Re: question on BLK vs. NOBLK compiler option...

Jon,

I appreciate your response. Let me rephrase my issue/question. Perhaps the 
compiler option BLK/NOBLK has nothing to do with what I am wanting to know.

We READ and WRITE quite a few sequential files (not indexed or relative) in 
many of our COBOL programs. Virtually all of the FD's for those files have 
the BLOCK CONTAINS clause specified. Many of these sequential files have 
variable length records.

My question is what is the best combination to use in the iSeries world?

1) Remove the BLOCK CONTAINS altogether and let the compiler block and 
de-block as it sees fit? Does it do a good job?

2) Retain BLOCK CONTAINS and use what was originally specified in the VSE 
world?

3) Retain BLOCK CONTAINS and but specify a new value that is more correct 
for the iSeries world? Is there a rule of thumb for this?