J3/02-301 Date: 13 November 2002 To: J3 From: Aleksandar Donev Subject: Miscellaneous F2x items Reference: J3-007R3 ______________________________________ Summary ______________________________________ Some items that have come up on J3's mailing lists and elsewhere are registered as requests for further clarification/discussion. I believe these need to be addressed before the final standard is released. Some are simple typos, other are requests for extra functionality or changes. ___________________ 56, Note 4.50 error ___________________ The IF/END IF conditional needs to be replaced with SELECT TYPE in order for this example to be correct: SELECT TYPE(B) CLASS IS(POINT_3D) ...! Body of IF END SELECT ___________________ NONKIND is a bad choice ___________________ Van Snyder and others have raised the very salient point that NONKIND is a very bad choice of word for the type parameters we have in F2x. A much more appropriate phrase is EXTENT, since later we may add other kinds of type parameters, such as other types, characterized by other keywords. I am raising this here just in case Van Snyder does not follow up on it, and will write a separate paper if needed. ___________________ Request for ``sizeof'' intrinsic ___________________ A discussion about the possibility of adding an intrinsic to return the ``storage'' extent of an object, ike the C sizeof function, came about on J3's list. I believe the committee should form its opinion on this matter and possibly recommend that such an intrinsic be added to F2x. The details of this intrinsic are to be discussed. ___________________ Unsigned integers in F2x ___________________ F2x does not address the numerous and lingering complaints from users about the lack of real standardization and proper treatment of unsigned integers in F2x. I register this complaint here and request that J3 explain its standing on this issue. My personal feeling is that the standard needs to do more to standardize the way unsigned integers are treated by Fortran compilers. ! EOF