Thibault,

Thanks for the review.  I've only been paying sporadic attention to the ROS2 process of late, and this was a useful reminder of the state of some things.

I'm looking forward to hearing more about the status of ROS2 next week at ROSCon and, in particular, it would be great to get a response from the developers on some of the things in this document.  As an example, rebuttal 1.1 says "This will eventually be okay when everyone uses ROS2.", which I believe, but claim 3.2 suggests that this will not happen for a "long time".  It would be more reassuring if I had some idea of whether a "long time" is a few months or several years.

Primarily, I worry about the community splitting in the time required to do the migration, and then never coming back together again.  My fear is that everyone will pick one version to work in, and it will lead to two communities (perhaps academic and industrial).  This would undermine one of the core strengths of ROS: it's community.

I'll also note that the word "hopefully" appears in 40% of the rebuttals.  Many of these have a claim of the form "I think that X will be a problem", and a rebuttal of "Hopefully X will not be a problem".  Hope, as Rudy Giuliani said, is not a strategy.

-- Bill

On Fri, Sep 25, 2015 at 11:49 AM, Thibault Kruse via ros-users <ros-users@lists.ros.org> wrote:

Hello all,

I would like to present a review of the strategy taken to create ROS2.

Recently OSRF announced the release of an 'alpha1' ROS2 milestone [1].
The list of missing features is still quite long. That also means some
opportunity to still influence decisions.

OSRF has been promoting ROS2 at ROSCon2014 [2], and provides
documentation [3]. As a reminder, major goals include improving real
time robotics, embedded robotics, Windows-compatibility, messaging
over unreliable networks and multi-robot scenarios.

All changes come at a cost, there are tradeoffs to be made. I have
initiated several discussions in the NG mailing list [4] to preview
the impact of ROS2.

The short version is that currently ROS2 has completely separate
sources and requires different core tools (e.g. a buildsystem that is
not compatible with catkin), and many APIs have breaking changes. The
migration to ROS2 will take similar effort as migrating all ROS
packages to a different middleware. A long transition period is
likely. Supporting packages in parallel for both ROS1 and ROS2 will be
very hard. Because of the lack of backwards compatibility, the
transition to ROS2 will probably be a large disruption to everyone
using ROS (https://i.imgflip.com/rl3g1.jpg).

The long version is here:
http://wiki.ros.org/sig/NextGenerationROS/StrategyReview

I announced that wiki review page one week ago on the NG mailing list
and tried to include feedback. Thanks to all who gave feedback.

Please use the NG mailing list for feedback about ROS2:
https://groups.google.com/forum/#!topic/ros-sig-ng-ros

regards,
  Thibault



[1] https://groups.google.com/forum/#!topic/ros-sig-ng-ros/B4BAQY5c3xs
[2] http://www.osrfoundation.org/wordpress2/wp-content/uploads/2015/04/ROSCON-2014-Why-you-want-to-use-ROS-2.pdf
[3] http://design.ros2.org/
[4] https://groups.google.com/forum/#!forum/ros-sig-ng-ros
_______________________________________________
ros-users mailing list
ros-users@lists.ros.org
http://lists.ros.org/mailman/listinfo/ros-users

_______________________________________________
ros-users mailing list
ros-users@lists.ros.org
http://lists.ros.org/mailman/listinfo/ros-users