Same keyword 'domain' is used for both ROS_DOMAIN_ID and DOMAIN_NAME

[taken-yjyoon]

I was looking for a method how to use domain names not an raw ip address. And I think there is a problem. Same keyword domain is used for both ROS_DOMAIN_ID and DOMAIN_ADDRESS.

In ROS2, we have ROS_DOMAIN_ID, which is an integer used to separate communication networks between nodes that shouldn't interact. On the other hand, for internet, we use domain names (like www.google.com) as string-based addresses so people don't have to type IP addresses which is hard to remember (like 74.125.200.100).

Reading through the documentation, it seems this program doesn't distinguish between these two concepts. The keyword domain is used ambiguously: in some contexts, it refers to the ROS_DOMAIN_ID (an integer), and in others, it refers to an internet domain name (a string).

This ambiguity leads to confusion for users and developers. For instance, since one is a number and the other is a string, especially in a C++ program, I can see this causing significant problems in the future. It's already difficult to search the documentation when trying to find information specifically about one versus the other.

For me, reading or creating configuration files is confusing because the same term is used for two very different things. While it might be understandable when actively reading the documentation, or in the zone, but it's easy to forget or mix them up later. I can see myself accidentally do the opposite.

At a minimum, this could cause the program to fail. More seriously, mishandling type mismatches, particularly with how strings works with C++, could potentially introduce security vulnerabilities if not handled properly.

I wish you use distinct terms for these very different fields, such as ROS_ID with domain_name, or ros_domain_id with dns_name, or the other clear combination.

In best regards, YJ.

If you want to see where I came to this reason, see 3.5. Participant Configuration of the document.
More specifically, 3.5.1. Domain Id and 3.5.7. Network Address

[juanlofer-eprosima]

Hi @taken-yjyoon, thanks for sharing you thoughts. We understand your concern and will take it into account. However it seems unlikely we will be addressing this in the short term.