.bk A9
.fl K730jlw3.j
.fd Documentation of problem with "E" Program
.ei jlw
.ed K730
.rd K730
.ri jlw

-pr While processing the journal files for MZ10A9, several problems involving the "J" and "E" programs were discoved.  Among others were that a file that satisfatorily passed the "J" program's requirements would not run on the "E" program.  Most of the problems were those of formatting.  For example, the J program accepted a typical entry of "ei jo" vice ".ei jo", while the "E" program rejected "ei jo".  Since the "E" program does not precisely specify the nature of the errors it finds, it is much harder to identify and correct them at that stage.  There are two systematic problems in the "E" program.  First, some files with no apparent defects, are rejected when run the first time.  If immediatedly run again, they will be accepted.  Second, in the rare event that the first feature and the first aggregate are documented in the same file, data for both a1 and f1 will be entered in the "A" file and there will be no data for f1 in the "F" file.  A simple experiment that split the original journal file into two journal files that separated the data for a1 and f1, proved that the problem could be worked around and that once the output files were properly established, the "E" program distributed subsiquent data to the proper place.  Specifically for A9 file H620jo.j will be split into two files, K730jlw1 containing data for f1, and K730jlw2 containing the data for a1 and incidental entries.