HPRC tells me that their definitions for the ANSI standard's
'implementor-defined' file status codes 9x are the same as FSERRs.
Looking at some of the FSErrs, I am less than convinced. Is this in fact
the case? All I am concerned with is what I can get with ANSI85 COBOL;
we are not calling intrinsics, were are using OPEN, READ, WRITE,
CLOSE... I am used to documentation telling me that 22 is for random
access or relative and indexed, but not sequential files, or that I will
get x=123 when doing something wrong with an xSAM (KSAM, ISAM, VSAM,
IAMSAM SAMIAM) file.

Several places in the manuals refer me to "Chapter 5 of the HP COBOL
II/XL Programmer's Guide", which states:
"The x is an ASCII character whose numeric code is an integer between 0
and 255, inclusive, and represents a file system error. For more
information, see the[REV BEG] MPE XL error message catalog or the MPE XL
Error Message Manual (Volume 1 or Volume 2).[REV END]"
Not exactly up to date, nor do I find it very helpful. The MPE XL Error
Message Manual doesn't mention which ones I should concern myself with.
I have browsed *some* of COBCAT.PUB.SYS; my eyes glazed over. I did grep
and find some entries similar to "641 FILE IN USE BY ANOTHER PROCESS
[9x] (COBERR 641)". OK, that's closer to what I want, except what would
the value of "x" be in this case? 641 is not "between 0 and 255,
inclusive".