Re: Why are Ada compilers difficult to write ?

["Dan'l Miller" <optikos@verizon.net>]

On Friday, June 29, 2018 at 11:36:07 AM UTC-5, G. B. wrote:
> Dan'l Miller wrote:
> > C++ is close to surpassing Ada as the programming language with generous
> > amounts of concurrency built •directly• into the language, not such
> > rinkydink library wrappers as it was formerly.  Wake up.
> >
> 
> Also, cleverly the language seems to continue its tradition of playing into
> the hands of programmers who enjoy building a multitasking solution rather
> than using one.

Very astute observation.  Perhaps the Ada community can utilize this as a differentiator/path-to-victory in the forthcoming explanations'/articles'/books'/bake-off-competitions' battle of the minds for mindshare.

One of the weak links in the chain of C++'s forthcoming new era of concurrency in the language is how precisely it fits with each OS's differing existing practices.  Programmers go through crawl, walk, run phases of utilizing new tech.  At the ‘crawl’ level, C++'s new threads support and multireader-uniwriter locks are obviously mapped to POSIX equivalents.  But on the other ‘walk’/‘run’ extreme, it is not clear at all how all of C++'s new concurrency-in-the-language features fully map onto, say, Apple's Grand Central Dispatch concurrency, if at all.  I predict (as G.B. seems to, as well) that C++'s roll-your-own approach is going to get overwhelming for the average programmer to get entirely correct in anything other than the plain-vanilla cases at the ‘crawl’ stage.