About the reason why we should use conda-pack

[석진환]

Hi everyone,

I have a few questions on the reason why conda-pack is necessary.

I usually work in a Spark environment on YARN and come across cases where I have to distribute python packages. 

So far, I've zipped the virtual environment directories and then distributed them, but I recently learned about the existence of `conda-pack`. I'm curious about the difference between my approach and conda-pack. 

The official document say that the conda is not relocatable, but I don't understand what you mean, so can you explain it with a simple example?

Thanks in advance!