J3/01-276 Date: August 3, 2001 To: J3 From: Dick Hendrickson Subject: General questions about 007 Here are four general topics that didn't seem to fit into the chapter by chapter approach for detailed comments. I'm also doing separate papers for each of the first seven chapters with detailed comments. References: Should we get rid of all of the references? They are spotty at best. If not, let's assign somebody not named Richard the task of going through and making sure that we consistently reference things that need to be referenced. I think we should delete them all and turn them into index entries instead. See e.g. 2.2.3.4 references to 12.3 and 12.4; nobody will follow them forward. See also last 3 sentences on page 29. The terms are all in the glossary/index Other changes such as Page 12, 2..2.3.2 remove reference to R213. adds nothing to the text Page 13, 2.2.3.3 remove reference to R211 adds nothing to the text make me think we would be better off to start over. Enhancements Could we add an appendix E or so that has the BNF rules and constraints? Ala F77 railroad tracks? Page 64 C522. Why don't we allow initialization of allocatable objects? We allow an allocatable object to be of derived type and allow every component in the type to be initialized. Why not allow the easy case? R525, etc. Isn't this a case where -list could be used on the right of the BNF? I've noticed a few cases where we have BNF like thing1 IS thing2 [,thing2]... but thing2-list is equivalent.