[J3] [EXTERNAL] [BULK] Re: A couple more ideas for features
Jeff Hammond
jehammond at nvidia.com
Thu Apr 6 20:25:07 UTC 2023
Why not add a fully general linear solver package first and then let users define their one \ operator that invokes it?
I think the syntax of \ is the least of our problems here. We’re talking about putting a decent chunk of LAPACK into Fortran.
Sent from my iPhone
On 6. Apr 2023, at 23.21, Clune, Thomas L. (GSFC-6101) via J3 <j3 at mailman.j3-fortran.org> wrote:
External email: Use caution opening links or attachments
“\” is problematic because it conflicts with a common use of line-continuation with FPP. It might break lots of existing code. Maybe FPP would ignore it outside a macro definition, in which case it’s not so bad.
From: J3 <j3-bounces at mailman.j3-fortran.org> on behalf of j3 <j3 at mailman.j3-fortran.org>
Reply-To: j3 <j3 at mailman.j3-fortran.org>
Date: Thursday, April 6, 2023 at 3:49 PM
To: j3 <j3 at mailman.j3-fortran.org>
Cc: Van Snyder <van.snyder at sbcglobal.net>
Subject: [EXTERNAL] [BULK] Re: [J3] A couple more ideas for features
CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC.
On Thu, 2023-04-06 at 13:53 +0000, Clune, Thomas L. (GSFC-6101) via J3 wrote:
Someone in the NASA community recently reached out to me with a few questions about the standard and potential new features. After disabusing him of the ones that seemed non starters, I’m left with 2 that I feel obligate to float with the committee.
The first is a request to allow “^” as on operator. The individual actually wanted to use it for numerical exponentiation, but I considered that to be a small benefit. But it is true that Fortran has a relative paucity of symbols that can be used as operators, and most of those cannot be repurposed for use with intrinsics. I suspect with generics, there may even be increased usage of user-defined operators, but admittedly this is purely conjectural. I know we are frugal with allowing new characters, but maybe others have some solid use cases waiting to throw at this?
My colleagues asked for "\", as in MatLab, so they could define an operator useful to write matrix expressions involving multiplication on the left by a matrix inverse.
The other request was for a DO … UNTIL(cond) to complement DO WHILE(cond) … I have no doubt that this request has come up before, so someone can quickly rattle off why it was not pursued?
This was on the list from my colleagues, e.g. section 2.1.7 in 03-258r1, but not a feature request at that time. 04-112 and 04-159 were feature requests. JOR recommended no action (see 04-265r1.xls).
His other requests were all either outside the standard (command line flag to change default behavior) or were features we actually already have and the user was merely unaware.
Cheers,
- Tom
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20230406/8ead12ea/attachment.htm>
More information about the J3
mailing list