No we're talking about directX. So let's see:
1) DirectPlay - deprecated and has not changed much since DX8. Certainly doesn't require any new driver models.
2) DirectAudio - crippled in Vista (no hardware acceleration), no major changes. OpenAL still works fine and Creative now has drivers which interceptes DirectAudio calls and translates them to OpenAL. Again, all can work fine with new and old driver models.
3) DirectInput - no changes since DX8.
4) XInput - no changes since DX9.
5) Direct3D - all the touted improvements.
So everything except D3D can readily be ported back to XP.
Nah. Just the foundation was just rewritten from the ground up. It interacts with drivers differently. (New driver model remember.) Features that used to be in the drivers and in applications are now in directx, and the kernel. But the developer facing api's are much the same, so I guess nothing's changed much.
Sure, but the foundation is not a big part of DX - it needs to deal with low-level things like surface setup and memory management. Almost everything other should not be changed much.
Grab WINE sources and see their DX emulation layer. The base code is fairly small compared to things like shader support.
And you can get Compiz on Redhat 3 too if your willing to patch it enough. No valid technical reason you can't.
There's a small difference - almost nobody uses RedHat 3 now, while XP is the most used desktop OS.
At some point that line of argument becomes silly. We've crossed the mark. The amount of work to get directx10 (not JUST direct3d) into XP would pretty much amount to upgrading XP to Vista.
No. The rest of DirectX has not changed much at all. So it's trivial to port it. The main changes are in D3D.
D3D can also be ported without problems, at least parts relevant to gaming. You won't get Aero on XP, but DX10 games should work fine.
DirectDraw/Direct3D is not DirectX. It is a small (but high profile) part of directX. So? We're not speaking about DirectSound (which was _crippled_ in Vista, BTW) or DirectPlay. We're speaking about D3D, which is supposedly impossible to port.
DirectX is MORE than than just some 3d api stuff. So? Other DirectX part were not changed much in Vista.
Well aren't you clever. But Compiz doesn't run on XP now does it. So I fail to see how that would be a solution for XP. Still don't understand? DX10 can be ported to XP (and other platforms, in fact) as an OpenGL-DX translation layer, no fancy new graphical architecture is necessary. That means XP is perfectly capable to support all relevant DX10 capabilities.
Of course, Microsoft wouldn't even need to bother with OpenGL-DX translation - they can directly add missing feature to DX core.
It's only a matter of marketing - MS doesn't want DX10 on XP. There are no valid technical reasons for this.
Not quite the same ballpark. In terms of MS Office dx7-dx8-dx9 is Office 2000 to Office XP to Office 2003. DX10 is Office 2007 with docx and ribbons.
So? DX5/6/7 came out roughly every 1.5 years and driver developers somehow managed to write good drivers. And now they have several years to port DX10.
I'm sorry. I meant simultaneously hardware accelerated d3d. You know, so if one program has a spinning rendered textured and shaded cube at 120fps in one window, and you switch to another program in another overlapping window with its own rendered texture mapped shaded spinning regular polyhedron, the cube in the first one doesn't drop to a framerate you can count on your fingers... its 2008. They should both be able to spin at full speed. While a movie is playing in a 3rd window, on a desktop with 3d shadow effects if that's what the user wants.
That was supported since late 90-s. You can create several accelerated graphical contexts and they will work along nicely. Try to run several 3D-graphical applications on XP - it just works. Now, XP heavily balances CPU/GPU power in favor of the foreground application (which makes sense), but it's a purely tuning matter. If you don't believe me - look at Linux, Compiz can work along nicely with 3D applications.
Yep. DirectDraw Clippers were available SINCE DIRECTDRAW 2.0 IN 1996 - the earliest version of DirectX (DirectDraw 1.0 was known as Game SDK). I know, I myself wrote applications for science graphics rendering with several graphical contexts.
Only someone in marketing would suggest that. "Hey, lets take all the revolutionary big features out of DirectX10, backport it to windows 98; and claim we've got directX10 working on Windows 98" Because, hey, you could do that. You could even show some program that checks for directx10 and makes a couple directx10 api calls to prove your programming mojo.
But, sorry, that isn't directx10.
Sorry, but what is DirectDraw/Direct3D? I somehow thought that it was a 3D API. 3D applications don't care about hotswapable graphic cards, they only care about that 'several API calls'. That API calls can certainly be ported to Windows XP, there's no great technical barriers.
See above. That isn't dx10 emulation. That's adding support for some dx10 api's using dx9/ogl.
Nope. New DX10 features are already present as OpenGL extensions. So these projects just build DX10 API on top of OpenGL. It's not emulation, it's translation.
That's great if you want to run Halo on XP or something, but try something actually impressive... get AeroGlass running on XP, while playing a DVD movie in one window and WoW in another. Then click the start menu without having the other two windows choke up.
Not a problem. I can run Compiz while playing Quake 3 and running a DVD player in Linux. All with current OpenGL.
Sure, you'll need new drivers for DX10, there's nothing unusual. You also needed new drivers for DX9, DX8, DX7 and so on.
And of course, I think most of people will gladly agree to leave DRM parts of DX10 unported.
You don't need new driver architecture for DX10, it can work well enough with the old one. You just won't get hot-swap support and other goodies.
Oh, and using multiple D3D applications simultaneously was supported since DX2 (via DirectDraw Clipper object). Vista allows to make _composite_ applications, i.e. a D3D surface which is in turn mapped into another surface.
So there's NO good technical reason not to port DX10. In fact, there are projects to make DX10 emulation using OpenGL features.
QT is _GPL_, not LGPL. So I can't use it in my commercial software. Yes, I know that all software must be free - but it's an imperfect world.
I don't really mind reasonable restrictions on VisualStudio and Windows. After all, I make money from selling software, so it's fair if Microsoft gets (a rather small) cut of my money. And with Microsoft I can be reasonably sure that they are not going to disappear next year.
Of course, I like to use OpenSource stuff, and I contribute bugfixes and features to several projects.
GTK actually works quite fine on Windows and Mac. Certainly, it can be compiled by newer MS compilers without much problems - I did it myself about 3 years ago with MSVS2003. But GTK just doesn't look as a native application.
wxWidgets is great if you don't need to create custom controls or a very complex GUI. I know, I tried both. And sometimes you need to deal with issues of underlying platform.
========= Is Launchpad Free Software/open source? If not, why not?
Like Sourceforge and Google Code Hosting Launchpad is not open source. Unlike those other services, we have committed to making Launchpad Free Software.
Launchpad exists not just to make development easier within projects, but also to encourage collaboration between projects. With Launchpad, distributors and upstream developers can share bug reports, translations, and code. Until there are standards and APIs that let standalone sites send these things to each other, having a single Launchpad instance is the most practical way of maximizing the collaboration that Launchpad is designed to achieve. It also helps in funding Launchpad development in the short term.
We have released the code for [WWW] CSCVS, which Launchpad uses to mirror CVS/Subversion trees to Bazaar. We also make many contributions to other Free Software components we use, including [WWW] Zope, [WWW] Bazaar, [WWW] SQLObject, and [WWW] Twisted. =========
By the way, they can gladly send you Launchpad source if you sign NDA/non-compete.
Why would Microsoft shoot themselves in the foot? Unlike Trolltech, Microsoft doesn't need VisualStudio development to support itself. And developers are the greatest asset of Microsoft, and Microsoft really understands business. And in the worst case, developers will have 'good enough' tools for about 5-7 years if Microsoft starts charging $100000 per seat for MSVS tomorrow.
So in practice, choosing Microsoft is almost risk-free. It's not the same with QT, nobody knows what is going to happen with them a year from now.
As for single desktop... It doesn't make sense to write _only_ for Linux.
However, today Microsoft tools cost almost nothing. You can get Visual Studio Express for free and professional version for something like $150. I don't think that this price is going to change in foreseeable future. And you don't need to upgrade it often - I still use VisualStudio 2003 for my C++ development, and a lot of people still use ten years old VS6!
On the other hand, QT costs a lot (I'm not going to pay that much money for a desktop toolkit) and its price IS going up with each release.
GTK produces revolting UIs on Windows and Mac, so it's not good either.
The only other somewhat good cross-platform UI toolkit I know is wxWidgets. It doesn't look like shit on Mac, at least.
Just use a good laptop. I can easily reach touchpad on my laptop with right thumb with only a little movement of my hand. It's certainly much faster than using a mouse.
OK, your cell provider can limit number of messages per hour/day. But how is it different from plans with 'texting packages'?
Voice calls, actually, consume very little control traffic - less than one message per minute on average (I worked as engineer in a cell provider). Also, most of messages are processed locally on BSC (Base Station Controller) or just one hop away.
BTW, SMS message floods are known to crash cell networks. I saw it myself once on New Year eve when everybody started sending SMS messaged almost at once.
Yes, I'm comparing Spring to all other IoC frameworks (Spring also can be used with annotation-based configuration requiring zero XML configuration files, BTW).
The biggest Spring feature - it can use the whole gamut of Java technologies. For example, my current application can be used in EJB container or standalone with Atomikos distributed transaction manager. With zero changes in code.
But I was not even able to _find_ a distributed transaction manager for RoR! And it's absolutely essential if you need guaranteed consistency in transactions involving several data sources (several databases, for example).
SMS is NOTHING like a movie you download over your cable.
Mostly because SMS messages have to go over the signaling channels of providers (which are also used for other messages - like voice channel setup, routing information exchange,etc.), not over the main voice/data channels. That's because SMS are implemented as a store_and_forward control message. As a result too many SMS messages can overwhelm cell provider network.
IoC is nothing new. However, good IoC _frameworks_ is a relatively new development.
Spring blows almost everything out of water - it's very powerful and easy to use, and it can be integrated easy enough with _everything_. Even for very complex applications.
"The group includes both left and right-wing opposition leaders as well as mainstream liberals such as former world chess champion and United Civil Front leader _Garry_ _Kasparov_, former Prime Minister of Russia and People Democratic Union leader Mikhail Kasyanov and Russian Republic Party leader Vladimir Ryzhkov, _as_ _well_ _as_ _the_ _National_ _Bolshevik_ _Party_ _with_ _its_ _leader_ _Eduard_ _Limonov_ and far-left Vanguard of RedYouth."
There's also a good photo: http://duma.lenta.ru/news/2007/12/02/others/ - Kasparov is on the left and Limonov is on the right. It shows Kasparov and Limonov putting corrupted (with several marked candidates) voting papers in a ballot box.
I live in Russia, the association of Kasparov and NBP is very well-known and alienated a lot of Kasparov's supporters.
Slashdot Mirror
1) DirectPlay - deprecated and has not changed much since DX8. Certainly doesn't require any new driver models.
2) DirectAudio - crippled in Vista (no hardware acceleration), no major changes. OpenAL still works fine and Creative now has drivers which interceptes DirectAudio calls and translates them to OpenAL. Again, all can work fine with new and old driver models.
3) DirectInput - no changes since DX8.
4) XInput - no changes since DX9.
5) Direct3D - all the touted improvements.
So everything except D3D can readily be ported back to XP.
Nah. Just the foundation was just rewritten from the ground up. It interacts with drivers differently. (New driver model remember.) Features that used to be in the drivers and in applications are now in directx, and the kernel. But the developer facing api's are much the same, so I guess nothing's changed much.
Sure, but the foundation is not a big part of DX - it needs to deal with low-level things like surface setup and memory management. Almost everything other should not be changed much.
Grab WINE sources and see their DX emulation layer. The base code is fairly small compared to things like shader support.
And you can get Compiz on Redhat 3 too if your willing to patch it enough. No valid technical reason you can't.
There's a small difference - almost nobody uses RedHat 3 now, while XP is the most used desktop OS.
At some point that line of argument becomes silly. We've crossed the mark. The amount of work to get directx10 (not JUST direct3d) into XP would pretty much amount to upgrading XP to Vista.
No. The rest of DirectX has not changed much at all. So it's trivial to port it. The main changes are in D3D.
D3D can also be ported without problems, at least parts relevant to gaming. You won't get Aero on XP, but DX10 games should work fine.
DirectX is MORE than than just some 3d api stuff. So? Other DirectX part were not changed much in Vista.
Well aren't you clever. But Compiz doesn't run on XP now does it. So I fail to see how that would be a solution for XP. Still don't understand? DX10 can be ported to XP (and other platforms, in fact) as an OpenGL-DX translation layer, no fancy new graphical architecture is necessary. That means XP is perfectly capable to support all relevant DX10 capabilities.
Of course, Microsoft wouldn't even need to bother with OpenGL-DX translation - they can directly add missing feature to DX core.
It's only a matter of marketing - MS doesn't want DX10 on XP. There are no valid technical reasons for this.
So? DX5/6/7 came out roughly every 1.5 years and driver developers somehow managed to write good drivers. And now they have several years to port DX10.
I'm sorry. I meant simultaneously hardware accelerated d3d. You know, so if one program has a spinning rendered textured and shaded cube at 120fps in one window, and you switch to another program in another overlapping window with its own rendered texture mapped shaded spinning regular polyhedron, the cube in the first one doesn't drop to a framerate you can count on your fingers... its 2008. They should both be able to spin at full speed. While a movie is playing in a 3rd window, on a desktop with 3d shadow effects if that's what the user wants.
That was supported since late 90-s. You can create several accelerated graphical contexts and they will work along nicely. Try to run several 3D-graphical applications on XP - it just works. Now, XP heavily balances CPU/GPU power in favor of the foreground application (which makes sense), but it's a purely tuning matter. If you don't believe me - look at Linux, Compiz can work along nicely with 3D applications.
These things don't even begin to get near where I'm talking about:
http://msdn2.microsoft.com/en-us/library/aa919937.aspx
Yep. DirectDraw Clippers were available SINCE DIRECTDRAW 2.0 IN 1996 - the earliest version of DirectX (DirectDraw 1.0 was known as Game SDK). I know, I myself wrote applications for science graphics rendering with several graphical contexts.
Only someone in marketing would suggest that. "Hey, lets take all the revolutionary big features out of DirectX10, backport it to windows 98; and claim we've got directX10 working on Windows 98" Because, hey, you could do that. You could even show some program that checks for directx10 and makes a couple directx10 api calls to prove your programming mojo.
But, sorry, that isn't directx10.
Sorry, but what is DirectDraw/Direct3D? I somehow thought that it was a 3D API. 3D applications don't care about hotswapable graphic cards, they only care about that 'several API calls'. That API calls can certainly be ported to Windows XP, there's no great technical barriers.
See above. That isn't dx10 emulation. That's adding support for some dx10 api's using dx9/ogl.
Nope. New DX10 features are already present as OpenGL extensions. So these projects just build DX10 API on top of OpenGL. It's not emulation, it's translation.
That's great if you want to run Halo on XP or something, but try something actually impressive... get AeroGlass running on XP, while playing a DVD movie in one window and WoW in another. Then click the start menu without having the other two windows choke up.
Not a problem. I can run Compiz while playing Quake 3 and running a DVD player in Linux. All with current OpenGL.
Sure, you'll need new drivers for DX10, there's nothing unusual. You also needed new drivers for DX9, DX8, DX7 and so on.
And of course, I think most of people will gladly agree to leave DRM parts of DX10 unported.
You don't need new driver architecture for DX10, it can work well enough with the old one. You just won't get hot-swap support and other goodies.
Oh, and using multiple D3D applications simultaneously was supported since DX2 (via DirectDraw Clipper object). Vista allows to make _composite_ applications, i.e. a D3D surface which is in turn mapped into another surface.
So there's NO good technical reason not to port DX10. In fact, there are projects to make DX10 emulation using OpenGL features.
You do have it, a classical bicameral parliament.
It's called 'Senate' and 'Congress'.
No, I meant it in FSF way. I'd like _my_ software to be free, but it's impossible for now.
Actually, MSVS prices were quite high in late 90-s, significantly more than now.
There's an opinion that Microsoft held high prices because it _wanted_ the competition to attract more developers to Windows platform.
QT is _GPL_, not LGPL. So I can't use it in my commercial software. Yes, I know that all software must be free - but it's an imperfect world.
I don't really mind reasonable restrictions on VisualStudio and Windows. After all, I make money from selling software, so it's fair if Microsoft gets (a rather small) cut of my money. And with Microsoft I can be reasonably sure that they are not going to disappear next year.
Of course, I like to use OpenSource stuff, and I contribute bugfixes and features to several projects.
Sorry, I meant 'Standard Edition': http://www.amazon.com/Microsoft-Visual-Studio-2008-Standard/dp/B000WM1Z46 - it can be bought even cheaper in quantity. We got it for about $150 for 10 developers.
Team software suite is not necessary for a lone developer or a small company.
GTK actually works quite fine on Windows and Mac. Certainly, it can be compiled by newer MS compilers without much problems - I did it myself about 3 years ago with MSVS2003. But GTK just doesn't look as a native application.
wxWidgets is great if you don't need to create custom controls or a very complex GUI. I know, I tried both. And sometimes you need to deal with issues of underlying platform.
QT is much better, but it's commercial.
Quoting https://help.launchpad.net/FAQ#head-34295746b9c12bbe42eee4a9bd5e2656306fd796:
=========
Is Launchpad Free Software/open source? If not, why not?
Like Sourceforge and Google Code Hosting Launchpad is not open source. Unlike those other services, we have committed to making Launchpad Free Software.
Launchpad exists not just to make development easier within projects, but also to encourage collaboration between projects. With Launchpad, distributors and upstream developers can share bug reports, translations, and code. Until there are standards and APIs that let standalone sites send these things to each other, having a single Launchpad instance is the most practical way of maximizing the collaboration that Launchpad is designed to achieve. It also helps in funding Launchpad development in the short term.
We have released the code for [WWW] CSCVS, which Launchpad uses to mirror CVS/Subversion trees to Bazaar. We also make many contributions to other Free Software components we use, including [WWW] Zope, [WWW] Bazaar, [WWW] SQLObject, and [WWW] Twisted.
=========
By the way, they can gladly send you Launchpad source if you sign NDA/non-compete.
Why would Microsoft shoot themselves in the foot? Unlike Trolltech, Microsoft doesn't need VisualStudio development to support itself. And developers are the greatest asset of Microsoft, and Microsoft really understands business. And in the worst case, developers will have 'good enough' tools for about 5-7 years if Microsoft starts charging $100000 per seat for MSVS tomorrow.
So in practice, choosing Microsoft is almost risk-free. It's not the same with QT, nobody knows what is going to happen with them a year from now.
As for single desktop... It doesn't make sense to write _only_ for Linux.
However, today Microsoft tools cost almost nothing. You can get Visual Studio Express for free and professional version for something like $150. I don't think that this price is going to change in foreseeable future. And you don't need to upgrade it often - I still use VisualStudio 2003 for my C++ development, and a lot of people still use ten years old VS6!
On the other hand, QT costs a lot (I'm not going to pay that much money for a desktop toolkit) and its price IS going up with each release.
GTK produces revolting UIs on Windows and Mac, so it's not good either.
The only other somewhat good cross-platform UI toolkit I know is wxWidgets. It doesn't look like shit on Mac, at least.
Just use a good laptop. I can easily reach touchpad on my laptop with right thumb with only a little movement of my hand. It's certainly much faster than using a mouse.
Subj.
Nothing to add.
OK, your cell provider can limit number of messages per hour/day. But how is it different from plans with 'texting packages'?
Voice calls, actually, consume very little control traffic - less than one message per minute on average (I worked as engineer in a cell provider). Also, most of messages are processed locally on BSC (Base Station Controller) or just one hop away.
BTW, SMS message floods are known to crash cell networks. I saw it myself once on New Year eve when everybody started sending SMS messaged almost at once.
The problem is - you can easily write hundreds of SMS messages. There are even "SMS Chat" applications.
Making a hundred voice calls is much more challenging.
Yes, I'm comparing Spring to all other IoC frameworks (Spring also can be used with annotation-based configuration requiring zero XML configuration files, BTW).
The biggest Spring feature - it can use the whole gamut of Java technologies. For example, my current application can be used in EJB container or standalone with Atomikos distributed transaction manager. With zero changes in code.
But I was not even able to _find_ a distributed transaction manager for RoR! And it's absolutely essential if you need guaranteed consistency in transactions involving several data sources (several databases, for example).
No, it won't be for free. It means you've paid for unlimited voice plan.
Of course, if you use more than average number of phone calls - then your phone company starts losing money. But on average it still makes a profit.
SMS is NOTHING like a movie you download over your cable.
Mostly because SMS messages have to go over the signaling channels of providers (which are also used for other messages - like voice channel setup, routing information exchange,etc.), not over the main voice/data channels. That's because SMS are implemented as a store_and_forward control message. As a result too many SMS messages can overwhelm cell provider network.
Current breeder designs use NaK alloy for cooling - it melts at about room temperature, though it's a bit more dangerous to work with.
RTGs? That's puny!
Russian radar satellites had the whole _nuclear_ _reactors_ (one of them crashed somewhere in Canada)!
See: http://en.wikipedia.org/wiki/RORSAT
IoC is nothing new. However, good IoC _frameworks_ is a relatively new development.
Spring blows almost everything out of water - it's very powerful and easy to use, and it can be integrated easy enough with _everything_. Even for very complex applications.
Look here: http://en.wikipedia.org/wiki/The_Other_Russia
"The group includes both left and right-wing opposition leaders as well as mainstream liberals such as former world chess champion and United Civil Front leader _Garry_ _Kasparov_, former Prime Minister of Russia and People Democratic Union leader Mikhail Kasyanov and Russian Republic Party leader Vladimir Ryzhkov, _as_ _well_ _as_ _the_ _National_ _Bolshevik_ _Party_ _with_ _its_ _leader_ _Eduard_ _Limonov_ and far-left Vanguard of RedYouth."
There's also a good photo: http://duma.lenta.ru/news/2007/12/02/others/ - Kasparov is on the left and Limonov is on the right. It shows Kasparov and Limonov putting corrupted (with several marked candidates) voting papers in a ballot box.
I live in Russia, the association of Kasparov and NBP is very well-known and alienated a lot of Kasparov's supporters.
Oh, no problem.
What do you want to remove next?