A lot of us haven't even gotten far enough to figure that out. Microsoft just switched us all over to a new UI toolkit (Windows Presentation Foundation) with the release of .NET 3 five or so years ago. We've invested a lot of time and money into getting up to speed with this new technology. And now we're already being asked to migrate away from it and onto yet another Next Big Thing.
Meanwhile, Microsoft zinged its mobile developers (the very ones they're so desperate to court) even harder by deciding that the new mobile platform they brought out only two years ago would not be compatible with the new mobile platform they're bringing out now.
It was twelve years ago this month that Steve Ballmer did his famous monkey dance while repeatedly shouting "Developers!" and apparently slipping into some sort of fugue state. Well, it turns out that all those developers he wanted to make such a big deal of saying are such a big deal? They're are all kind of pissed off and disgruntled right now.
Bingo! Talked about disrespecting your developers/users by changing the rules of the GUI every couple of years. Only the hardcore fans, pioneers and draftees are jumping into Windows 8. (The rest of us are keeping it locked up in virtual machines.) There's enough work in the Windows development universe, a developer can ignore Metro till it grows up or decent third party tools come out. For f--ks sake, Redmond, I've got real apps, (both web and desktop) to write.
And for a long time, it made everyone happy - because jumping on the new thing kept you employable, and actually shipping working products mattered less.
Microsoft did not change its ways. The world did.
As anyone noticed how many MacOS X APIs have been deprecated since the early days?
Or the API changes in Android?
Every corporation plays this game, as you point out.
I still remember the early COM days, where the only API was C based, and COM was called OLE.
The API changing game is another problem - it does not allow you to use existing application with newer system. As far as I know, this does not happen on Android (for now).
That is why you have to import android.support.* versus android.app. Which forces you to have some form of preprocessing in the Java code.
Plus the API for the action bar is still missing from the support library.
Except that the old widget sets never die. Xt, Athena, Motif, Lesstif, Tcl, gtk, Qt. Even swing and related tools if you include Java.
And so ... they're still in use. While significant new apps may not be getting developed with these tools, plenty of old ones are maintained, and the toolkits are largely serviceable -- they may be used if desired.
They actually provide a bit of useful visual feedback (if you're versed in Linux/Unix history), in that the style and design of an application gives you some idea of how it's going to work and what its quirks may be.
Consistency? We've heard of it.
edit: typo
Still, if you want I can post all the Linux kernel ABI changes.
The changes done in Apache during the 1.x to 2.x transition.
Or many other examples, as they are quite easy to find.
Nice to bend the reality to fill into your argumentation.
I've written Linux device drivers. They're easy to write, and API changes are not a concern at all.
private as "private/internal API" nothing to do with secret.
> I've written Linux device drivers. They're easy to write, and API changes are not a concern at all.
The whole point is that it changes! So it is ok for Linux to change the Kernel ABI, but for companies not?!
New platform, new kits. Same as any other company.
In contrast, Microsoft will just go and kill whole platforms from time to time (e.g Visual Basic) rendering useless years of experience building deep understanding of a platform.
"Classic" VB was the PHP of the client/server era: ugly, reasonably capable, and everywhere.
Imagine the screams you would be hearing across the Internet if the PHP core devs all decided en masse to go work on something else, and PHP was under a closed, proprietary license so nobody could fork it, and you get a sense for what it was like when Microsoft killed VB.
If a vendor who only supplies a single product decides to torpedo the APIs that you rely on, they create an easy opportunity for you to switch to another vendor entirely. Whereas historically Microsoft could wreak havoc upon an entire development platform (Visual Basic) knowing that probably all you'll be able to do about it is huff and puff around a bit on the Internet. After all you couldn't jump ship on Windows development - all your customers were Windows users, or your entire IT department was built around Windows, or your corporate intranet is all dependent on IE6, or whatever.
Take this game bundled with Windows 95. ftp://ftp.microsoft.com/deskapps/games/public/AAS/Hover.exe
It runs even on Windows 8. Upgrading from Windows 1.0 to Windows 7.
http://www.youtube.com/watch?v=vPnehDhGa14
As I said in another comment, I still see some ASP and Visual Basic usage even after all these years, those applications are still supported in Windows 8 and Windows Server 2012. Microsoft stopped making new versions because it wanted to move on to better technology. Want to take a poll on HN about what developers think of Visual Basic and Classic ASP? A case of damned if they do, damned if they don't.
[1] by which, I don't mean they are remotely in danger of being dethroned. Just that they can't dictate application standards anymore, because most activity happens on the web.
Using it to put developers within your own ecosystem off-balance is an antipattern.
WinRT 1.0 is basically a subset of Silverlight with a few specific controls thrown in.
That reminds me: WPF is not even a superset of Silverlight; There's functionality I need in Silverlight (regarding media codecs) which is not available in WPF, and my only way to get it is to instantiate a browser, run a silverlight app inside, and somehow bridge it all underneath in my app.
Pure madness.
You can still write WPF apps for Windows 8, of course, but they won't be Metro apps, they won't run on ARM tablets and, significantly for the current conversation, they won't be allowed in the Windows app store.
WP7 apps will run on WP8 without any modifications or even without anything needing to be done by the developers.
>Well, it turns out that all those developers he wanted to make such a big deal of saying are such a big deal? They're are all kind of pissed off and disgruntled right now.
Citation needed and content farms like TNW who will "print" anything to drive clicks don't count.
I know it's hip on HN to rag on Microsoft, but cmon, two thousand apps for a platform that is a month away from release(only available to MSDN users right now) and everyone is predicting doomsday? Perhaps developers are making use of the last month to further enhance apps before submitting them?
Almost every Metro style application demo I've seen is some variant of a "social" app. Do they really think that the millions of windows developers all suddenly started working on social apps?
Every one of the pivots that they have made in the past few years has been in reaction to some external threat and the developers are in the crossfire. Windows developers are too weary to be excited about yet another wide ranging change.
Contrary to popular belief by many geeks, game consoles do not support OpenGL as well. When they do, like the PS3, is a kind of subset that most studios avoid using anyway, and prefer to make use of the lower level APIs.
Re: networking, you are of course right. At the same time, I can sympathize with Microsoft's decision to abandon Winsock. Cross-platform use of BSD sockets is mired in a deplorable mess of #defines and '90s incompatibilities. For new code, the non-blocking WinRT API is a million times better -- the question is, what can Microsoft do to make porting easier?
What I find mind-blowing, though, is that Intel's new architecture Haswell which will come out mid 2013, and is supposed to be even more GPU-oriented than previous chips, will only support OpenGL 3.2, which will be 4 years old by the time the chip launches (some sources say OpenGL 4.0 - still 3 years old though)...and yet it will support DirectX 11.1. I find that incredibly disappointing.
I guess Intel just doesn't care about OpenGL. I'm not sure how they are supposed to make us believe that they are serious about entering the mobile market with that kind of attitude towards OpenGL.
They've publicly stated that the hardware is OpenGL 4 capable. (Google it.)
Bingo. Microsoft was never very fond of the idea of people porting Windows apps to other platforms. If people develop portable apps that run everywhere, the network effects achieved by having OEMs bundling Windows on every PC are neutralized.
(Note, i'm actually a little excited about the Surface tablet as a potential device that crosses the chasm between full tablet and laptop. But I don't have very many apps that I use now on my iPad.)
EDIT: Also, I don't know much about any game development SDK engines that abstract devs from these proprietary libraries and runtimes...however I think they must exist?
I have a Transformer Prime (#2, Transformer Infinity is the newest one) and the keyboard dock. Android apps/games, plus terminal, VNC, etc for doing real work.
And Android 4.0 looks more like Windows 7 than Windows 8 does (black "taskbar" on the bottom with clock and notifications on the right, home button, back button, and app switcher button on the left, and the 'desktop' can have icons and widgets).
That is not a reason to not code and offer a code wrapper for BSD-sockets-like API based on WinRT sockets.
Of course, no need to cover 100% of the API and they could use lots of warnings to force developers to slow transition.
Oh, the irony. If you actually look at those #ifdefs, the vast majority of them are #ifdef __WIN32__ - because Microsoft didn't really care about compatibility when they first adopted BSD code. Compared to Unix/Windows, the differences between BSD and Linux are almost nonexistent.
Wow. That's pretty ridiculous. So DirectX is allowed?
OK if you're a triple A studio with a cross-platform engine. Show stopper for most other devs.
DirectX11 with 2 compatible modes (D3D9 and D3D10)
Of course if you're doing anything interesting then it's probably a nightmare.. but the vast majority of applications aren't doing anything interesting, so Win RT is great for them.
Decoding does not do system calls, except for hardware decoding, so no worry.
File input is worrying because of the sandbox, but should be doable with some work.
But networking is a nightmare. 20 libraries are using WinSock2 (including some codecs like FLAC), plus libVLC's core. Sure HTTP access is fine, but RTP, RTSP, FTP, HTTPS, multicast UDP are weirder beasts...
To add it to the pain, VLC is a mostly C99 project, but C99 is too new a standard for Microsoft to implement it :)
Good luck with it – I’m sure you know it’s one of the Killer Apps on any platform.
Considering the last time MS released a C compiler I can fully believe it.
GCC has recently migrated to C++ and CLang is developed in C++.
So Microsoft reasoning is that C89 is good enough for them, and you are free to buy a compiler from another vendor if you really want to use C99.
Hah, better than that, Herb Stutter explicitly stated they don't give a flying fuck about C99: http://herbsutter.com/2012/05/03/reader-qa-what-about-vc-and...
> If you really need [...] features in C95/C99/C11 that are not part of ISO C++[, use] a different compiler
I read in a lot of places recently (this year) that XNA is 'dead' (search Google for instance). Is this true? What replaces it?
If you want to see what's possible you can check them out:
1) Adlib: http://apps.microsoft.com/webpdp/en-US/app/adlib/9cc0809e-b0...
2) Petunk http://apps.microsoft.com/webpdp/en-US/app/petunk/c03ecbca-4...
You can also read a great deal more about the process as well as watch me convert a game live here: http://www.reddit.com/r/windows8/comments/10g4r2/check_out_m...
I am in this very situation, with an app with has dozens of millions of users on Windows (and millions daily).
The first reason is the most important. WinRT requires a lot of rewrites of your code. (and the rationale behind some limitation are often political), and Microsoft is not helping or making it easy for you Notably for networking and multimedia.
Networking: everyone uses the BSD sockets, Win32, Linux, OSX, iOS, Android, Symbian, Bada, etc... And you cannot use them on WinRT. There is a very high chance that your code or one of your library uses it, of course. But the bad part is that Microsoft does not make easy with a wrapper exposing WinSock-like API based on their new API. Not even a wrapper than does 80% of the needs.
Multimedia and Games: your videogame that works with a C++ engine and OpenGL ES for casual games that works on iOS, Android, Bada,... does not work under WinRT, because OpenGL is not allowed... And I don't even speak about Video acceleration.
Yes, they need to do a massive change in their APIs, but helping the transition would have been a good idea...
So yes, writing yet another Twitter client is easy on WinRT, for more complex applications, well, it is way too hard and difficult to do it this fast.