From bbuckley@agb.RoyalRoads.ca Thu Aug 19 04:08:03 1993 Received: from Post.RoyalRoads.ca (d3.RoyalRoads.ca) by dkuug.dk with SMTP id AA08254 (5.65c8/IDA-1.4.4j for ); Thu, 19 Aug 1993 20:08:23 +0200 Received: from agb.RoyalRoads.ca by Post.RoyalRoads.ca (5.61/5.17) id AA03822; Thu, 19 Aug 93 11:13:13 -0700 Received: from localhost by agb.RoyalRoads.ca (4.1/1.34) id AA03807; Thu, 19 Aug 93 11:08:03 PDT Message-Id: <9308191808.AA03807@agb.RoyalRoads.ca> To: SC22 WG5 Subject: vote on varying string module Date: Thu, 19 Aug 93 11:08:03 -0700 From: bbuckley@agb.RoyalRoads.ca X-Charset: ASCII X-Char-Esc: 29 Re: Vote on Varying String Module I was very interested to read John's public vote. I had read the revised copy carefully, and had a number of concerns. However, I was reluctant to vote NO at this stage, partly because of the tremendous amount of work Lawrie has put into this document. Nonetheless, in part because of John's comments, that is what I now plan to do. This is my personal vote only. As requested by John, I therefore will vote NO, but with the expectation that these largely editorial matters can be quickly resolved. As support for John, I too will vote publicly. In addition to the comments John made, I would like to add the following. I have skipped those detailed comments which are already covered by John. I would like to add my agreement for all of John's global points, except for 'Global B' and 'Global I'. I do not feel that 'B' is necessary. With regard to 'I', I think it is high time we started to get away from this attachment to upper case, and this is a good place to start. I had already suggested 'Global C' to Lawrie. Bert Buckley Detailed comments: Page line comment ----- ---- ------- iii 2.4 'Procedure' should be 'procedure' in several cases. This applies to the section titles where they appear as well. 3 23 Change 'Note that,' to 'Note that', 'forms,' to 'forms'. 3 23 Replace the final ',' on the line with ' and' and note that the commas in the list of operator forms are in the wrong font. It should be noted that these same changes should be made in several other places, as will be briefly noted. 4 13 Same as 3/23 (but 'and' should be 'or'). 4 19 Same as 3/23. 4 19 Change 'that,' to 'that'. 4 23 Change '.' to ' and is false otherwise.'. Cf 2.4.10. 4 29 Here 'and' is in the wrong font. 4 29 Same as 3/23. 5 1 I agree with John about the name of CHAR, and voted that way earlier. 5 15-19 The description fails to says what the *contents of result are in case (ii). It only tells us the length. Add at the beginning 'The result has the value of the characters of {\tt string}.' 6 24-27 The description is technically ambiguous. As written, it would be correct to return 'ello h' as the result of adjustl(' hello'). After the full stop on line 25, add 'The characters must be shifted as few positions as possible.' 6 37 See 6/24. 6 2-4 I agree with John. When the functionality is the same, the description should be copied *verbatim* from 1539. Otherwise there is the possibility of inconsistency, or an implicit conclusion by some readers that there is some subtle difference. Observe that I have already noted in 6/24 that the new description is not correct. 7 7-8 Punctation of lists is inconsistent. Here there is a comma after the first line, and a full stop after the second. These are not present in 2.4.2. 7 13-14 Delete 'A negative ... permitted.'. This is already stated on line 9. 7 26-28 Inconsistent punctation, as in 7/7. Cf 2.3.1, 2.3.2. 7 33 Change 'Standard,' to 'Standard'. 7 34 Change 'length' to 'lengths'. Cf 2.3.3. 7 36 Change 'defined ... ,' to 'in the standard set,'. Cf 2.4.4. Also, which standard is referred to. Both define character sets. 8 8-10 'and' is in the wrong font. 8 8-10 Same as 7/26. 8 11 Change 'back ... scalar' to '{\tt back} must be a scalar'. There is no list of points here, so the notation is inappropriate. 8 14-25 I agree with John about the revised wording for several of these. Since it is straightforward to define the result in terms of the Fortran 90 intrinsics for fixed length character strings, why not do so. It is briefer, and removes the possibility of inconsistency. 8 15-16 Remove the comma from the end of each line. 8 15-16 I dislike mixing Fortran syntax with English text, as I have stated before. Here line 16 is a logical expression, so it has a value true or false. In this case, to make sense of the sentence, it is necessary to add 'is true' to the beginning of line 17. Then we have a proper meaning: "the result is ... if *the expression* is true or zero if ...". 8 19 Avoid the syntax issue. This is Fortran so '=' needs to be '==', but it would be better to change '=' to ' is '. 8 22 Change ':' to ','. Remove the final comma, and add 'is true' to the start of line 23. 8 25 Change '=' to ' is '. 8 32-24 Same as 7/26. 8 35 Same as 8/11. 9 8-10 Same as 7/26. 9 11 Same as 8/11. 10 10-14 Same as 7/7. 10 31 Change 'is caused' to occurs'. Is caused by whom? 11 8-10 Same as 7/7. 11 8-10 The style of the text changes here. Change 'type' to 'of type' to be consistent with 1539 and other parts of this document. (4 times) 11 28-30 Same as 11/8, both parts. 12 15-17 Same as 11/8, both parts. 12 24-26 Change 'and the' to 'and'. Consistency of style. (3 times) 12 37-43 Same as 11/8, both parts (many changes). 13 5 Change 'in' to 'into'. 13 6 Change 'beginning with' to 'before' -- this is an insertion. 13 12 Replace '.' with 'and substring is inserted before the first character of the copy of string.' Cf 12/26. 13 16 Replace 'including ... ,' by 'inclusive,'. Cf 2.7.4. 13 39-41 Same as 11/8, both parts. 14 7-9 Same as 11/8, both parts. 14 20 Change 'a two' to 'two'. 14 21 Change 'occurence' to 'occurrence'. 14 24-28 Same as 11/8, both parts. 14 24-25 Why can string and word not be of type default character? I was opposed to adding Split, but since it is here, it should be consistent with all of the other routines and should allow arguments of all combinations of varying-string and default character. I think this is the only exception. 14 30 Change 'occurence' to 'occurrence'. 14 34 Change 'character' to character,'. 14 36 Change 'zero-length' to 'zero length'. Consistency. 15 14 Write 'straightforward'.