J3/01-277 Date: August 4, 2001 To: J3 From: Dick Hendrickson Subject: Chapter 1 and Intro comments 1) Intro XIV. Need to add regularization of I/O keywords on OPEN. Format and I/O statements to (4) 2) Page 2, section 1.4(6). Don't we specify IEEE rounding properties on those processors which support IEEE. Maybe we need to add an "except for IEEE Computational Rounding And Processing" 3) Page 2, section 1.4(7). Does the stream I/O stuff implicitly specify the properties to be "C like"? 4) Page 2, section 1.5(2). It looks to me like all of the obsolescent features are syntactical features. Should we make this sentence stronger and simply say the processor must report all of them, deleting from "insofar..." … 5) Page 2, 1.5.1, last paragraph. Doesn't F95 say PAD= is 'UNDEFINED'? Isn't that something we perverted at meeting 157 via the misuse of the interp process? 6) Page 3, 1.5.2. We need to say that MOD(I,J) has been changed in F2K (and also in F95). I'm not sure if it is a compatible extension or not. Changing from "processor dependent" to "prohibited" seems to me that it could break a program that relied on a particular processor dependent implementation. The last paragraph of page 2 seems to me to say this is OK (otherwise how would the clock and random number routines be used?). 7) Page 3, 1.5.2. Do we need to say something about IEEE arithmetic compatibility? I seem to recall some change to X**Y, or maybe X/Y, from prohibited to processor dependent or something like that. In any case, changing to a system that allows NaN instead of overflow/abort should be mentioned. This touchy-feely section is to give a flavor of what to expect. Perhaps a non-normative note. 9) Page 4, 1.5.3, bullet (3). Can also specify PAD= on a READ I believe. 9) Page 6, 1.6.5 (5). Could we delete this one. Saying we do something consistently adds nothing. Otherwise maybe add something like "as opposed to substring things" 10) Page 7, 1.7 1st line of last paragraph. Change five to six to account for PAD= change.