Re: How to get Ada to ?cross the chasm??

[Lucretia <laguest9000@googlemail.com>]

On Sunday, 29 April 2018 01:39:10 UTC+1, pat...@spellingbeewinnars.org  wrote:

> Is Ada not catching on because of a GPL licensing issue? Really? I don't even know what is missing from the FSF version vs the Adacore GPL version, it's certainly not an issue for me, my main problem has been the limited libraries and the difficulty in interfacing with C to create the libraries that do not exist with Ada.

There are issues, but it's mainly when you hit a bug box in the FSF version which doesn't happen in the GPL version that it mainly becomes apparent.
 
> For Ada to succeed I believe it has to SCALE DOWN and be more friendly and accepting and we have to interface more easily with C.

But we also need people to stop looking at one piece of Ada and saying "ew! Pascal!" or "I can't read this, it's too verbose!" These are comments I've come across a lot. More people need to be more open minded about programming languages and get out of the "it has to have braces or I won't use it" mentality. To me, that is snobbishness to the extreme, it's also incredibly stupid.

> Again I am grateful for the help her but is this not a very judgmental place? If someone doesn't like the formatting of your code, your going to hear about it. Regular posters seem to be consistently engaged in posturing, trying to prove that they are better Ada programmers than the other regular posters.

Try hanging out on #C or #C++ (whatever their IRC channels are called), you'll come up against a lot worse people there.
 
> Ada is huge and then there is the standard library to learn and potentially another build system with GPR.

Ada is large, GPR is just one of many project systems that exist, GPR is specific to GNAT though.

> I am not using Ada right now but I just posted to the list about using shared objects because I am slowly trying to figure out how to write up a document(or possibly a book as I have a publishers account) that will introduce simple ways to mix Ada with C and to introduce a subset of the language to new programmers. We don't have a large enough community to support the libraries that probably a lot of programmers need. IMHO it's critical to be able to write your own bindings.

Binding generation can be very time consuming, see my SDL2 bindings. If I decided to do a quick thin binding, it'd feel all wrong, it would just be like writing C in Ada. I decided to take a mixed level binding approach, by adding thin where it doesn't really matter much, but also by making Ada types and mapping them to the C "types," i.e. the numeric types. For other things, they get wrapped in Ada and are higher level. The library isn't done and probably never will be as SDL2 is still evolving, and always will be.

> And there there is the hugeness of the language. How could any dentist or young person be expected to cover it all? I think it's important to learn a subset. I am personally trying to only use Ada83 and this is the subset I would like to eventually include in a book or PDF.

You don't need to cover it all, you can do a lot with types, subprograms, generics and packages. I haven't touched tasking since uni.

I think limiting yourself to Ada83 is not the best way, you miss out on various things added, i.e. interfaces for OO, extended return for assignment in place for limited types, just to mention a few. Also look into "pragma Restrictions" if you want the compiler to stop you using some feature.
 
Luke.