And my MF guru's response:
No, haven't heard that before, but he probably was referring to "Undefined File Access Method", in which the file has attributes of "undefined" (not fixed length, not variable length with a length code preceding each record to tell you its length, but rather just a bunch of bytes). You have to do a read of a block of data, then interrogate the length code returned by the operating system (it tells you how many bytes were read),
Either that, or it's a specialized proprietary access method used by some third-party software product. IAM, for example, is Innovation's Access Method that they use for their dump/restore product. See if you can find out if this is a vendor-supplied product.
It's a bag-o-bytes with custom code that handles it. Or it is something so obscure it is not on the web anywhere. It certainly is not listed in any detail on IBM's Redbook (core docs) site.
Of course custom code beats any 3rd party indirect access method, be it DB2/SQL or ISAM, etc. All packages will sacrifice something for portability or supportability or ease of code.
That would be like me saying I've just spent the last 10 years custom hacking a C program that has indexes tuned to my style of data being far better than Oracle simply because it is faster for a single well defined activity that I've been coding to for the last 10 years.
Just silly!
