Porting all of ROS 2 to rust

edit

So why can't we start off with rust?

I won't answer this, as it would seem to go into the direction of "why tabs, not spaces?", but as far as ROS 2 compatibility with Rust, please see Announcing Rust bindings for ROS2.

Even for ROS 1, there are several implementations of client libraries for Rust and Go-lang.

For a more generic overview of the available client libraries for ROS 2, see Home > ROS2 Overview > ROS 2 Client Libraries.

Edit:

Also anyways we are building ROS2, and hence rewiring most of the other software's too to work with it, [..]

This may not be apparent, but ROS 2 has been in development for quite some time already, with a significant code base developed, maintained and released for quite some time as well. What you're suggesting is to essentially throw all of that away and start over (or perhaps do some conceptual reuse).

From a plain resource perspective (ie: time, money) that wouldn't be feasible I believe, as nice as rust (or <insert-other-language-here>) is.

The two industrial projects I worked before that called themselves "ROS centric" mostly used timer and messages. They also used some of the ROS infrastructure for building everything with the single command and running together multiple nodes.

Some other parts traditionally provided by ROS (especially the logger and the configuration parser) can be replaced by dedicated open source libraries. ROS implementations are not unique or significantly better.

That much can probably be replicated in some independent project. C++ integration is still vital as some nodes were interfacing various devices via they C based SDKs.