I'm confused by this discussion. A PDS and a library in my experience are the same thing with different names. It's like car and automobile. A source library may be more frequently referred to as a PDS while a load library is more frequently referred to as a library. Either way they are both partitioned datasets designed to contain multiple, uniquely named "members".
A PDS or library may have differet attributes. Typically a source PDS has a fixed block record format (RECFM=FB), perhaps with a logical record length (LRECL) of 80 and an SMS optimized block size (BLKSIZE) of 27920. Typically a load library uses unformated records (RECFM=U) and a block size of 6160.
Joined: 14 Mar 2007 Posts: 8657 Location: Back in jolly old England
major differ is pds will lead to SE37 after the space declared is allocated fully whereas in a pdse u wont be getting a SE37 error.
I think that you may be getting confused with dynamic directory block allocation rather than space allocations here.
A PDSE is like any other dataset that will abend once all of its allocation has been exhausted. The fact that a PDSE is reorged automatically each time a member is deleted does tend to lessen the frequency of any space related abends.