Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.compilers > #2380 > unrolled thread

Braceless Compilers and Languages Forum

Back to article view | Back to comp.compilers

Contents

  Braceless Compilers and Languages Forum Mikael Egevig <mikael@egevig.org> - 2019-11-22 12:41 +0100
    Re: Braceless Compilers and Languages Forum rockbrentwood@gmail.com - 2019-12-06 18:33 -0800

#2380 — Braceless Compilers and Languages Forum

Hi fellow compiler enthusiasts,

I'd like to introduce you to the new Braceless compiler and language
forum at https://forum.braceless.org.

The forum exists both as a forum for the "dream" future language
Braceless and for general announcements and discussion of programming
language features and compiler implementation.

If the forum turns out to be a success, I'm ready to keep it online
for the coming decades just to provide a place where newcomers and
experts can exchange views and knowledge about programming languages
and compilers.

In particular, it is my hope that experts will help us newbies by
revealing how to implement advanced language features, things that are
only very rarely covered by textbooks on compiler construction.

The forum is only a few days old so there's not much going on at this
point in time, but feel free to register and post.  If you need
additional categories or the like, just let me know and I'll be happy
to add them.

NOTE:
We will not use your email address for anything beyond registration,
logins, password recovery, and perhaps a ban if you post spam or
deliberately act as a troll.


Sincerely,
Mikael Egevig

[toc] | [next] | [standalone]

#2383

On Saturday, November 23, 2019 at 2:27:42 AM UTC-6, Mikael Egevig wrote:
> Hi fellow compiler enthusiasts,
>
> I'd like to introduce you to the new Braceless compiler and language
> forum at https://forum.braceless.org.
>
> The forum exists both as a forum for the "dream" future language
> Braceless and for general announcements and discussion of programming
> language features and compiler implementation.

You'd be well-advised to read the design philosophy discussed by Stroustroup
in his article on the development of C++; particularly the issue of "dream
languages"

http://www.stroustrup.com/hopl2.pdf

More importantly, as you mention this being a forum meant be around
"for decades", you'd be more strongly advised to watch the PBS
Frontline 2019 Episode 17 ("In the age of AI") (currently here:
https://www.pbs.org/video/in-the-age-of-ai-zwfwzb/ also on YouTube).
There is a good possibility that the tidal surge described in this
episode - all of which is now on your doorstep - will sweep away all
considerations of programming languages and programming language
design; all of which are largely superseded by the emerging AI
paradigms.

The one thing I can tell you with certainty - which even the PBS
episode did not take into account - is that the wave of automation
will sweep over the tech sector -- including even (and especially) the
fields of Software Engineering and Artificial Intelligence themselves;
rending both work/job/livelihood and business/marker/commerce in these
fields superfluous.  In other words, even the the cast of "big
players" in the PBS episode are going be swept under by the tide.

Program synthesis, program design synthesis, program normalization,
refactoring, reengineering, recoding - these are all fields in and
applications of Artificial Intelligence, as well. The true future of
compilers and compiler designs lies with intelligent compilers; those
with sufficiently advanced intelligence that they can literally make
their own compilers (and languages and operating systems) and
implement them; as well as creating the specification and
implementation for other for software at both the system and
application level.

This all transcends programming languages and programming language design.

[I'm much less persuaded that programming will go away.  The point of
a programming language is to let you tell a computer unambiguously
what to do.  While I can certainly imagine areas where big data
approaches that synthesize the instructions by crunching the
environment would work, there are plenty of applications, particularly
financial ones, where there is one correct result and anything else is
wrong.  (I can tell you stories about programming bond yield routines
to match the synthetic 360 day calendar bonds use.)  That seems a much
better fit for programs that say exactly what to do rather than AI
that usually gets close enough. -John]

[toc] | [prev] | [standalone]

Back to top | Article view | comp.compilers

csiph-web