```suggestion for use in robot-agnostic packages. ```

```suggestion This REP aims to define best practices for robot-related ROS parameters ```

```suggestion `robot_name` stores the robot's unique name. It should be set to make sure that this robot instance is unique on the system, in the same way that computers are recommended to have unique hostnames. For the case of PR2 robot, the default robot name is 'robot' [1], but other individuals have named 'james', 'gutsy' etc... ```

```suggestion The `robot_type` should capture the name of the class of robot in a way that a user would understand the scope of the robot. If it's a commercial product it is recommended to use the production model name. For example 'pr2' in PR2 robot case [1], 'pepper' in Pepper robot case and 'fetch' in Fetch robot cases. ```

```suggestion the URDF format. ```

```suggestion parameter server, helping applications that use these parameters to more easily run on a greater range of platforms. ```

```suggestion among the ROS community. This REP proposes similar conventions for robot-specific information to be stored on the ```

```suggestion As the number of ROS-supported robots increases, we will write ```

You changed one of the two instances of `robot/info` to `robot_info` but I still think that if we're not going to define any schema there's no point in including it in this standard. That's not to say it can't be there, but until we can desctibe what goes inside it then we shouldn't try to add it here. We can always amend/extend/replace this in the future when it's clear what should be in that field.

Might I suggest adding something to the effect of `It should be set to make sure that this robot instance is unique on the system in much that computers are recommended to have unique hostnames.`

Can you add a justification for lower case? It seems that `PR2` would be fine? I could also see manufactures considering adding some level of versioning in here. For example http://library.isr.ist.utl.pt/docs/roswiki/Robots(2f)Kingfisher.html vs https://robots.ros.org/kingfisher/ would likely want to be distinguished as something like `Kingfisher M100` vs `Kingfisher` (though it does appear to be renamed the Heron for disambiguation now)

It would be good to define the type more clearly. One example is not enough to clearly define what should be here.

I don't think that this is currently defined well enough to be useful. I see the name and type as equivalents to `lsb_release` type of information. It would be easy to have a command line tool that just lists robots on the system and displays their name and type. But the info could be printed, but not much else. If we want to keep this field we'd have to define some level of schema to make it useful.

I think that the namespacing here doesn't really make sense. To be able to introspect/discover this field a longer /more unique name would be better. I'd suggest `robot_name` and `robot_type`. This also gets rid of the ambiguity of what namespace it applies to. The keys will be in the root of the namespace of the robot. So if you're searching for multiple robots, you don't have to look for a `name` field, and then traverse upward looking for a namespace `robot` and then traverse upward once more to find the `robot_description`