Subject: Re: [Boost-build] Son of b2 - suggestions
From: Robert Ramey (ramey_at_[hidden])
Date: 2016-10-29 12:09:15
On 10/29/16 4:57 AM, Paul A. Bristow wrote:
+100 Excellent, Excellent, Excellent. Should be read by every
software development making any kind of product or library!!!
I was tempted to make a point by point comment. But this is so close to
perfect that any attempt to improve would only dilute the power of the
original exposition. I'm guessing that in being associated with boost
14? years - this is the only time I've felt this way.
It also makes it pretty clear that the discussions regarding a successor
to b2 are focusing on issues which are not currently relevant and
destined to result in anything useful.
> PS I admire your courage - unseating CMake will be hard now. Son of b2 needs to be more user-friendly as well as better. A better
> (google-friendly) name than b3 would be good too, but I've no good ideas.
Modern software development is almost dominated by the task of composing
opaque components. The build system is a big part of this task. The
best attempt is still the original make. But no one has been able to
extend or replace make with something which addresses the current
"build" problem. I understand the appeal of CMake. It's better than
most alternatives - including b2. But it still has a lot of problems,
quirks and violations of the rules set forth by Mr. Bristow. So I don't
think that CMake will last forever. Someday, someone will step back and
realize that the task of software development is more than just
programming. It has to address
1) Some sort of formal idea of functional design which has yet to be
2) which borrows and reconciles ideas going back 50 years such as
program correctness/provability, literate programming, layered
abstractions, types, proper role of object oriented programming, etc.
3) and provides a basis for agreeing on what a good/bad program is.
I'm looking at the stuff we're trying to build with software today: self
driving cars, talking home appliances, automated medical diagnostics,
web applications which knit together billions of people, and I'm just
blown away that we've come so far with the crude tools we currently have.
As far as C++/Boost is concerned, I feel that we have a language which
can express just about anything. But we don't really know what we want
to use it for.
Sorry, I love to rant.
> Paul A. Bristow
> Prizet Farmhouse
> Kendal UK LA8 8AB
> +44 (0) 1539 561830
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk