Hats off to the senior guys here who clarified a lot of things.

Well - One last question before I go ahead and implement LMT's in my production database.

1. Since my schema contains tables of diferent sizes ranging from 0.5M to 100M (Database size is roughly 10G), shall I use the autoallocate option or the uniform size.

2 - When I tested autoallocate, the extent sizes were so different for each segment (first 8M, second 1M etc).
Will this not cause fragmentation again?

3. I am planning an export, creation of LMt's and import of data into that LMT? Is that the right way to go about it?