On 05/06/2015 20:09, Andrei Alexandrescu wrote:
On 6/5/15 12:02 PM, Bruno Medeiros wrote:
The traction and support that Rust gained, even before 1.0, showed that
having your primary toolchain based on GCC or LLVM is the only
sustainable way forward
Well I'm wondering how that inference works. -- Andrei
Indeed I wasn't clear on that reasoning. It's not the traction that Rust
gained that is the *cause* for LLVM "being the only sustainable way
forward". Rather, it's simply what showed me that this was the case. Let
me be more clear.
First when I said "traction and support", I wasn't referring to the Rust
user base, but rather the number of *paid* developers[1] that Mozilla
has working on the Rust project. Besides the core team, they also seem
to be willing to also hire programmers on a contract basis to work one
certain aspects of the toolchain (see for example:
https://michaelwoerister.github.io/2014/02/28/mozilla-contract.html).
But more importantly: when I finally started looking into Rust, was also
what caused me to look into LLVM properly, particularly by subscribing
the LLVM weekly newsletter. And that's when I gradually realized all the
work LLVM having done, with so many people working on it.
The recent news that Microsoft is going to work on a LLVM based compiler
(http://developers.slashdot.org/story/15/04/14/1529210/microsoft-starts-working-on-an-llvm-based-compiler-for-net),
is another sign of this tidal wave we are seeing in the horizon: the
future for non-proprietary languages is to be based on LLVM or GCC, or
they won't be able to compete.
And toolchain always wins. If won't matter if the language is superior.
[1]: Note the point here isn't that paid developers are better than
volunteer ones, the point they will usually will have more time to work
on things than volunteer. The point is about the time and effort they
are able to dedicate, not that they are paid per se.
--
Bruno Medeiros
https://twitter.com/brunodomedeiros
