kiss-icp depends on plyfile (MIT vs GPL v3)

[Samahu]

Hi,

During a review of the software licenses used by some of our projects we noted that kiss-icp list plyfile as one of their dependencies:

Kiss-ICP is listed as an MIT license but plyfile is GLP v3 as shown in the table below:

 kiss-icp            1.1.0       MIT License                                        
 plyfile             1.0.3       GNU General Public License v3 or later (GPLv3+)    

This may be seen as conflict between the two licenses (depending on how you interpret GPL, I am not a legal expert). So we were wondering if that is something you guys already went over and decided that the use of plyfile doesn't infringe the kiss-icp license.

[nachovizzo]

Hello there! thanks for pointing this out.

Indeed, none of us is expert in licensing. We will try to investigate a bit how to solve this, but meanwhile

• the plyfile library is only used in the "python" interface, in case you are only using the C++ core library form kiss-icp, or the ROS wrapper, none of this code is being used. So I'd assume we are safe on that matters
• if you are using the python package, but you have your own dataloaders, then you can delete the pairs luco dataloader, which is the only one which uses this library (

  kiss-icp/python/kiss_icp/datasets/paris_luco.py

  Line 28 in 19e57cf

  from plyfile import PlyData

  )
• @benemer , @tizianoGuadagnino 3rd option would be to remove the dataloader completely or remove the dependency, I think we did this in a rush and probably can use any of the other 3rdparty dependencies to load this data, wdty?

[tizianoGuadagnino]

I vote to completely wipe out the dataloader guys . I honestly didn't even remember it was there. Probably we use it twice in our lives. @Samahu, thanks a lot for pointing this out

[benemer]

I am also OK with removing it. We do not even report the results in the paper. Good catch @Samahu!

[tizianoGuadagnino]

Handled in #426