On 09.04.2012 20:39, Sönke Ludwig wrote:
(IMO) one of the biggest obstacles for truly broad adoption of D currently is the weak platform support on end user platforms. The two mobile platforms that came up recently (iOS and Android) are two examples. And indeed I think that support for mobile platforms could be a real stepping stone because of D's extraordinary convenience and language power - the alternatives to C/C++ are pretty thin here and cross-platform development in general has come to a grinding halt recently with all the proprietary languages and APIs. If D could step up here...
But mobile platforms aside, Windows support is something that in general has always been neglected a bit, especially regarding 64-bit support. Starting with Windows 8 there will arise additional problems because Metro application will only be able/allowed to use the COM based WinRT and the VisualStudio runtime. DMD with its use of snn.lib is out of the game here, just as the any other runtime library.
Not true at all, in every talk I've seen on WinRT so far C++ CRT is still shipped side by side with WinRT. Basically every language has his own runtime. It wouldn't be Microsoft if they haven't got a solid reserve of backwards compatibility. Simply put WinRT is a major update on COM technology and even here it's backwards compatible with the old COM. The fact that OS API is expossed through this new COM interface is just a nice feature. I was kind of wondering when they will finally ditch Win32 API.
Right now, if we don't catch up here, D will slowly degrade to a pure server and command line application language which surely wouldn't do it justice. In consequence this means that there is one more reason to raise the priority of COFF output from DMD (together with 64-bit codegen) - or possibly the alternative to make OptLink COFF-capable to at least be able to somehow link against the VS runtime. Another such thing - although this can be worked around - would be direct support for Objective-C classes like in Michel Fortin's dmd modification. I think these GUI application related functionalities are by far the most important things for D's mass adoption. And personally, I would even be willing to donate a (for me) considerable amount of money to help bringing this forward because many things I would like to realize with D are currently (almost) impossible.
-- Dmitry Olshansky
