the soul of a refreshed Ada momentum

the soul of a refreshed Ada momentum

[Optikos]

(Apparently due in part to my match-making between Microsoft's C++/WinRT and Microsoft's (nee Xamarin's nee Mono's) Embeddinator-4000 and my related PitA nagging mid-2018) Microsoft is expanding its C++/WinRT endeavors to not be merely C++ and not be merely WINMD (Windows meta-data).  The generalization is called xlang, which might be pronounced ecks lang or cross lang or trans lang or don't-care lang; I don't know which one, (but I am rooting for either cross lang or trans lang).  The following directory contains 8 design notes that at first blush might be interesting for Ada community to participate in at this early stage to try to get more automatedly-produced API projections into the Ada language, or more {C++, C, C#, Python, Rust}  API projections of well-crafted Ada libraries.  Major portions of this project were pushed by internal-Microsoft to public GitHub just 8 days ago, i.e., the portions that have the check-in comment "Final C++/WinRT project structure for OS ingestion (#193)".  This is fresh raw red meat thrown to the programming lions to devour.

https://github.com/Microsoft/xlang/tree/master/design_notes

In particular in XDN03, it seems that xlang would do better to have Ada's more-visionary approach to types & subtypes instead as the lingua franca (or PL/1's approach to building up all types from bit).  At the very least, Ada community should embrace xlang enough to steer how Ada's more sophisticated approach to typing and subtyping and range-restriction would map onto xlang's more C-ish (DEC PDP machine-word) oriented types, such as a range-restricted int mapping to C++ parameterized-type integer-type with lower & upper bounds as integer parameters to the type.

Opportunities like this to expand Ada's marketshare into non-Ada programming-language territory come around only once per generation of human beings or so.  Not since the 1990s-era dual opportunities of 1) Ada bindings to POSIX and 2) the out-reach to unchecked C constructs in Ada95 has Ada had even the remotest high-visibility opportunity to tout its wares directly to all non-Ada programmers in any kind of lingua franca that tears down the barriers between programming languages, so that the best language for the task gets utilized, or better yet so that the barrier is lowered for surrounding legacy nonAda code with better modern Ada code as part of transforming a system to eventually incrementally chip away at the nonAda guts.

At the very least, Ada community should watch xlang to see whether xlang is good enough to interop with or a stinker that should be done better via an Ada-centric replacement for xlang.  Food for thought for what a refresh of Ada momentum could look like.

the soul of a refreshed Ada momentum

[Optikos]

This xlang might be a good platform for automated façade code generation in language m to •hijack• libraries in language n, so that m is not hampered by the lack of n’s rich trove of libraries, where most notably m=Ada and n∈{C,C++,C#,Rust,Go,Python,ObjectiveC,OCaml/Coq}

Many people nowadays seem to choose programming languages for the repertoire of libraries, not the core language.  This xlang automated API/ABI projection might be a game-changer by divorcing library lock-in from choice of language, just as its core/predessor (C++/WinRT) did for C++ vis a vis C#.  Think of xlang as a potential golden era of taking library lock-in shackles off.

Ada in the 2020s:  the great hijacker language

Re: the soul of a refreshed Ada momentum

[alby.gamper]

On Thursday, March 14, 2019 at 8:17:54 AM UTC+11, Optikos wrote:
> This xlang might be a good platform for automated façade code generation in language m to •hijack• libraries in language n, so that m is not hampered by the lack of n’s rich trove of libraries, where most notably m=Ada and n∈{C,C++,C#,Rust,Go,Python,ObjectiveC,OCaml/Coq}
> 
> Many people nowadays seem to choose programming languages for the repertoire of libraries, not the core language.  This xlang automated API/ABI projection might be a game-changer by divorcing library lock-in from choice of language, just as its core/predessor (C++/WinRT) did for C++ vis a vis C#.  Think of xlang as a potential golden era of taking library lock-in shackles off.
> 
> Ada in the 2020s:  the great hijacker language

An interesting set of articles on xLang can be found here 

https://kennykerr.ca/

Alex