On Mon, Apr 4, 2022 at 12:05 PM William Linkmeyer
It would be fitting, I think, to use boost-based projects for the website’s back-end I’m currently working on a boost::beast-based server. It hosts my website, but that’s not exactly mass-tested.
While I am flattered, trying to use Boost.Beast and C++ to build a modern website may not be the most efficient. That's a lot of new code which few people will understand. I haven't deployed any servers recently, but I suspect that the economical and pragmatic solution is to use off-the-shelf software parts that are widely understood and for which it is easy to find reasonably skilled maintainers.
I am curious in re. the thought that Boost is on a decline. If the point of comparison is high-impact library creation and adoption, then yes, maybe boost is on a decline. But, aren’t there only so many fundamental libraries to create — and doesn’t boost cover virtually all of those fundamental libraries?
Human creativity is essentially infinite, and software is infinitely frustrating, so I think that we will not be covering all the potentially useful libraries any time soon, if ever :) You do raise a good question, how do we measure the success of Boost? Is it the number of installs? The number of programs using Boost? The amount of participation on the lists? I think software development is an inherently social phenomenon. It is the vibrant exchange of knowledge and ideas that drives the state-of-the-art forward. Thanks