jdb%s1-c@sri-unix.UUCP (06/21/83)
I have been out of touch for about six months, so perhaps my perception of the current state of affairs with regard to APL is no longer accurate. Nonetheless... The APL which Berkeley distributed with 4BSD and 4.1BSD is derived from a program originally written by Ken Thompson at Bell, many years ago (to the best of my knowledge, sometime before V6). This program was part of an early distribution tape. Another variant of the same program followed another path and ended up at Purdue University. There it was modified, first by Jim Besemer and later by me, under the guidance of Dr. Anthony P. Reeves. The Purdue version is not wholly a superset of the version distributed in 4BSD (in particular, it does not use the two-character mnemonics to represent APL characters), but it is generally more robust. It also will run on both (I&D) PDP-11's and VAXes (workspaces are not directly interchangeable but can be converted by a utility program). [I once ran it on 32/V, but principally developed it for 4.[01]BSD.] Until last fall we were distributing APL to anyone with a UNIX source license >= V6 (it no longer runs on vanilla V6 but it contains no V7 code of Bell origin). Since then, both Dr. Reeves and I have left Purdue. Dr. Reeves expressed an interest at that time in continuing the development of APL at his new location, Cornell University. If you are interested in a copy you should contact him; he can be reached at ...!decvax!cornell!reeves reeves@cornell (607) 256-4296 In answer to the original question, the Purdue version does provide some low-level I/O functions for UNIX I/O to/from files. By combining these with the "execute" and "format" functions you can perform ASCII I/O. (Alas, these facilities are not as complete as I'd like; however, at the time I was working on APL there were more important things to do [such as my thesis]). John Bruner Lawrence Livermore Laboratory (415) 422-0758 ...!decvax!decwrl!mordor!jdb jdb@s1-c