J3/01-223 Date: 21 May 2001 To: J3 From: Dan Nagle Subject: Issues 127, 314 & cetera While reading Ch 9 in pursuit of Issue 127, I found the following which resulted in comments: (Edits to follow in another paper or a revision of this one.) In Chapter 9, 4th Paragraph: Does WAIT have a category? Should it? 5th Paragraph: record files are composed of records, stream files are composed of a sequence of file storage units. Surely similar language should be used in both descriptions. 9.2.2.3 unformatted (2) & formatted (4) Add: There may be a last fsu.? 9.2.3 "first record or file storage unit" is unsymmetric try "first record or first file storage unit"? likewise "last record or file storage unit" is unsymmetric try "last record or last storage unit"? Solutions to J3 Note 127: There is a logical eof fsu (not preferred) -or- There is a last fsu; reading any fsu after the last one is eof (preferred) (unless you want reading _way_ past the last fsu to be an error?) Also: Is there a eof when the file is unbounded? for example, a keyboard or an external device which simply produces fsus? a REWIND, BACKSPACE or ENDFILE following a partial record write should complete the record. 9.4 pg 171 Note that this means that INQUIRE returning -1 via the NUMBER= specifier is affected. 9.5.4.4.3 io_mode is mentioned (in Ch 9) pg 194 and in the arg lists of the DT procedures pg 192 UI 295 (9.5.4.4.3) Is 12.3.2 the right reference? pg 205 UI 308 an Interp is beyond this paper Edits to 1.5.1, add a paragraph something like: "previous standards supported a PAD= specifier on data transfer statements with a default value of NO, if INQUIRE for an unconnected unit, NO was returned. Now it's UNDEFINED"