question about organized point cloud

[Zhang-Qinghui]

Please complete the following information:

• OS and Version: Ubuntu 16.04
• ROS Version: Kinetic
• Built from Source
• Version: commit 7208295

Describe the bug
hi, organized clouds could be accessed in a matrix fashion and I have tried, but I have one thing don't understand. Does the point under the same index number in the ordered point cloud of different frames come from the same position in the actual space?

the picture shows the point at Column 8 line 180 of pointclouds from the different frames,and the velodyne is fixed, the environment is constant. But the point data is different. Shouldn't the data be the same？ thank you in advance.

Expected behavior
I want some explanation about the problem.

[JWhitleyWork]

@Zhang-Qinghui - There is an inherent problem with the current driver's implementation regarding the size of a "scan" - see #285 for explanation. The organized point cloud feature can only organize the data that it is given and, since the driver is inconsistent on the size of a "scan," the precise location of a point within a scan is currently non-deterministic. If you would like to work on this PR to fix this issue, help would be welcome.

[Zhang-Qinghui]

Thanks for your reply. I think it's because the rotation speed of the radar is unstable.This is a hardware problem.