While the STL is a powerfull base of collection structures and algorithms it's not as robust as what Qt has to offer in some respects. std::string, for example, is pretty bare bones as far as it's implementation and developers end up having to create to a whole slew helper functions just to do basic formating operations.
At this point they have only a couple of options: wait for the standards comitte to catch up with the real world (out of the question for a business); standardize on somebody else's library, like boost; or develop their own to fit their customer's needs exactly. I don't think TrollTech is losing sleep over their choice.
C does not support generic programming (templates) in any way. Sure C, in being turing complete, can perform any function that C++ can produce but the amount of extra textual expression required in C to match what the C++ compiler does for you is staggering. Less code equals less chance of bugs, lower maintenance and faster development (well, that last point is arguable given you really need a C++ language lawer at times). This is a very important difference in these languages that some people fail to remember when comparing.
Sure I'll give you that C++ is huge and messy with differing implementations but it is also the most cabable native-compiled language on the market with mass acceptance (i.e. portable). In a market such as GUI toolkits, using an expressive, natively compiled and portable language is crucial. This is where C++ is actually a sane choice over, C, Java, Object Pascal, OCaml and many others.
Is there really a future for "horizontal launches of reusable spacecraft"?
The lingo sounds great. I mean "Mojave Spaceport" certainly has a ring to it. But would this be any more than a red tape cut for private experimentalists?
Posted AC originally but I thought it needed better visibility...
The biggest reason for look-and-feel cloning is to make migration from Windows, with it's 98% control of the desktop market, to Linux as pain free as possible.
Besides, imagine trying to sell off the idea of Linux migration to a Fortune 500 company saying that "Oh, yeah, you'll have to retrain all of your staff who will be using the new Linux installation because we feel our WM and Desktop environment is cooler/slicker/13373r than that Windows crap." as opposed to "The computing environment will be familiar requiring minimal retraining and no loss of productivity."
designing fonts is not rocket science, but it comes pretty close. typography might even be the equivalent of rocket science in design.
That's very true. I made a suggestion in the KDE forum before but I never did send it to Dre at the League. http://dot.kde.org/1026495633/1026507253/1026514 88 6/
I have been watching the KDE development mailing lists for the last year and a half and I can tell you that the activity within the KDE development process is so active that it is certain that proper support for G++ 3.1 and the C++ standard library will be quickly implemented. At least, that is, if Waldo Bastian would have anything to say about it.
Which is why the UnitedLinux initiative is so important
When that next release of KDE comes out — with SVG icons, tabbed browsing and other such goodies — a single binary package would be able to satisfy all “Powered by UnitedLinux” distros. A standard that could support software that would be as easy to install as clicking an “install-me” icon, regardless of your distro, would be a Good Thing™.
The thing that bothers me, tho’, is how the UL standard will keep up with bazaar style, bleading edge software that is so signature of Linux and Open Source.
PureFiction is coordinating this convention. There seems to be a mentions of Alpine. There was a story about it before with threads containing a little history of PureFiction.
Some would call it "shameless self promotion".
I'd just call it being stoked about what you're doing
Hancom is about to release their Linux/Win32/MacOS X office package very soon. It's all build with Qt meaning it will blend in very well with the KDE environment (cut and paste, etc...).
Given the cross platform nature of it, it is also very likely you'll see it stocked on shelves where as pure linux apps tend to be neglected by retailers due to their low sales. Hancom Word is a very mature application and is quite popular in Asia. The pro package a will contain applications from theKompany who contributes to open source(especially with their very usefull PyQt). Even tho it isn't open source it should be a good thing for linux in general.
Maybe they are lucky (or not) that it is in about the same timeframe as going from gcc2 to gcc3. All c++ binaries will be broken with or without qt2/qt3 in most major distro's.
No that's not luck. That was a major reason why the desicion was made to move to Qt3 so quickly. This was heavily discussed about 4 months back and finally decided to timeframe the release with g++ 3.1 (because the minor release of g++ will break BC again).
Don't put it pass the KDE team to coordinate efforts with other projects.
That's what kinit is for. It's just a process to load libs and forks out kde applications so that load times aren't so harsh. The real problem though is how ld loads c++ generated binaries. That's why kde-2.2. has an obj-prelink tool
From what I hear there are some speed improvements in KDE 2.2.
But aside from that comparing Win2K GUI speed to KDE/Qt GUI speed is pretty pointless at this time. The real problem isn't KDE or Qt themselves but the graphics platform that they are running on, X. Not that X "sucks" or is "bloated" but the simple reason the the drawing architecture, Xlib, does not easily allow for hardware accellerated graphics routines. Sure, there are drivers for XFree86 that have these accellerations but they are available only where they fit, mainly in the Server code, they are not accessable from the client applications due to the Client/Server design of X. Accellerated Raster Op's, strech blitting, bit blitting are all unavailable to your client KDE/Qt applications, things that win32 API's can take for granted. If you really want to see the GUI speed of KDE/Qt or even Gnome/GTK+ improve you should think about giving support to Keith Packard's Xrender extensions to the X protocol (what's resposible for the anti-aliasing available now) or a port of the graphics API's to something like DirectFB.
Re:Mosfet's Liquid Style Engine
on
KDE 2.2 Released
·
· Score: 1, Informative
There is a version of liquid in the KDE CVS. It's called MegaGradient. The code was rolled into the CVS and forked from the original, which is one of the reasons that Mosfet has stopped developing in the KDE CVS and is now releasing his work under the QPL.
With the Qt libs available on Mac OS X, it wouldn't be hard for the KDE team to cut what little X dependencies they have and port KDE3 to said operating system.
The group who wrote the classics "Unreal" and "Second Reality" known as the Future Crew, started a little company back in '95 called Remedy. You may have heard of their game, "Max Payne". Well anyway when they first released a benchmarking tool called Final Reality (sounds like Second Reality, eh) it was such a success that Remedy spawned off the Mad Onion. Think of it as commercial demos.
Slashdot Mirror
While the STL is a powerfull base of collection structures and algorithms it's not as robust as what Qt has to offer in some respects. std::string, for example, is pretty bare bones as far as it's implementation and developers end up having to create to a whole slew helper functions just to do basic formating operations.
At this point they have only a couple of options: wait for the standards comitte to catch up with the real world (out of the question for a business); standardize on somebody else's library, like boost; or develop their own to fit their customer's needs exactly. I don't think TrollTech is losing sleep over their choice.
C does not support generic programming (templates) in any way. Sure C, in being turing complete, can perform any function that C++ can produce but the amount of extra textual expression required in C to match what the C++ compiler does for you is staggering. Less code equals less chance of bugs, lower maintenance and faster development (well, that last point is arguable given you really need a C++ language lawer at times). This is a very important difference in these languages that some people fail to remember when comparing.
Sure I'll give you that C++ is huge and messy with differing implementations but it is also the most cabable native-compiled language on the market with mass acceptance (i.e. portable). In a market such as GUI toolkits, using an expressive, natively compiled and portable language is crucial. This is where C++ is actually a sane choice over, C, Java, Object Pascal, OCaml and many others.
Idealism loses out to realism in business
If Inkscape or Sodipodi causes true believers like Everaldo and Jimmac find no reason to hop on their Macs then there you have it.
eDonkey and its variants - now thats a dumb-"ass" name
That's it... commence beat-down.
but what makes this special?
Is there really a future for "horizontal launches of reusable spacecraft"?
The lingo sounds great. I mean "Mojave Spaceport" certainly has a ring to it. But would this be any more than a red tape cut for private experimentalists?
Posted AC originally but I thought it needed better visibility...
The biggest reason for look-and-feel cloning is to make migration from Windows, with it's 98% control of the desktop market, to Linux as pain free as possible.
Besides, imagine trying to sell off the idea of Linux migration to a Fortune 500 company saying that "Oh, yeah, you'll have to retrain all of your staff who will be using the new Linux installation because we feel our WM and Desktop environment is cooler/slicker/13373r than that Windows crap." as opposed to "The computing environment will be familiar requiring minimal retraining and no loss of productivity."
Examples of usage are here
It's like succession... except legislation replaces rifles and geeks replace rednecks.
Proof positive /. moderation points have been ransacked by Windows users who only have half a clue.
That's very true. I made a suggestion in the KDE forum before but I never did send it to Dre at the League.4 88 6/
http://dot.kde.org/1026495633/1026507253/102651
I have been watching the KDE development mailing lists for the last year and a half and I can tell you that the activity within the KDE development process is so active that it is certain that proper support for G++ 3.1 and the C++ standard library will be quickly implemented. At least, that is, if Waldo Bastian would have anything to say about it.
Which is why the UnitedLinux initiative is so important
When that next release of KDE comes out — with SVG icons, tabbed browsing and other such goodies — a single binary package would be able to satisfy all “Powered by UnitedLinux” distros. A standard that could support software that would be as easy to install as clicking an “install-me” icon, regardless of your distro, would be a Good Thing™.
The thing that bothers me, tho’, is how the UL standard will keep up with bazaar style, bleading edge software that is so signature of Linux and Open Source.
PureFiction is coordinating this convention. There seems to be a mentions of Alpine. There was a story about it before with threads containing a little history of PureFiction.
Some would call it "shameless self promotion".
I'd just call it being stoked about what you're doing
Hancom is about to release their Linux/Win32/MacOS X office package very soon. It's all build with Qt meaning it will blend in very well with the KDE environment (cut and paste, etc...).
Given the cross platform nature of it, it is also very likely you'll see it stocked on shelves where as pure linux apps tend to be neglected by retailers due to their low sales. Hancom Word is a very mature application and is quite popular in Asia. The pro package a will contain applications from theKompany who contributes to open source(especially with their very usefull PyQt). Even tho it isn't open source it should be a good thing for linux in general.
Maybe they are lucky (or not) that it is in about the same timeframe as going from gcc2 to gcc3. All c++ binaries will be broken with or without qt2/qt3 in most major distro's.
No that's not luck. That was a major reason why the desicion was made to move to Qt3 so quickly. This was heavily discussed about 4 months back and finally decided to timeframe the release with g++ 3.1 (because the minor release of g++ will break BC again).
Don't put it pass the KDE team to coordinate efforts with other projects.
That's what kinit is for. It's just a process to load libs and forks out kde applications so that load times aren't so harsh. The real problem though is how ld loads c++ generated binaries. That's why kde-2.2. has an obj-prelink tool
that the idea of teaching a "child" system has been used for AI research.
Cog anybody?
From what I hear there are some speed improvements in KDE 2.2.
But aside from that comparing Win2K GUI speed to KDE/Qt GUI speed is pretty pointless at this time. The real problem isn't KDE or Qt themselves but the graphics platform that they are running on, X. Not that X "sucks" or is "bloated" but the simple reason the the drawing architecture, Xlib, does not easily allow for hardware accellerated graphics routines. Sure, there are drivers for XFree86 that have these accellerations but they are available only where they fit, mainly in the Server code, they are not accessable from the client applications due to the Client/Server design of X. Accellerated Raster Op's, strech blitting, bit blitting are all unavailable to your client KDE/Qt applications, things that win32 API's can take for granted. If you really want to see the GUI speed of KDE/Qt or even Gnome/GTK+ improve you should think about giving support to Keith Packard's Xrender extensions to the X protocol (what's resposible for the anti-aliasing available now) or a port of the graphics API's to something like DirectFB.
There is a version of liquid in the KDE CVS. It's called MegaGradient. The code was rolled into the CVS and forked from the original, which is one of the reasons that Mosfet has stopped developing in the KDE CVS and is now releasing his work under the QPL.
Actually,
With the Qt libs available on Mac OS X, it wouldn't be hard for the KDE team to cut what little X dependencies they have and port KDE3 to said operating system.
They have It's called 3DMark.
The group who wrote the classics "Unreal" and "Second Reality" known as the Future Crew, started a little company back in '95 called Remedy. You may have heard of their game, "Max Payne". Well anyway when they first released a benchmarking tool called Final Reality (sounds like Second Reality, eh) it was such a success that Remedy spawned off the Mad Onion. Think of it as commercial demos.
The demoscene will never die. It just keeps changing. To an Amiga cracker/intro/demo coder the scene has been truly dead since the early 90's. Try telling that to the sceners of today.