From fsmart@singnet.com.sg Mon Feb 17 13:02:06 1997 Received: from melati.singnet.com.sg (melati.singnet.com.sg [165.21.1.15]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id NAA22270 for ; Mon, 17 Feb 1997 13:01:23 +0100 Received: from Singnet.singnet.com.sg (ts900-5818.singnet.com.sg [165.21.162.38]) by melati.singnet.com.sg (8.8.5/8.8.5) with ESMTP id UAA09029; Mon, 17 Feb 1997 20:00:22 +0800 (SST) Message-Id: <199702171200.UAA09029@melati.singnet.com.sg> From: "Fred Marton" To: "Jerrold L. Wagener" Cc: "'vsnyder@math.jpl.nasa.gov'" , Subject: Re: F2000 Wish List : Default types used in intrinsics, plus ... Date: Mon, 17 Feb 1997 19:58:44 +0800 X-MSMail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1155 MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Jerry, thank you for the reply. > > 7. Allow pointers to point to PARAMETERS > > Do you have some reason in mind? Off hand I would think that this might > prevent implementing parameters as if they were simple macros, which is > currently an allowed implementation. -jlw > > In reference to item 7, allowing pointers to point to parameters, I run into the following situation quite often: In a simulation, where the range of input parameters is quite wide, there are many ( actually most of them ) calculations that require parameters dictated by the user inputs. Some of these parameters are fixed, where I use a parameter to assure that when it is used it has the value built-in the program. However, with some user inputs the same parameter must be given a variable value ( that is a non-built-in value ). Pointers would be useful in this situation. In F90, however, either there is logic ( read macaroni code if not spaghetti) to use or not use pointers with PARAMETERS, or not use pointers at all ( my choice up to now ). The proposal allows the flexibility of using pointers to variables or parameters as needed. However, as hinted by your note, I cannot comment if this proposal would negate some already implemented property of pointers. It does not seem to do so, but I am no expert. As a separate point, the comment about the size of the constituency brings forth the following thoughts. Say a canvas is done on the usage of each Fortran statement. I am very confident that if such a canvas were done, that there would be a bell curve of Fortran statements vs. usage. After taking the canvas, should we drop all statements used by less than 1% (or 0.1%) of the users? I do not think so. The strength of the language comes from its richness, not its usage. Well, that is my humble opinion. Thanks again ( and I hope some of the proposals get to be discussed ). Fred Marton ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Logic is a method whereby we can be wrong in confidence ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ---------- > From: Jerrold L. Wagener > To: Fred Marton > Cc: 'vsnyder@math.jpl.nasa.gov'; sc22wg5@dkuug.dk > Subject: Re: F2000 Wish List : Default types used in intrinsics, plus ... > Date: Monday, February 17, 1997 5:54 AM > > Fred, > thanks for your note, which just showed up and was waiting for me when I > returned from the WG5/X3J3 meeting last week. I wish we had had it last week, > as it touches on some of the decisions made. See comments interspersed below, > and I'll copy this reply to the committees. > > Jerry Wagener >