Open davetcoleman opened 11 years ago
+1
+1
I looked at the XML for the transmission stuff. Has this changed at all from the format we had a while back? There are two options as I see it:
I expect (2) will take a while (at least a few weeks), so my suggestion would be to do (1) for now. If the XML format for transmissions is stable, we can start on (2) as well.
To clarify, option 1 is basically do nothing and leave things the way they are. We are already using
Just today there was some discussion on the controls mailing list about modifying the transmission format to allow multiple hardware interfaces to be supported. So I guess we aren't completely stable yet. However, in the current parser I made it load the unrecognized elements into a tinyxml variable so that its easy for custom tags to be added for more complex transmissions.
On this note, I was looking into the integration of my transmission parser into urdfdom and urdfdom_headers and I noticed in urdfdom_headers that there is currently no tinyxml dependency. Would it be ok to add this so that the data structures can store the custom tags as I just mentioned?
Overall, I think we should be planning to do option (2) as soon as the discussions on transmission format settles down.
+1, but not just yet. Urdfdom is a stable package (version 1.7.1), and I'd like to see the transmission parsing code get used a bit more before commiting to an API and having to tick-tock any changes.
I'm currently working with this code, and in fact I might propose (minor) changes that affect both the public API and the XML representation. These are changes that I had not foreseen in previous design discussions, but only became apparent when getting down and dirty with the implementation.
Meanwhile, if you try to parse an urdf with urdf_parser_py
it throws an error if your <transmission><joint>..
tags don't have a type
attribute. #36
As suggested by @mikeferguson, it might be more useful to move the transmission parsing code into urdfdom so that the xml standard for urdf is centralized.