OK, half of your complaints are key mappings. If you want to truly exploit the software, learn VS's key mappings. I looked thoroughly and found no way to map keys to the functions I complained about, nor to many more.
Lies, lies and more lies. There is a difference between a lie and not knowing something which is not immediately obvious.
You can compile across multiple processors in the same project with VS 2005 express edition. Use the/MP compiler option. I have, with 2005 express edition. Thank you for telling me.
If you go to thesubversion website, you'd see no less than three integration options with Microsoft Visual studio. I do not believe they work with the Express editions, and again, I'm not going to put down a large sum of money just to see if it begins to approach what I have with Eclipse. If we are comparing free to free, Visual C++ lacks the features Eclipse has, due to the lack of plugins. If we are comparing several hundreds or several thousands of dollars to free, I still fail to see the benefit I would have from switching to Visual C++, and still lack many features I have with Eclipse.
Funny, because I've had the opposite experience (heavy number-crunching codes) on identical hardware, and I've tried to have rational conversations with people to figure it out (because not only is the compiler faster, but the resulting code seems faster on Windows as well for the same number of cores, but the clusters I compute on are Linux), so I can't figure out if it's a gcc issue or if I am just dumb. Mostly playing with the Intel compilers right now to get a third opinion. I think it depends on the libraries used to a large extent. The test we did was a 2Ghz Core 2 Duo machine running GCC on Linux under Eclipse vs. a 3Ghz Core 2 Duo machine running Visual Studio 2005 Professional on Windows XP. Both were making clean debug builds of the same project (the start of a now much larger project) with the same major library versions. The Linux machine came in at 0:18 and the Windows machine came in at 1:16. I expected the Linux machine to come in last by far, but was pleasantly surprised.
If you disagree, please clarify, because maybe I'm just missing something with Eclipse... I disagree. Visual Studio lacks so many of the features I use every day in Eclipse, and the ones it does not lack are generally terribly done anyway (the exception being the debugger). A few examples of things Eclipse has but Visual C++ does not:
* Ctrl+Shift+R for opening a file in the project or workspace. Ctrl+Shift+R, partoffilename, enter and the file is open. I am constantly reaching for this whenever I use VC++ and missing it.
* Ctrl+Tab for switching header/implementation. I thought this would be a useless feature when I first started using it, but again, I am now constantly reaching for it in other IDE's.
* Multi-CPU compiling of the same project. Visual C++ charges (at least before 2008 came out, perhaps they dropped it - I doubt it) $2,500 for the luxury with the professional edition or above, and I'm not willing to drop $2,500 on -any- software just so it doesn't refuse to do something so basic.
* SVN integration right in the project view with Subversive. I can't tell you how much time this saves me. On Windows I use TortoiseSVN, and I'm constantly frustrated by the lack of integration and by the lack of features. In Eclipse, I can right-click a project, click Compare With->Base From Working Copy, and have a complete list of changed or added files. If I double-click a file, I get a very nice graphical diff with syntax highlighting. I use this every single day, and miss it in Visual C++.
* Compiler speed. GCC is much faster than Visual C++ 2005 Professional on the project I am working on.
* Mylyn is simply indispensable. It's hard to describe what it does; just try it. It's the best bug tracking IDE integration I've seen.
* Everything can be done with keystrokes. I hate my mouse; at least for development work, a keyboard is much faster. With Visual C++, I have not found a way to do many things without using the mouse (see Ctrl+Shift+R above); in Eclipse I can't find anything that is impossible to do without a mouse.
* AutoCompletion that actually works. Visual C++'s IntelliSense routinely falls over for me, and I have to hunt for.pcb files to delete before it will work again, if then. Eclipse's has never failed on me, and works just as well as or better than that of VC++ (but Eclipse folks, please sort the list by public members, then public inherited members, and -then- private/protected members and not by alphabetical order - that's my only gripe).
* I like working under Linux, and Eclipse's C++ toolset works best under Linux and OS X. Visual C++ doesn't at all. I would be willing to switch operating systems to get better tools, but not the other way around - as a result, I stay on Linux because the better tools are, in my opinion, here.
Again, disclaimer, this is only my opinion etc. etc. but I hope that this has explained why I personally find Eclipse superior.
But I don't kid myself that anything that can run on linux (whether kdevelop or eclipse, or *even* vim) is half as good as visual studio. I have tried to use Visual C++, and it used to be my IDE of choice, but after using Eclipse 3.3 (they improved the speed dramatically in 3.3, from unusable to usable) for a year or so I simply cannot bring myself to use anything else. I recently tested out Visual C++ 2008, and it really pales in comparison to what I love with Eclipse. About the only thing it did better was debugging, and I would gladly trade excellent debugging support for decent debugging support with SVN, Mylyn, Ctrl+Shift+R to open a resource,and IntelliSense that actually -works- without fail every time (Visual C++'s IntelliSense just fell over after about 2 hours of using it and never came back, and refused to give any information on any library - not even the standard C++ libraries.)
About the only thing I see VC++ having on Eclipse 3.3 for plain C++ development besides debugging is speed. Visual C++ is comparably fast, but Eclipse isn't slow in the "it takes a while to respond" sense, but rather in the "there is a 7 second splash screen when you start it" sense. I can live with that.
I was rather disappointed, really. I had thought that I had given up a great IDE (VC++) to go to a great OS (Linux) with a moderately decent IDE (Eclipse); in reality, looking back I have up a decent IDE to go to a great OS with a great IDE. A few years ago I loved Visual C++ 6, and I suppose I'm just sad to see its ancestor thwarted on my desktop so easily by a free program.
(Disclaimer: The statement above is merely my opinion. Your opinion may not match it.)
John Carmack spoke about this at QuakeCon 2007, not from a controlling in games perspective but rather from an optimization point of view. The eye really only focuses on one point at once; everything else in the field of vision is blurry. His idea was to only make what the eye was looking at on-screen high-detail, thus allowing much more detail in scenes.
And he disapproves of it so much because there is no acceptable alternative license that allows the original author to distribute it in a way that allows people to make and sell their own improvements. That's the whole point of using a copyleft license in the first place - to prevent people from making improvements that aren't shared with everyone. If you benefited from the software, so also should people benefit from your improvements to it. That's what it's all about.
That's a very good point, and I had forgotten about that. He can't seem to fathom the concept of people developing software because they enjoy it rather than because they want money in return.
I think his point was that if you choose Open Source software to modify and base your commercial software on and it is under the GPL, you must also Open Source your commercial software. He is arguing that that prevents proprietary improvements, and that that is wrong. What he can't seem to understand is that it -isn't their work to make proprietary in the first place-, and you can't base a commercial project off of Microsoft's code unless they specifically allow you to either. He's completely missing the fact that it's no different with the software he himself produces.
You're free to use GPL's tools to write proprietary software (John Carmack used the GNU toolset to write Quake on NEXT, and later donated $20,000 to the FSF as thanks for use of their tools), but you can't take a GPL'ed program, add a few lines of code, and sell it as a proprietary package. Bill Gates sees this as wrong, but somehow doesn't see that not being able to get the source code for Windows, add a few lines, and sell it as a new OS is the same darned thing.
Open source, he said, creates a license 'so that nobody can ever improve the software, It's pretty amazing that anything gets done, since what he describes as impossible is almost the only way Open Source software improves.
Do the original terms of the sale/license agreement permit Microsoft to do this?
That clause should go over well. "We reserve the right to deny you use of what you paid for whenever and however we wish."
They are so quick to apply property metaphors to data (copying as "theft", intellectual works as "intellectual property", finite distribution of an infinite object, etc.). Why aren't they applying it here? How would you feel if your toaster disappeared because Sunbeam decided not to make toasters anymore? Furthermore, how would you feel if a little card came with your toaster saying that might happen?
Eclipse is excellent for C++ development; just make sure you get 3.3 or higher since 3.2 is considerably slower and a large number of distros still only have 3.2 in their repositories. Make sure you have the Sun JRE (some distros come with other JVM's which run Eclipse at a crawl). On Ubuntu, I just install the "sun-java5-jre" package and download the Eclipse C++ development package from the www.eclipse.org site.
I'm not a huge fan of the fact that it is written in Java (due to speed, but it's honestly not slow at all on a halfway decent system), but at least it compiles quickly if you want to modify it for your needs (I compiled all of the C++ plugin from CVS in about 30 seconds).
I use it for daily development work and can't bring myself to go back to Visual Studio, even on Windows. It has, with a couple simple free plugin installs, built-in CVS and SVN support right in the project browser, a web browser (useful for looking up API documentation without switching windows), a task-oriented work plugin (Mylyn) which you really have to use to understand the power of (or view the presentation on Google Video from back when it was called Mylar), and more. Paired with GCC, I get multi-CPU compiles (even over a network with a bit of tinkering and distcc) for free, rather than the $2,500 Microsoft charges you for the luxury. I've simply never used a better tool.
I've been using the new bar for over a month now and it works perfectly for me. The behavior is predictable, and it takes several times of you picking a different, same selection before your top spot is taken over by something you don't expect (in practice it's never happened to me - things sorted themselves out and it always shows what I want most. I've gotten used to typing the first two letters of a favorite site, pressing down and Enter, and having it take me there predictably).
To turn your point around, what if Microsoft's software doesn't do what you wish? Then you're -really- hosed, since not only will they not modify it for such a (nearly statistically certain) "small" company, but you can't modify it yourself either.
I agree, but it doesn't get around the fact that it is a violation of basic freedoms.
Perhaps the reason we have trouble thinking of a just way to sell content is that we are trying to justify US copyright law, something which from it's humble beginnings as a control on publishers using printing presses has grown out of control. As Eben Moglen put it (paraphrasing), "Modern copyright law only makes sense if you take the issues and squint at them just right so they form an image that can't be seen looking at them directly." Treating an infinite resource as though it were finite can only make sense at a general level, and the closer you look the more it breaks down. We are trying to look closely at it, and it simply can't make sense at that level since it's foundation defies reality to begin with.
The same could be said of any law. "Uh-huh. Sure. That relys solely on trust. And the law is currently quite effective at dealing with people who smoke pot, isn't it? I somehow don't see a trust-based system working at all. Oh, and how will the law know you've smoked it?"
The alternative is having a police officer-type mechanism - whether a person or a machine - watch every move we make in our lives to make sure we don't violate any laws. That is -not- freedom.
Resale can't work without the stringent DRM that's being forced on consumers. If you put it that way, the same is true of copyright law itself. "Copyright can't work without the stringent DRM that's being forced on consumers." That simply isn't correct. If you sell a song, you delete it. If you fail to delete it, law takes over, not some omnipresent nanny software.
Notice how I use "free software" instead of "open source" - When the Web 2.0 bubble comes, it doesn't matter if you can just look at the source code from the tool you used after the company went bankrupt. You need to have the legal rights to keep modifying it and/or let someone else do it. Open Source is trademarked, and the definition mandates that an Open Source application not only reveal source code but "the license must allow modifications and derived works". http://www.opensource.org/docs/definition.php
"It looks like an object element that's pointing to a non-existent page (http://webstandards.org/404/) should be falling through, but that page is now returning a response status of 200 (success) instead of 404. Was it previously correctly returning a 404 response?"
So the OS is supposed to magically know how to poke it's fingers into a process and divide the workload of that process among multiple cores/CPU's without incurring performance loss or data corruption? If you know how, many people would sure be happy to know your secret.
GNOME for me requires only 128MB of ram at idle. I tested it on a machine with 256 megs of ram and 128MB of memory was being used at idle and only 32MB of swap was being used. *shrugs*
Slashdot Mirror
* Ctrl+Shift+R for opening a file in the project or workspace. Ctrl+Shift+R, partoffilename, enter and the file is open. I am constantly reaching for this whenever I use VC++ and missing it.
* Ctrl+Tab for switching header/implementation. I thought this would be a useless feature when I first started using it, but again, I am now constantly reaching for it in other IDE's.
* Multi-CPU compiling of the same project. Visual C++ charges (at least before 2008 came out, perhaps they dropped it - I doubt it) $2,500 for the luxury with the professional edition or above, and I'm not willing to drop $2,500 on -any- software just so it doesn't refuse to do something so basic.
* SVN integration right in the project view with Subversive. I can't tell you how much time this saves me. On Windows I use TortoiseSVN, and I'm constantly frustrated by the lack of integration and by the lack of features. In Eclipse, I can right-click a project, click Compare With->Base From Working Copy, and have a complete list of changed or added files. If I double-click a file, I get a very nice graphical diff with syntax highlighting. I use this every single day, and miss it in Visual C++.
* Compiler speed. GCC is much faster than Visual C++ 2005 Professional on the project I am working on.
* Mylyn is simply indispensable. It's hard to describe what it does; just try it. It's the best bug tracking IDE integration I've seen.
* Everything can be done with keystrokes. I hate my mouse; at least for development work, a keyboard is much faster. With Visual C++, I have not found a way to do many things without using the mouse (see Ctrl+Shift+R above); in Eclipse I can't find anything that is impossible to do without a mouse. * AutoCompletion that actually works. Visual C++'s IntelliSense routinely falls over for me, and I have to hunt for
Again, disclaimer, this is only my opinion etc. etc. but I hope that this has explained why I personally find Eclipse superior.
I use C++ with it day in and day out and it's far and above better in my eyes.
About the only thing I see VC++ having on Eclipse 3.3 for plain C++ development besides debugging is speed. Visual C++ is comparably fast, but Eclipse isn't slow in the "it takes a while to respond" sense, but rather in the "there is a 7 second splash screen when you start it" sense. I can live with that.
I was rather disappointed, really. I had thought that I had given up a great IDE (VC++) to go to a great OS (Linux) with a moderately decent IDE (Eclipse); in reality, looking back I have up a decent IDE to go to a great OS with a great IDE. A few years ago I loved Visual C++ 6, and I suppose I'm just sad to see its ancestor thwarted on my desktop so easily by a free program.
(Disclaimer: The statement above is merely my opinion. Your opinion may not match it.)
John Carmack spoke about this at QuakeCon 2007, not from a controlling in games perspective but rather from an optimization point of view. The eye really only focuses on one point at once; everything else in the field of vision is blurry. His idea was to only make what the eye was looking at on-screen high-detail, thus allowing much more detail in scenes.
That's a very good point, and I had forgotten about that. He can't seem to fathom the concept of people developing software because they enjoy it rather than because they want money in return.
I think his point was that if you choose Open Source software to modify and base your commercial software on and it is under the GPL, you must also Open Source your commercial software. He is arguing that that prevents proprietary improvements, and that that is wrong. What he can't seem to understand is that it -isn't their work to make proprietary in the first place-, and you can't base a commercial project off of Microsoft's code unless they specifically allow you to either. He's completely missing the fact that it's no different with the software he himself produces.
You're free to use GPL's tools to write proprietary software (John Carmack used the GNU toolset to write Quake on NEXT, and later donated $20,000 to the FSF as thanks for use of their tools), but you can't take a GPL'ed program, add a few lines of code, and sell it as a proprietary package. Bill Gates sees this as wrong, but somehow doesn't see that not being able to get the source code for Windows, add a few lines, and sell it as a new OS is the same darned thing.
That clause should go over well. "We reserve the right to deny you use of what you paid for whenever and however we wish."
They are so quick to apply property metaphors to data (copying as "theft", intellectual works as "intellectual property", finite distribution of an infinite object, etc.). Why aren't they applying it here? How would you feel if your toaster disappeared because Sunbeam decided not to make toasters anymore? Furthermore, how would you feel if a little card came with your toaster saying that might happen?
Eclipse is excellent for C++ development; just make sure you get 3.3 or higher since 3.2 is considerably slower and a large number of distros still only have 3.2 in their repositories. Make sure you have the Sun JRE (some distros come with other JVM's which run Eclipse at a crawl). On Ubuntu, I just install the "sun-java5-jre" package and download the Eclipse C++ development package from the www.eclipse.org site.
I'm not a huge fan of the fact that it is written in Java (due to speed, but it's honestly not slow at all on a halfway decent system), but at least it compiles quickly if you want to modify it for your needs (I compiled all of the C++ plugin from CVS in about 30 seconds).
I use it for daily development work and can't bring myself to go back to Visual Studio, even on Windows. It has, with a couple simple free plugin installs, built-in CVS and SVN support right in the project browser, a web browser (useful for looking up API documentation without switching windows), a task-oriented work plugin (Mylyn) which you really have to use to understand the power of (or view the presentation on Google Video from back when it was called Mylar), and more. Paired with GCC, I get multi-CPU compiles (even over a network with a bit of tinkering and distcc) for free, rather than the $2,500 Microsoft charges you for the luxury. I've simply never used a better tool.
Or, even better, "Linux Native".
I've been using the new bar for over a month now and it works perfectly for me. The behavior is predictable, and it takes several times of you picking a different, same selection before your top spot is taken over by something you don't expect (in practice it's never happened to me - things sorted themselves out and it always shows what I want most. I've gotten used to typing the first two letters of a favorite site, pressing down and Enter, and having it take me there predictably).
I wasn't going to post this (bored lunch break GIMP work), but now I must. http://i29.tinypic.com/25rjrew.png
To turn your point around, what if Microsoft's software doesn't do what you wish? Then you're -really- hosed, since not only will they not modify it for such a (nearly statistically certain) "small" company, but you can't modify it yourself either.
I agree, but it doesn't get around the fact that it is a violation of basic freedoms.
Perhaps the reason we have trouble thinking of a just way to sell content is that we are trying to justify US copyright law, something which from it's humble beginnings as a control on publishers using printing presses has grown out of control. As Eben Moglen put it (paraphrasing), "Modern copyright law only makes sense if you take the issues and squint at them just right so they form an image that can't be seen looking at them directly." Treating an infinite resource as though it were finite can only make sense at a general level, and the closer you look the more it breaks down. We are trying to look closely at it, and it simply can't make sense at that level since it's foundation defies reality to begin with.
The same could be said of any law. "Uh-huh. Sure. That relys solely on trust. And the law is currently quite effective at dealing with people who smoke pot, isn't it? I somehow don't see a trust-based system working at all. Oh, and how will the law know you've smoked it?"
The alternative is having a police officer-type mechanism - whether a person or a machine - watch every move we make in our lives to make sure we don't violate any laws. That is -not- freedom.
If you want to take that road, copyright isn't a right either.
As it turns out, the Acid2 test is broken lately due to a server going down causing an incomplete image to be displayed. The article: http://dailyapps.net/2007/12/internet-explorer-8-clears-acid2-really/
"It looks like an object element that's pointing to a non-existent page (http://webstandards.org/404/) should be falling through, but that page is now returning a response status of 200 (success) instead of 404. Was it previously correctly returning a 404 response?"
"Well that is it, the Acid2 test page is broken at the moment (...) try out the alternate test page over at http://www.hixie.ch/tests/evil/acid/002/"
I'm running a few-days-old Minefield (SVN build) release and it passes perfectly.
So the OS is supposed to magically know how to poke it's fingers into a process and divide the workload of that process among multiple cores/CPU's without incurring performance loss or data corruption? If you know how, many people would sure be happy to know your secret.
But I sure like my newfound ability to compile multiple source files at once and finish a 5-files-changed compile in a few seconds.
GNOME for me requires only 128MB of ram at idle. I tested it on a machine with 256 megs of ram and 128MB of memory was being used at idle and only 32MB of swap was being used. *shrugs*