gdburns@TBAG.OSC.EDU (Greg Burns) (12/21/89)
Would it not be nice if we had more standards in our corner of the parallel computing world? 1) a standard object and executable format 2) a standard concurrency library for non-Occam languages 3) a standard, basic message passing library 4) a standard description language for multicomputer topologies I think the only thing we do have is a standard file system interface, UNIX. Am I wrong or can everybody handle read(fd, buf, length); You would think that with one chip we could get binary portability, but I would settle for source level. The Intel/Alliant PAX business got me thinking about this. ------------------------ Greg Burns gdburns@tbag.osc.edu Trillium Diving Team (614) 292-8492 Research Computing The Ohio State University Trollius - hand made - fine tailoring
stein@dhw68k.cts.com (Rick 'Transputer' Stein) (12/29/89)
In article <8912202119.AA23740@tbag.osc.edu> gdburns@TBAG.OSC.EDU (Greg Burns) writes: >Would it not be nice if we had more standards in our corner of >the parallel computing world? >1) a standard object and executable format >2) a standard concurrency library for non-Occam languages >3) a standard, basic message passing library >4) a standard description language for multicomputer topologies Well, your suggestion is a good idea. They (meaning the IEEE/ANSI folks) have a POSIX standard. Isn't that about UNIX? Why not have one for frigging message-passing extensions to it? Or an altogether new standard for the message-passing OSs? Lot's o' problems with this one. Lobby your local IEEE office and see what happens. :-) -- Richard M. Stein (aka, Rick 'Transputer' Stein) Sole proprietor of Rick's Software Toxic Waste Dump and Kitty Litter Co. "You build 'em, we bury 'em." uucp: ...{spsd, zardoz, felix}!dhw68k!stein