J3/13-349 To: J3 From: Malcolm Cohen Subject: Further editorial changes Date: 2013 October 08 1. Introduction This paper contains additional suggested editorial changes to 10-007r1. 2. Discussion The way that constraints sometimes have or not have a syntax rule qualification is confusing. In my opinion it is better for the constraints to be written in such a way that qualification is unnecessary. 3. Syntax rule cross reference This is completely broken in the published standard; for example, does not appear in C509. In 13-007r0d0 this reference has changed to C507 which is also wrong, the actual constraint being C4107. In 10-007r1, access-id is cross-reffed to C563 but it is actually C562 (13-007r0d0 gets this right). Other obvious mistakes: (a) is not cross-referenced at all; (b) is cross-referenced as only appearing in C852, but it also appears in R850; (c) is reffed in R438 but is xreffed as only appearing in C462; (d) is also not xreffed correctly. Since no-one has ever complained, my conjecture is that this is largely unused. It is certainly not useful to have a broken cross reference! This is produced by a short perl program. If no-one steps forward to read this (it looks utterly trivial) and find out why it is producing garbage, we will have to delete D.2. I can see why C4107 turns into C507 (4*100+107), but not why other things are not appearing. 4. Edits to 10-007r1 [throughout] Index and hyperlink every statement reference and requirement. {These are frequently not indexed and/or not hyperlinked.} [throughout] Index "error termination" nearly everywhere it occurs. Hyperlink to "2.3.5 Execution sequence". [28:46+] 2.1 High level syntax, R212 other-specification-stmt, after "<> " insert "<> ". {It was impossible to have a CONTIGUOUS statement.} [43:13-14] 3.2.3 Constants, R308 and associated constraint C303, delete. {The term appears nowhere else in the standard.} [44:22] 3.2.5 Statement labels, p3, After "Any statement" insert "that is not part of another statement, \obs{and that is not preceded by a semicolon in fixed form},". {Contradicts 46:29 and 47:24. This change does not conflict with the existing edit in 13-008r1.} [45:90] 3.3.1p3, replace the first sentence "There ... fixed." with "\obs{There are two source forms.} Subclause 3.3.2 applies only to free source form. \obs{Subclause 3.3.3 applies only to fixed source form.}". {Remove duelling permissions and requirements - subclause headings do not qualify any text within the subclause! Also fix wonky fontness.} [90:3] 5.3.2, C517, delete "(R507)" from the constraint. {The constraint is completely self-explanatory, it does not need linking to the syntax rule.} [152:4] 7.1.12p1, item (6), Insert "and" between "NUM_IMAGES," and "THIS_IMAGE,". {Conjunction went missing when we added coarrays.} [153:23] 7.2.1.2p1, item (8), after "unless the variable" Change "is an allocatable noncoarray and" to "is allocatable, is not a coarray, and". {"noncoarray" is not a very good noun.} [186:4] 8.1.10, C844, "(R850) If a appears," ->"If a appears on an EXIT statement,". {Reword to avoid the need for specifying the syntax rule.} [187:33-188:1] BNF productions , and should be separate "paragraphs" (i.e. they are too close together). [250:31] "numericor" -> "numeric or" {Typo.} [309:16] Delete numbered blank line. {Typesetting error.} [333:16] 13.7.24 BESSEL_JN, Result Characteristics, Case (i), "is scalar" -> "has the same shape as X". {This case is elemental, thus not scalar for an array X.} [334:16] 13.7.27 BESSEL_YN, Result Characteristics, Case (i), "is scalar" -> "has the same shape as X". {This case is elemental, thus not scalar for an array X.} [344:13+x] Delete "Bits" line. {Conditional error.} [354:36] ".." -> ".". {Conditional error.} [365:4] "Bits" -> "Integer". {Conditional error.} [468:17: "USE(rational_number)" -> "USE rational_number". {Typo.} [572-583] D.2 Syntax rule cross-reference, delete. {An incorrect cross-reference is worse than none.} ===END===