From jls@uxb.liv.ac.uk Wed May 5 10:30:51 1993 Received: from ib.rl.ac.uk by dkuug.dk with SMTP id AA13336 (5.65c8/IDA-1.4.4j for ); Wed, 5 May 1993 10:30:51 +0200 Received: from mail.liv.ac.uk by ib.rl.ac.uk (IBM VM SMTP V2R1) with TCP; Wed, 05 May 93 09:32:44 BST Received: from uxb.liverpool.ac.uk by mailhub.liverpool.ac.uk with SMTP (PP) id <08217-0@mailhub.liverpool.ac.uk>; Wed, 5 May 1993 09:30:42 +0100 From: jls Message-Id: <22677.9305050827@uxb.liv.ac.uk> Subject: X3H5, HPF and F90 To: SC22WG5@dkuug.dk (SC22/WG5 members) Date: Wed, 5 May 93 9:27:33 BST X-Mailer: ELM [version 2.3 PL11] X-Charset: ASCII X-Char-Esc: 29 Haveing read fairly carefully HPF and had a brief but reasonable look at the X3H5 draft Jerry distributed, and I would claim a pretty good understanding of F90, I am more than a little concerned. The degree of functional overlap between X3H5 and HPF but with widely different syntax is serious. Even more serious I would have thought was the number of examples in X3H5 which simply re-express in much more cumbersome form things that are relatively easily written in pure F90. My personal view is that we need both HPF and X3H5 like a whole in the head. Of the two I would plump for HPF although the X3H5 document is written in better standard speak (I have not yet seen the final revision of HPF only the last draft for comment). If either were coming up for CD ballot under ISO rules I would be voting NO. In X3H5s case I think on principle. I dont think this work should be proceeding, or if it is it should be merged with HPF. We cannot afford two conflicting standards to help optimisation of programs on parallel architectures. If an architecture independent optimisable language is not yet possible then it is too early to standardise any of them. -- Dr.J.L.Schonfelder Director, Computing Services Dept. University of Liverpool, UK Phone: +44(51)794 3716 FAX : +44(51)794 3759 email: jls@liv.ac.uk