ISO/IEC JTC1/SC22/WG5 N1825 Result of letter ballot on N1814 Key for the Result line: Y Yes. C Yes, but I recommend the following changes. N No, for the following reasons. A Abstain. Y Bader Y Chivers C Cohen N Corbett Y Donev Y Gorelik C Long Y Maclaren Y Moene Y Morgan C Muxworthy Y Nagle Y Martin Y Reid Y Sleightholme C Snyder A Takata Y Whitlock Y Xia Totals: Y 13, C 4, N 1, A 1 Recommended changes Cohen 1 Introduction, page xiv, "Programs and procedures:" bullet point, after "A null pointer" insert "or unallocated allocatable". Reason: Missing feature from the feature list. Note: I strongly recommend this change. Cohen 2 10.10.4 "List-directed output", paragraph 13 (page 266), after "Except for" insert "new records created by explicit formatting within a defined output procedure or by" Reason: This paragraph should be identical to the one for namelist formatting at 10.11.4.3p5; the latter was changed by interp F03/068 which had too narrow a scope and the inconsistency was not spotted at that time. Cohen 3 Introduction, page xii, bullet-point "Execution control:", append new sentence to paragraph: "The ERROR STOP statement initiates error termination." Reason: Even though the difference between STOP and ERROR STOP might only be interesting in the parallel environment, it is not obviously a synchronization construct so not covered by "Parallel execution:". Long 1 In Note 8.45 (page 194 in 8.5.6 LOCK and UNLOCK statements) replace the first comment in the example code: ! Lock to manage the work queue with ! Lock on each image to manage its work queue [This suggested change arose when someone familiar with distributed locks (so-called MCS Locks) thought the coarray variable queue_lock represented a single, distributed lock, rather than a separate lock on each image. The proposed replacement wording makes it more clear that LOCK_TYPE does not have some magical global variable property.] Muxworthy 1 There are bad page breaks at the bottom of pages 52, 380 and 565. Muxworthy 2 Similarly on pages 151, 486, 492, 522 but of much lower, possibly negligible, priority. Muxworthy 3 Table 2.1, split between pages 31 and 32, is not well presented; it was not so split in F95 or F03. This could be resolved by additional white space on pages 30 and 31 or by interchanging subclauses 2.3.2 and 2.3.3. Muxworthy 4 The diagram of the plasma module would be improved if it were not split over two pages (504 & 505). Also it should now use continuous lines rather than ASCII characters. Snyder 1 In 13.7.136p3, p380, insert "a" before "default" in the description of the SIZE argument. Snyder 2 In 13.7.136p3, p380, insert a full stop after "N" in the description of the GET argument. Reasons for NO votes and comments with abstentions Corbett I vote no because of the problems with the proposed standard that were found during March 2010, namely, the conflicts between the expanded definition of the term "not present" in Section 12.5.2.12 and the specifications of some of the intrinsic functions, and the problem with generic resolution using the POINTER attribute to disambiguate references. I am concerned that approving the standard and then resolving the problems through the interpretation process will favor solutions that involve small changes to the standard over the best overall solution. I shall change my vote from no to yes, perhaps with comments, if a note is added explaining the problems and making it clear that a possibly significant change to the approved standard is expected. Takata Although I share Bob's concern, such a note suggested by him would be inappropriate, even if it is not illegal.