Compressibility has a lot to do with repetition of patterns and long sections of the same byte, so far as I know.
So source can compress very well, due to all the white space we often have on each record.
The same for text fields in physical files in general. Now varying fields MIGHT not compress as much, I'm just thinking, if their allocated length is fairly short, since the exact data is stored unpadded in an auxiliary space.
*PGM objects won't compress much, typically.
Just some thoughts.
----- Original Message -----
I totally get that. I suspect that on my machines that are primarily
stream files that a bulk of them are Domino .nsf files and I hope that
they try to keep those rather compressed.