09-286 To: J3 From: Malcolm Cohen Subject: Final editorial fixes for now. Date: 2009 August 13 1. IEEE terminology We use "IEEE infinity" and "IEEE NaN" quite a few times, without any definition. [11:12+] After the definition of "host scoping unit", insert new terms "IEEE infinity IEC 60559:1989 conformant infinite floating-point value IEEE NaN IEC 60559:1989 conformant floating-point datum that does not represent a number" [150:39] "IEEE inquiry function" -> "inquiry function from the intrinsic modules IEEE_ARITHMETIC and IEEE_EXCEPTIONS". {Avoid use of nonstandard term.}. [256:1] After "IEEE rounding on conversions" insert "(14.4)" i.e. "(\ref{D14:The rounding modes})". {Cross-ref needed to understand.} [347:14+2] Change "IEEE single precision" to "IEC 60559:1989 single precision". {Correct reference to standard.} [398:4] "exception handling" -> "floating-point exceptions". {Improve accuracy of witter.} [403:20+2] In Note 14.2, "non-IEEE processor" ->"processor whose arithmetic does not conform to IEC 60559:1989". "IEEE machine" ->"processor which does support IEC 60559:1989". {Improve accuracy of note.} [404:4] "Further IEEE support is available through" ->"Additional IEC 60559:1989 facilities are available from". [404:42] "IEEE features"->"IEC 60559:1989 features". [408:4] Before "floating-point format", "IEEE"->"IEC 60559:1989". [408:6,22,26,28,29-30,31,32,38,40] "IEEE result"->"result specified by IEC 60559:1989". [408:44] "IEEE facilities"->"IEC 60559:1989 facilities". [409:15+5,15+18] "IEEE rounding"->"rounding", twice. [411:23,34] "IEEE flag"->"exception flag". [412:14,18] "IEEE rounding"->"rounding", twice. [416:22-33] "IEEE real type"->"IEC 60559:1989 floating-point format". [416:36-37] "IEEE single precision"->"IEC 60559:1989 single precision". [420:3] "the processor supports IEEE division for X" ->"division of operands with the same kind as X conforms to IEC 60559:1989". [421:12-13] Delete "the processor supports IEEE", "as described in"->"conforms to". [421:18] "the processor supports IEEE base conversion for X" ->"formatted input/output base conversions conform to \theIEEEstd{}". [422:38-423:1] "the processor supports both IEEE and non-IEEE kinds of reals" ->"some but not all kinds of reals conform to \theIEEEstd". This is not all of the problematic language, but it is a start. 2. Missing constraints Two constraints for procedure pointers seem to have been dropped. Since their absence lead to ridiculous things like putting modules into common blocks, they should be reinstated. [107:19+] Insert constraint "C577a A shall have the EXTERNAL attribute." [108:6+] Insert constraint "C577b A shall be the name of a procedure pointer.". 3. Other miscellany [4:13] After "scoping unit" insert "or construct". {Another forgotten BLOCK construct.} [295:4] Index "default initialization" here. [377:13+something] Index "default initialization" here. ===END===