More Pick info (was Prime 2250)

Prime Infomation was certainly a pick variant - I spent a lot of time (about 7 years) on Prime's and Prime Information at AB. When pick first came out, is was what we called "vanilla" which meant that the machine booted pick - there was no other operating system. Later other manufacturers (licensees of pick systems) decided to strip out the monitor and missionary/native code architecture, and gut everything that was OSlike rather than RDBMSlike. They then put the remnants on top of some other operating system (layered rather than vanilla). Unix was common, as was DOS, Windows, PRIMOS, etc. To answer someone elses question, BASIC was the programming language (you had no choice). There was also PROC which was not that much more than batch files in DOS. Then there was the assembler. The assembler verb definition was removed from many later versions but you could define a pointer into the ABS area to call it up. BASIC was compiled into DCD or Pcode, and the result was run interpretively. The DCD/Pcode was rather elegant - it implemented a stack architecture to speed code generation and execution in RPN. At the end of each BASIC statements resulting code the state machine was completely back to its initial value. In a certain sense, each basic statement resulted in a separate DCD/Pcode subroutine that was reentrant as the symbol table and variable allocations were made out of the users workspace. All the Pick variations handle file I/O the same - I've never seen any programming language even close to pick basic in the elegance of the file/record interface. The assembler was virtual, in that the instruction set really didn't exist on any machine. So - assembly was a 3 pass process. First, the Pick assembler turned your assembler file into "virtual machine code". Then a BASIC program turned the virtual machine code into an assembler program in the native assembly language of the cpu it was running on. Finally the native assembler was run to generate executable object code. Somewhere I still have some listings where you can see the same program turn into 68000 assembler and then compile on a different machine and see each instruction turn into 8080 assembler. Nifty AND educational :) I forget if prime information was created by ex Vmark engineers, or if Universe was created by ex Prime engineers - one way or the other ;) I have to admit, next to freeBSD, Primos is a really nice OS IMHO. ISTR that the baud rates for the AMLC lines (serial cards) were controlled via the AMLC.COMI file, located on the system volume ( <SYSTEM>AMLC.COMI or <MFD 0>AMLC.COMI). This was a command input file which set the line characteristics. Can't recall if the console speed was set there, but the other serial lines were. I have a complete Primos and Prime Information manual set at the office. I'll check how to set the console speed for you. If you have no docs and are playing with the primos machine, just remember, it's "A " rather than "cd ", and "ED" rather than "vi" <grin> oh - emacs was VERY common on primos, so that might be there as well for you to use. One final note - I do NOT want to represent that I wrote the Pick operating system, or that I was on the original development team. Dick Pick and Chandru Murthi would just scream if I said I was. Pick was developed in the late 60's by those two under a government project for maintaining the data on a helicopter project. The original name (according to the US Government) for the Pick OS was G.I.R.L.S. (Generalized Information Retrieval Language). No Kidding. So - I didn't invent it, I was merely a freelance consultant who worked on modifications and enhancements to the virtual code and monitor code in the Pick OS for some of the pick licensees. Cheers! Jay West -----Original Message----- From: John Lawson &lt;jpl15(a)netcom.com&gt; To: Discussion re-collecting of classic computers &lt;classiccmp(a)u.washington.edu&gt; Date: Sunday, April 11, 1999 7:43 PM Subject: PR1ME 2550 Up and Running