The idea is to reuse the parser in gazebo and other tools. The library will most likely be implemented in C / C++.
Consider an standalone RNDF viewer. Does it make sense to do a web viewer? The main drawback I see here is that we wouldn't be able to reuse the parser.
RNDF files should be generated in a separate file. Then we should have a gazebo model that points to that file to import it (similar to what Gazebo does today to import e.g. collada files).
Even though not in the near roadmap, always keep in mind that at some point we may need to move to Opendrive.
Define interface for exporter / importer and internally for the importer (parser / visualization).
Some things to consider: