rodrigovaras
commented
6 years ago Hey, I'm glad you get interested in this project, it is truly a gem, we spend many resources/hours on it. I architect the idea from the ground up and we spend many years perfecting the final goal and components. About your questions:
-
I would say this is Version 3 of the original product. The V1 was used (and is still being used) by the Skype Of Business (former Office LYNC). Large parts of that project heavily use a very similar runtime to power the Model and View Model objects that the Office UI render, the project was a complete success ! . V1 only runs on Windows and desktop machines. The V2 & V3 were planned to run on phones devices and be cross platform without loosing the Windows API core strategy. The V2 was intended to be used on the next generation Skype clients, but the project was cancelled at the last minute in favor of other alternatives outside the native approach.
-
About the format of the interface and flow, yes the Dijinni is similar but there are important differences. The pmod is based on the intrinsic support of 'reflection' on native objects which can be projected to different languages and platforms. In Dijini the code generation for java is C++ JNI code, with pmod you never see a piece of JNI code again. The pmod architecture allow different layers of code generation. Also Djini does not support any 'observable' pattern and in pmod most of the important objects already have this idea embedded, The pmod allow to fire events from the native layer into the UI layer, etc.. pmod support Windows platform E2E, including desktop, UWP, etc.. pmod also have an experimental support on Node, etc.. Djini is probably lighter in terms of runtime size/memory requirements.
-
Roadmap: I leave the group that was building this idea and most of the people are doing different things. I'm right now on the Visual Studio team doing other stuff, but pmod is my baby. I'm committed to maintain and invest more on this as long as people in the community want it. It was already super hard to get authorization to publish this as an open source project, i move mountains to make it happen. But this project require more people involved, the important pieces are there but the community needs to contribute to make it great,
Things on my mind that i have planned already:
- Invest in a better description language, maybe move away from C# and create a better and less buggy language parser.
- Better authoring library: the learning curve is not trivial, a better and simplified model would facilitate without the need to understand boxing/unboxing and reference count objects.
- Factory objects: allow the integration on how to create new objects. pmod is great once you have a the first reference. We want something that allow consuming languages to create new instances in a natural way, for example in C# var myModel = new MyModel("foo", 200);
- Better code generation for simple scenario that won't require any observable object but just a super simple projection which may people face today. ( i have a lot of C++ code that i need to be visible in other language)
Hope this helps and ready to answer more questions.
bilsou
Just found out about this gem and in all fairness I was surprised it did not get more attention, it looks really good. I have been through the FAQ but I wanted to clear some things up.
pmodbeen used in any production app(s) (or internally at Microsoft) ?pmodroadmap ? Will new features/bug fixes being introduced in the future ?Thanks in advance for your answers.