ted@grebyn.com (Ted Holden) (12/01/89)
From: Steve Vestal, Honeywell >As I recall, the Performance Improvement Working Group (PIWG) affiliated with >ACM's SIGAda intended to collect results for their benchmark suite and make >these publicly accessable. Does anyone know whether these results are >conveniently available (e.g., anonymous ftp)? Inconveniently available? >An article in the Nov/Dec Ada Letters by Asplund et. al. gives some benchmark >figures associated with interrupt handling (among other things). Does anyone >know if PIWG is thinking of addressing interrupt-related timings in any way? Just when I get ready to leave off the Ada group for awhile, they toss me something which nobody could resist. Normally I wouldn't cross-post this kind of thing, but I feel the people in the C++ group need to know the scale of the opportunity which is here. The idea of being able to standardize on a single programming language (assuming that language were good enough to standardize on) would be of enormous benefit even to many mundane organizations such as Census, where data has always been generated with mainframes and Cobol and analyzed with Fortran and Fortran related software, and the two groups can barely talk to eachother. The U.S. military has a far worse Tower-of-Babel situation and has attempted to solve it by creating a Frankenstein language for all purposes. The literature and every conversation I have ever had with real-world people who have been forced to actually attempt to USE Ada have totally convinced me of the failure of this project. By the same token, C++ could very nearly, if not entirely, fill the bill. PIWG??? I feel the word "working" should be left out of the group's title; it doesn't really fit, and it just makes the acronym sound bucolic. Kind of like Pappy Parker back in the Ozarks callin the "hawgs". Did you ever wonder why there isn't any such thing as a PIWG associated with Pascal, C, or C++? Why these other languages don't seem to REQUIRE a PIWG? I'll tell you: it's because these other languages were D-E-S-I-G-N-E-D right in the first place. For this reason, practitioners of C++ programming don't need to spend three years of "designing" around a language for every three months of coding to solve problems. They simply solve the customer's problems. Ted Holden HTE
rracine@ajpo.sei.cmu.edu (Roger Racine) (12/02/89)
The PIWG (Performance Issues Working Group) has a very useful set of tests for benchmarking Ada compilers. Ted Holden, you are wasting an awful lot of time and money sending these misstatements around the country. You do not have to like Ada. You may have to use it if you do work on new DoD projects. But what purpose does it serve to answer a request for information on a serious subject with a lot of rhetoric without any facts to back up your statements? Back to the PIWG. To attempt to compare compilers, operating systems, databases, and other software systems, people have been creating benchmarks for years (Whetstones, Dhrystones, etc.), no matter what Ted Holden says. Byte Magazine has a set of tests for comparing computer systems. PIWG created a set for testing Ada compilers and runtime environments. It is not easy to create standard tests for interrupts u knows the hardware. So the last time I used the PIWG, it did not support testing that. However, they have been working on it, and it could be available. The correct person to contact is Dan Roy (301) 464-6800. He chairs the working group. Roger Racine