the Win32 API that truly suck badly doesn't need to bother you, because those parts already have better .NET equivalents; you never need to use them.
.NET or Cocoa. But what I said also applies on Linux: C#+Gtk# and Python+PyGtk are much better combos than Objective-C+Cocoa.
Well, and your objection is bogus. The fact that something gets added as a Win32 API means that it is accessible from C/C++, whereas if it got added as a
AFAIK, Apple has never pretended that Cocoa is complete; that you use e.g. BSD sockets is promoted as a virtue, not a flaw.
But in the case of Apple, it is a flaw; in contrast to C#/.NET, where even the Win32 access can be managed, any BSD or other C interface used from Objective-C is intrinsically unsafe and unmanaged. Apple continues sinking into the tarpit of unmanaged code. Objective-C simply doesn't separate high level and low level code enough.
If I "should" be using
Microsoft seems to be doing the right thing: they add Win32 APIs for things that need to be callable from both C/C++ and C#. Nothing would be gained by making those APIs
Of course, I think you should be using neither
For me, the reason I like Solaris is that I've found I get less memory usage and better performance out of Apache/Tomcat. One webapp that was running well over 200MB on RedHat now runs in less then 50MB of memory on the Solaris setup under load.
And that doesn't make you the least bit suspicious? Are you naive enough to think that the use of Solaris gives you a fourfold reduction in memory usage in your web server because... it has a better kernel? Don't you think Google would switch over all their machines to OpenSolaris in a jiffy if they could shave off a few percent in memory and CPU usage, let alone a factor of 4?
I'm sorry, but the most plausible conclusion is that you simply can't read your VM stats or have some configuration difference in your web server.
I don't like Solaris because it has DTrace, ZFS, Crossbow, Zones, etc. I like it because it's developed by people that can come up with those technologies [...] Anyway, we're not going to agree on this but as an aside, I never understood the animosity Linux proponents had towards Solaris.
Probably because many of them were long-term Sun users, like myself. Until the mid-1990's, Sun workstations gave good cost/performance ratios and were fairly inexpensive, and that's why people bought them.
But Sun software always sucked: NFS, SunView, NeWS, the SunOS kernel, and Java were badly designed and had buggy implementations, all the while Sun marketing touted them as the best thing since sliced bread and Sun engineers went around writing academic papers left and right. Please, no more crappy Sun software.
Architecturally, the Linux kernel isn't going to win any prizes. But at least the Linux developers concentrate on stuff that matters, as opposed to Sun engineers, who are trying to set themselves yet another monument in software.
Nope. You err because you take "price" as being built upon "money", and worse, "money" as being only that is issued by a central bank. Neither of these two equalities is true.
I did no such thing. I just pointed out that the definition of free market only refers to the ability to set prices freely, nothing else. The government could require you to do all your contracts by wearing lederhosen and yodeling in Bavarian and it would still be a free market.
In both case, no government gets involved.
Even if your example were valid, so what? I'm not saying that the definition of "free markets" excludes the ability of having transactions without government interference, I'm saying that it doesn't require absence of government interference, except in a narrow and specific sense.
But your example is wrong anyway. Barter in the US has a monetary value and is taxable. Furthermore, the fact that people don't just hold you up at gunpoint and take your goods, and that the goods they barter with you generally don't kill you, is a result of government restrictions.
Taking from your list, I'd say that I believe a government must exist for enforcing contracts and protecting property, but not for the other aspects.
Sure, governments don't need to do any of the other things, and personally I'd also prefer if they stopped doing some of them. But when they do engage in them, it's still a free market according to the standard definition. You can argue against these other governmental activities, but you can't argue against them with free market arguments.
I dunno, increasing the percentage of solved violent crimes by 3% might well be considered "working" by many, in particular if they don't perceive any downsides.
You know, you can invent your own definitions for everything, but that doesn't make your definitions valid.
Here are actual definitions (emphasis mine):
http://en.wikipedia.org/wiki/Free_market
http://canadianeconomy.gc.ca/english/economy/freemarket.html
"Free market" refers only to the fact that people set prices freely. Free markets wouldn't even exist without extensive government interference: governments enforce contracts, governments create the infrastructure where economic activity takes place, governments issue money, and governments protect property. And governments create a free market in intellectual works through copyright (whether that's a good idea is another debate, unrelated to free market economics).
What you want is libertarianism or anarcho-capitalism, neither of which even necessarily results in a free market economy. You just like to call your position a "free market" position because that sounds so much less offensive than the actual name for what you espouse.
So you're claiming that Linksys took a loss on the WRT54G versions 1 through 4?
No.
I usually try to be polite here on Slashdot, but that's just flat-out stupid.
Yup, you're right: your statement and conclusion are incredibly stupid.
And I say that because I am a free market evangelist. For us, government must have no say on how the market works,
Just because you fancy yourself a "free market evangelist" doesn't mean you are one.
A proper free market evangelist, on the other hand, always places freedom in front, and would contend that even if central planning was in some sense "better" (whatever that means), it would still be a source of less freedom, thus something he'd be opposed to.
You're confusing free market economics with libertarianism. You're an economically naive libertarian, not an advocate of free markets.
All that expensive furniture and designer layout is going to be money that's not going to be available for growth. Yes, that affects the value of your stock options if you happen to work for a startup, possibly big time. Fancy offices are for big companies whose stock isn't going anywhere anymore.
Spend bucks on things that make people productive: fast machines and big screens. Spend bucks on things that let people have fun. For anything else, go to the surplus store and buy functional and sturdy.
That's because they factor the decrease in hardware cost over the lifetime of the product into their profit calculations. In different words, they had to move to the cheaper hardware for the mass market product.
The GL, having smaller sales volume, ended up being more expensive. It's still a good deal for what it is.
Sorry, but this isn't some evil master plan, it's simple economics and product life cycle.
RAW mode is very useful on small, noisy sensors. That's because RAW software running on a desktop can do a much better job doing the conversion than the in-camera converter.
Second, strict copyright enforcement is neither republican nor democrat, liberal nor conservative. It's an artificial control of the market, and as such it's bad according to the free market evangelists.
Quite wrong. To free market evangelists, copyright is intellectual property and strict copyright enforcement protects private property. In different words, strict copyright enforcement is what makes a free market in intellectual property possible. That is quite consistent with their beliefs. Strict copyright enforcement does not interfere with the market itself (e.g., Disney still can screw consumers any way they want).
The debate is about whether we want to have intellectual property in the first place.
Running code faster is not a good thing?
You have to trade that off against the costs: licensing, porting, incompatibilities, potential bugs. For most people, the tradeoff never made sense.
And it's the same with Solaris, DTrace and ZFS: no to small benefit for most people, some design problems, big switching costs, and big long-term risks.
I get it, you don't manage a lot of servers, you're not a developer
Another mistake on your part.
you say a lot of stuff with out backing it up
I'm an amateur compared to Sun shills like you, or Sun marketing.
Look, if the argument is going to be "Don't use Win32, use .NET" then .NET has to do all the things that Win32 can do. So which is it?
.NET, but to claim that Cocoa and Objective-C are better is ridiculous. Cocoa and Objective-C are obsolete. C# and .NET may be bloated, but at least they are on the right path.
Many people write C# programs without ever calling Win32 directly. And Win32 API calls can be safely wrapped up in C# classes. In contrast, in Objective-C, you have no choice at all but to call C code. In fact, even Cocoa and Objective-C APIs are unmanaged and unsafe.
I don't like Win32 or
Compare this to the current state of .NET, where developers have to constantly mix in win32 calls to do anything but the most basic applications
.NET programmers occasionally call out to C (and they rarely do), that's negligible compared to the unsafe mess that Cocoa and Objective-C programs are.
Are you joking? Objective-C is unmanaged code that mixes C and Smalltalk code everywhere. Even if
Apple IS abandoning Carbon.
No, they are merely not making a 64 bit version of it. Furthermore, whatever Apple calls it, there will continue to be huge numbers of C-based APIs. And even if there weren't, except for the method call syntax, Cocoa APIs are just as bad as C based APIs.
But you can't. MS has added new features to Win32 (and VC++/MFC), and MS is going to continue to do so in Windows Seven.
.NET allows unsafe C code to be wrapped safely once. In Cocoa, there is simply no getting away from unsafe C code. .NET is not my favorite environment, but it is still technically far ahead of Cocoa and Objective-C.
So? Apple keeps adding new features to its C-based APIs as well.
But having to drop into Win32 to call new features that have only just been added?
How's that different from Cocoa? Cocoa programmers constantly have to call plain C APIs and Apple keeps adding them. At least
Apart from the just plain wrong statistical reasoning, the experiment was done under uncontrolled conditions over the Internet. The sample table in the article actually had lines separating the columns and rows. Geez, with that, it's not surprising that the author finds no differences!
Zebra striping may or may not help significantly, but this paper won't tell you either way.
You were using tools and languages designed for those kinds of screens.
Failure to demonstrate a statistically significant improvement in accuracy is not the same as absence of such an improvement. For example, you can fail to demonstrate a significant improvement simply by making the sample size small enough or making the measurements noisier. In fact, I can pretty much guarantee that if you make the sample large enough, you will find a statistically significant difference; whether there is a statistically significant difference simply isn't the right question to ask. The right question to ask is that, once you have enough samples to demonstrate a statistically significant difference and measure it, whether it is practically important.
In different words, if you can't demonstrate a statistically significant result, you don't have a negative result, but you basically hae no result at all and should consider your paper unpublishable.
Altogether, the paper only shows one thing conclusively: that its author and the OzCHI reviewers are statistically illiterate.
That's from the zdnet blog I posted a link from.
Yes, and they essentially copied it from standard Sun marketing language.
Were you even around to use sun workstations "when they were actually popular"?
Yes. But if you actually think that Sun software is any good, you apparently weren't. Sun succeeded because they were the first to deliver affordable 32bit workstations, and they succeeded despite their operating systems, not because of them.
As for using GCC, that was usually because Solaris didn't come with the Sun compiler tools. You had to buy those seperately. And at the time, and still to this day, they tend to compile code that performs better than GCC.
You say that as if it's a good thing.
Have you ever used Cocoa?
.NET is your.
.NET over Cocoa (I tried both). Most of my stuff, I do in Python and Gtk+, which is far more pleasant than either .NET or Cocoa/ObjC programming.
Yes.
In fact I will have to wait an see what an MS Office looks like under it before I am entirely convinced that it is prime time yet.
Neither you nor I are going to be writing MS Office and that's simply not a good criterion. The measure of a toolkit is whether it makes implementing the day-to-day simple GUIs and software people need to implement simple. Note that MS Office is not written in Cocoa either.
I also think that Objective-C is a better OO C than c++.
That's true, but a pretty low standard. In the end, Objective-C is still a C-derived language with all the problems that that entails. Objective-C was a poor man's Smalltalk, a compromise that made sense 20 years ago, but doesn't make sense anymore.
By the sounds of it
No, not really. If I had to choose, I'd prefer writing
Since Apple is stuck with Cocoa's API syntax, their best bet would be to buy themselves a good Smalltalk implementation and move mainstream development over to that, with the option of dropping into Objective-C when necessary.
And how is this different from Carbon/Cocoa? Carbon/Cocoa libraries also don't check for the validity of memory passed to them, also have weird undocumented edge cases, etc.
.NET actually fixes this. Wouldn't it be nice if Apple also had a managed runtime and libraries...
In contrast to Carbon/Cocoa,
Again, I totally agree. The .NET Framework class libraries and the C# language are the definition of "The Right Way" to organize and structure code. They are technically sophisticated, architecturally beautiful, and make use of all of the best ideas at the culmination of nearly three (3) decades of object oriented programming theory and software engineering.
.NET is still better than Cocoa. But all that OOP theory and software engineering that has gone into .NET has made it a worse platform than it could be, and I can see why some people prefer Cocoa.
That's not necessarily a good thing, however: much of OOP and software engineering is ivory tower ideas that often don't work out in the real world. Look at the decades of failed S/E ideas about how to manage project and ensure correctness.
I think, on balance,
If you can't even read his argument without getting mixed up, then your future is in trouble. (OTOH you have a bright future as a Slashdot poster. Not reading TFA is henceforth passé; reading it and getting it entirely wrong is the new standard.)
.NET. That's a problem with TFA.
.NET against Cocoa, nothing else.
No, he is saying that the argument is confused because it mixes up Win32 and
If you want to compare Windows and OS X as modern platforms, you need to compare the modern APIs, that is,