I didn't do very comprehensive tests unfortunately. I did enough to satisfy myself;).
Tail calls are supported by the.NET CLR. I don't think the C# compiler currently emits tail calls (I could be wrong).
There are other things like asynchronous I/O, asynchronoos method calls,generational garbage collection and value types (stack based objects) that also help C# get a speed advantage over Java.
Speed was definitely a major concern for Microsoft. Much more so than it was for Sun.
Microsoft also obviously has the skill to write very fast VMs from their experience with Java.
Last I looked you couldn't 'draw' a form, double click to fill in event code like you would in VB and have VAJ generate JSP pages that would dynamically generate HTML/JavaScript representations of the form.
Also, does VAJ have edit and continue? Does VAC++ have edit and continue? I genuinely curious. Also does the latest version support intellisense? The one I tried last year didn't:|.
Windows Forms certainly uses windows primitives (like fonts) whereas Java seems to do those completely by itself (it just seems to need to be able to create windows and draw pixels...mostly).
If you mean, subjectively, your C# programs run as fast as your C++ programs, that's probably because C# relies on a lot of native code in its libraries
Um no. The.NET framework classes are all written in C#.
Some parts call on native code using P/Invoke (which is a heck of a lot better than JNI - which requires rewriting of existing libraries).
These parts are specialised like ADO.NET, ASP.NET and Windows Forms. But we aren't comparing the speed of these are we?
Well, so far, I haven't seen any such "great solutions" on Windows--just a lot of really mediocre software that it took companies a long time to develop and debug.
Well maybe you should look harder. Show me a better IDE than Visual Studio (in terms of debugging, edit and continue etc). Show me Linux's equivalent of Microsoft Access and show me how it is better. Tell the millions of people relying on windows around the world that they're wrong.
I use windows. I develop applications for windows (many of which are free).
I also develop Unix applications. I'm getting my degree from a Unix only university computer science department.
I prefer to use windows. I think the APIs are tidier, and better designed than in unix.
exhibit a:
CreateThread vs pthread_create
WaitForSingleObject vs pthread_wait, sem_wait, wait etc etc.
Tell me that I'm just stupid and don't know any better.
Is it a good design decision to automate this? That is an open question. If you automate boxing/unboxing, most programmers won't know it's happening, and they will wonder while their code runs so slowly. It also leads to unobvious behavior when people try to use inheritance with number classes.
Um, integers won't be boxed until they're used as objects (put in a collection or have a method called). So it won't be any slower than when you have to manually box in java.
Also, you can't inherit from number classed (they're sealed). In Java you can't inherit from java.lang.Integer, java.lang.Double etc cause they're sealed (finalised anyway) as well.
The greatest trick Bill Gates ever pulled, they will say, was that he made windows desktop software so ridiculously difficult to install, use, and maintain (via the windows registry).
I'm sorry, but I have to say the windows registry still remains one of the best things about windows. Don't give me that "it corrupts" shit. That was years ago. It's so much better to have a centralised storage database for settings.
You're telling me that the registry (structered settings database) is MORE complicated than the gazzilions of configuration files in Linux (all of which have different formats etc)?
Aparently you're required to connect to their servers everytime you want to view a webpage.
That's soooo wrong.
Things I've learnt from taco today:
1) Microsoft is evil.
2) Microsoft Passport is NOT some sort of unified authentication service. It aparently isn't like the orignal authentication services Microsoft used to use on their websites. It's just pure evil and the work of the devil.
3) You can't develop under windows unless you subscripe to passport:
4) Microsoft is evil.
5) Microsoft is evil.
Come on Taco, take off that second hand evangelical suit you're always wearing. It's got ketchup (and god knows what else) stains on it.
How is Passport different from the original authentication scheme they used for online MSDN subscriber downloads (except that it's the same system used for hotmail, windowsbeta, messenger etc).
In order for me to do lots of things on sun's java website, I had to send my details to them. Hang on, it's not called "Sun Passport". It must be ok.
When Microsoft's routers went down then it can't be their routers (routers don't fail), it must have been their DNS servers running Windows. Ofcourse they needed to go and get some BSD CDs from some friendly geeks and install BSD on their DNS servers in order to get their sites back up. Right.....
If Slashdot's routers go down then it's definitely their routers - no doubt about it.
It's different. The -turbo option basically just hides the window, and leaves the process still running. I can close all my IE windows (leaving no iexplorer.exe processes left) and IE will load much faster than mozilla. This is with "start in new process" turned on.
Funny, last I checked MS still hadn't bought a copy of my web site.
But it has nothing to do with Microsoft buying your site. It is the USER who is doing the highlighting by turning the option on. Microsoft simply supply the highlighter for the user to use (if they choose to).
Are you saying I can't view your website, print it out and then highlight it?
I expect Microsoft will be forced to shift from the exclusionary tag model to an inclusionary tag model where only sites with an inclusionary tag can be modified in this way. That way content owners have to give their eplicit permission to microsoft to edit their page in ways they would be completely unaware of.
*sigh* All these posts just prove that most drones here don't even know what smart tags is. Microsoft need not get exclusive rights from content owners to edit their page in ways they are completely unaware of. The pages aren't being edited. Certain keywords on the page are being hyperlinked.
For example, the word "hogwash" on a website could be linked to a dictionary definition. The links look different from standard HTML links and modify the HTML of the page in NO WAY.
Do you think there should be a META tag to allow Mozilla to increase the text size of a page? No. That'd be stupid, cause chances are the user will want to be able to ZOOM in on ANY PAGE. The same thing applies to smart tags. No matter what page I'm browsing, I may want nouns or foreign words to be highlighted so I can get dictionary definitions.
In the end it is the choice of the user. I could write a plugin to IE NOW that would scan the HTML of the currenlty loaded page and hyperlink important words (a program called flyswat already does this). Am I modifying the content of the original author? No.
If I buy a book, underline some words and scribble on the side of the page references and comments, would this be illegal? Am I modifying the original content of the author?
For god sakes people. The page isn't being modified. Think of it as a layer on top on the page.
Lets take an example of an accessibility tool. Lets say I write a tool which reads out definitions of words currently under the mouse. Would this be illegal?
Now lets say I write another tool that'll open up a webpage that gives me definitions of words that I click on with my mouse.
I find that the new tool I created is hard to use cause it is hard to figure out which words can be clicked on. I draw swiggles under these words. Now, is this illegal?
Now, another example. Windows XP has this new 'skinning' feature. It makes programs look and sometimes act differently from when they were originally written. Is this illegal?
Smart link links don't look anything like a normal link. They don't work like a normal link either.
It doesn't matter what microsoft does to the webpage as it appears in the CLIENT side as long as the user isn't deceived into thinking the links are made by the site's author.
That isn't happening here.
Lets say I write a program that searches all text all windows and hyperlinks words to dictionary definitions. This would work with word, ie, outlook etc. Would this be illegal cause I'm making word documents, webpages and HTML pages look and act differently?
BTW, there already exists such a program (it was formally known as flyswat and is very useful).
Slashdot Mirror
I didn't do very comprehensive tests unfortunately. I did enough to satisfy myself ;).
.NET CLR. I don't think the C# compiler currently emits tail calls (I could be wrong).
Tail calls are supported by the
There are other things like asynchronous I/O, asynchronoos method calls,generational garbage collection and value types (stack based objects) that also help C# get a speed advantage over Java.
Speed was definitely a major concern for Microsoft. Much more so than it was for Sun.
Microsoft also obviously has the skill to write very fast VMs from their experience with Java.
Um. Visual Age has a horrible UI.
:|.
Last I looked you couldn't 'draw' a form, double click to fill in event code like you would in VB and have VAJ generate JSP pages that would dynamically generate HTML/JavaScript representations of the form.
Also, does VAJ have edit and continue? Does VAC++ have edit and continue? I genuinely curious. Also does the latest version support intellisense? The one I tried last year didn't
Um, unsafe mode lets you use pointers to access memory (in arrays etc).
.NET that'll lead to better speed than Java (like tail calls).
Unsafe mode is hardly ever used.
I can tell you 'managed' code is certainly faster than java (from the experiments i've done).
I also tried using safe and unsafe code to copy from one array to another and didn't actually see much of a performance difference.
I guess when using unmanaged code from managed C++ you could see some improvement.
This is with JIT compilation BTW.
I'm sure with the IL native code compilers, the performance would be even better.
There are other features of
Yeah, agreed.
Windows Forms certainly uses windows primitives (like fonts) whereas Java seems to do those completely by itself (it just seems to need to be able to create windows and draw pixels...mostly).
Uh, I dare you to say that Haskell and C# are essentially the same langauge with different syntax.
Ditto with Prolog/Mercury and C#.
Is address auto-completion (like Netscape Messenger) something I need to turn on manually in Outlook
Uh Ctrl+K?
Yes but I was more concerned with raw processing speed of C# being faster than Java.
BTW, I'm pretty sure swing on windows eventually translates to Win32 API calls. It's just high level enough to abstract out all the windows stuff.
If you mean, subjectively, your C# programs run as fast as your C++ programs, that's probably because C# relies on a lot of native code in its libraries
Um no. The
Some parts call on native code using P/Invoke (which is a heck of a lot better than JNI - which requires rewriting of existing libraries).
These parts are specialised like ADO.NET, ASP.NET and Windows Forms. But we aren't comparing the speed of these are we?
Well, so far, I haven't seen any such "great solutions" on Windows--just a lot of really mediocre software that it took companies a long time to develop and debug.
Well maybe you should look harder. Show me a better IDE than Visual Studio (in terms of debugging, edit and continue etc). Show me Linux's equivalent of Microsoft Access and show me how it is better. Tell the millions of people relying on windows around the world that they're wrong.
I use windows. I develop applications for windows (many of which are free).
I also develop Unix applications. I'm getting my degree from a Unix only university computer science department.
I prefer to use windows. I think the APIs are tidier, and better designed than in unix.
exhibit a:
CreateThread vs pthread_create
WaitForSingleObject vs pthread_wait, sem_wait, wait etc etc.
Tell me that I'm just stupid and don't know any better.
JDK mostly compiles, and its native code compiler is very good, better than Microsoft's C# native code compiler.
And you know this how? Have you actually tried C#? I have. It's faster than Java I can tell you.
1: .NET doesn't rely on DCOM at ALL. No DCOM in sight.
.NET supports asynchronous IO. Asynchronous method calls (call now, get return result later). Java does not.
2:
Is it a good design decision to automate this? That is an open question. If you automate boxing/unboxing, most programmers won't know it's happening, and they will wonder while their code runs so slowly. It also leads to unobvious behavior when people try to use inheritance with number classes.
Um, integers won't be boxed until they're used as objects (put in a collection or have a method called). So it won't be any slower than when you have to manually box in java.
Also, you can't inherit from number classed (they're sealed). In Java you can't inherit from java.lang.Integer, java.lang.Double etc cause they're sealed (finalised anyway) as well.
Um, the registry contains much more than just software settings.
If you did a cp -r etc/* and did an upgrade then copied it back would it work? No. You have to be selective with the conf files.
Same with the registry. Just backup HKCU/Software/Whatever and your software settings will be sweet.
The greatest trick Bill Gates ever pulled, they will say, was that he made windows desktop software so ridiculously difficult to install, use, and maintain (via the windows registry).
I'm sorry, but I have to say the windows registry still remains one of the best things about windows. Don't give me that "it corrupts" shit. That was years ago. It's so much better to have a centralised storage database for settings.
You're telling me that the registry (structered settings database) is MORE complicated than the gazzilions of configuration files in Linux (all of which have different formats etc)?
Aparently you're required to connect to their servers everytime you want to view a webpage.
That's soooo wrong.
Things I've learnt from taco today:
1) Microsoft is evil.
2) Microsoft Passport is NOT some sort of unified authentication service. It aparently isn't like the orignal authentication services Microsoft used to use on their websites. It's just pure evil and the work of the devil.
3) You can't develop under windows unless you subscripe to passport:
4) Microsoft is evil.
5) Microsoft is evil.
Come on Taco, take off that second hand evangelical suit you're always wearing. It's got ketchup (and god knows what else) stains on it.
How is Passport different from the original authentication scheme they used for online MSDN subscriber downloads (except that it's the same system used for hotmail, windowsbeta, messenger etc).
In order for me to do lots of things on sun's java website, I had to send my details to them. Hang on, it's not called "Sun Passport". It must be ok.
Well that's the thing isn't it.
When Microsoft's routers went down then it can't be their routers (routers don't fail), it must have been their DNS servers running Windows. Ofcourse they needed to go and get some BSD CDs from some friendly geeks and install BSD on their DNS servers in order to get their sites back up. Right.....
If Slashdot's routers go down then it's definitely their routers - no doubt about it.
I wonder if encoding with BASE64 would help the situation....
;)
I'm too chicken to try it myself
I wonder how long your name is allowed to be. You could send DeCSS to mars :).
It's different. The -turbo option basically just hides the window, and leaves the process still running. I can close all my IE windows (leaving no iexplorer.exe processes left) and IE will load much faster than mozilla. This is with "start in new process" turned on.
-turbo IS NOT THE SAME process as what IE uses.
Sure, wait till Windows XP and IE6 are out of BETA and try again.
I wasn't the one who mentioned money. I was responsing to a post.
And if IE is there to CONTROL the internet, why does ASP.NET directly support netscape?
Funny, last I checked MS still hadn't bought a copy of my web site.
But it has nothing to do with Microsoft buying your site. It is the USER who is doing the highlighting by turning the option on. Microsoft simply supply the highlighter for the user to use (if they choose to).
Are you saying I can't view your website, print it out and then highlight it?
I expect Microsoft will be forced to shift from the exclusionary tag model to an inclusionary tag model where only sites with an inclusionary tag can be modified in this way. That way content owners have to give their eplicit permission to microsoft to edit their page in ways they would be completely unaware of.
*sigh* All these posts just prove that most drones here don't even know what smart tags is. Microsoft need not get exclusive rights from content owners to edit their page in ways they are completely unaware of. The pages aren't being edited. Certain keywords on the page are being hyperlinked.
For example, the word "hogwash" on a website could be linked to a dictionary definition. The links look different from standard HTML links and modify the HTML of the page in NO WAY.
Do you think there should be a META tag to allow Mozilla to increase the text size of a page? No. That'd be stupid, cause chances are the user will want to be able to ZOOM in on ANY PAGE. The same thing applies to smart tags. No matter what page I'm browsing, I may want nouns or foreign words to be highlighted so I can get dictionary definitions.
In the end it is the choice of the user. I could write a plugin to IE NOW that would scan the HTML of the currenlty loaded page and hyperlink important words (a program called flyswat already does this). Am I modifying the content of the original author? No.
If I buy a book, underline some words and scribble on the side of the page references and comments, would this be illegal? Am I modifying the original content of the author?
For god sakes people. The page isn't being modified. Think of it as a layer on top on the page.
Lets take an example of an accessibility tool. Lets say I write a tool which reads out definitions of words currently under the mouse. Would this be illegal?
Now lets say I write another tool that'll open up a webpage that gives me definitions of words that I click on with my mouse. I find that the new tool I created is hard to use cause it is hard to figure out which words can be clicked on. I draw swiggles under these words. Now, is this illegal?
Now, another example. Windows XP has this new 'skinning' feature. It makes programs look and sometimes act differently from when they were originally written. Is this illegal?
Uh. Running Windows XP and that doesn't happen. And it won't happen.
There are thousands of features in Windows XP which are disabled by default.
Microsoft spend hundreds of millions on IE and didn't make much money off it either...
Smart link links don't look anything like a normal link. They don't work like a normal link either.
It doesn't matter what microsoft does to the webpage as it appears in the CLIENT side as long as the user isn't deceived into thinking the links are made by the site's author.
That isn't happening here.
Lets say I write a program that searches all text all windows and hyperlinks words to dictionary definitions. This would work with word, ie, outlook etc. Would this be illegal cause I'm making word documents, webpages and HTML pages look and act differently?
BTW, there already exists such a program (it was formally known as flyswat and is very useful).
Um. No.
Turn font smoothing on in windows.
Font handling in X still sucks.