I am slowly getting a multi-vehicle (single master) fleet working under
ROS (Fuerte).
That has entailed a lot of work on the original code to make it work
under the new infrastructure, plus learning how all the various pieces
have to be chained together.
Just a little background in order not to appear as just a whining
student not able to do his own homework.
I started getting to know ROS around October 2012, and since then I
found the way information is conveyed and displayed to be far from perfect.
The beginner tutorials are one of the few good examples I found since
then: a comprehensive chain of ever more complex topics until it gets
you to a working node example being able to exchange messages.
That's about the last mildly inclined path, from there it's all but free
climbing without protection!
Build systems, transforms, publishers, urdf models...
Every argument has links on its page, but clicking a link is not equal
to learning.
A link means that an argument shares a connection with another, but this
connection */should/* be explained.
On the contrary the various arguments are just linked together usually
without any explanation as to how that link happens to be.
For example I had to fight an entire day to understand how the urdf
model, the robot state publisher and the various transform prefixes have
to share some commonality in order not to have unconnected trees or RVIZ
utterly refusing to show data because a topic is missing its reference
frame.
This is nowhere to be found (or at least I've never found it explained
anywhere).
And I'm still fuzzy about it, I wouldn't dare write a tutorial on it.
Besides I wouldn't even know what to file that tutorial under.
So I think that what we are seeing here is the usual very-open source
development problem: since there is (almost) no central control,
information generation and sharing is up to participants. Thus
information quality lacks.
I don't have a solution recipe, save to propose a little team to
(periodically) check the documentation status.
As a first step I suggest covering the basics in a *comprehensive*
tutorial that goes from installing the packages to have a full working
robot (even if only simulated).
Maybe an extension of the turtlesim with a simple urdf to also explain
tf trees and publishers would be nice.
My 2 cents
--
*Eng. Claudio Carbone
Embedded Systems Design*
P.IVA: 11688471009
tel: +393809017424
email: Send email
<
mailto:cla_carbone@tiscali.it>My linkedin profile
<
http://it.linkedin.com/in/embeddedesign/en>
My Portfolio
My portfolio site <
http://www.fusioncoredesign.it/>
(text/plain)