On Mon, 9 Jul 2018 16:41:40 -0300, Clark Morris wrote: >How would the WFM (Work Flow Manager IIRC) for the Burroughs B500 and >successor compare with IBM z/OS JCL and with VSE JCL. How does z/OS >JCL compare with VSE JCL? My memories of DOS360 JCL probably are >irrelevant.
I worked on a Burroughs 1726 in the mid-1970's. It was an interesting machine that included essentially the entire Master Control Program (MCP) that ran on the bigger machines. I remember that Work Flow Language was considerably more powerful than JCL, but I don't remember any specifics. I had been working with JCL for a few years at the time. I have a vague recollection that there were a few things that you couldn't do in WFL that you could do in JCL. I looked around on Bitsavers for a 1700 manual that covered WFL. I couldn't find one, but I did find a 6700/7700 System Software Handbook that has a section on WFL. http://bitsavers.trailing-edge.com/pdf/burroughs/B6500_6700/5000722_B6700_B7700_System_Software_Handbook_Jul73.pdf I also found a Unisys Work Flow Language Manual at https://public.support.unisys.com/aseries/docs/ClearPath-MCP-18.0/86001047-516 The 1700 was a very odd machine that allowed for bit-addressable memory. While the physical memory was organized as 8-bit bytes, the higher level languages all had their own "interpreter", each of which allowed for allocation of storage on any bit boundary and with arbitrary bit length. The interpreters were the microcode to support a given language. There was one for Cobol, another for Fortran, and yet another for SDL. I understood SDL to be a subset of Algol, and it was the language that was used to write the MCP. Each interpreter was optimized for the language that it was designed for. Compiled instructions were of variable size, depending on the requirements of the program. A program that referenced few data items would need fewer bits in the instruction to reference those data items than one that referenced many data items. Pages were similarly variable in size, depending upon the requirements of the program that was running. Storage could be allocated in arbitrary increments. I did some experimentation and IIRC, I found that I could do the equivalent of a GETMAIN for 1 bit and it would be honored. Regardless of the size, a 48-bit header would be defined to describe it. If you want a peek into this rather odd system, here is a link the the 1700 System Reference manual: http://bitsavers.trailing-edge.com/pdf/burroughs/B1700/1057155_B1700SysRefMan11-73.pdf And for a peek into the Master Control Program, the MCP reference manual can be found at: http://bitsavers.trailing-edge.com/pdf/burroughs/B1700/1088010B_1700MCPRefManAug75.pdf Perhaps another time I will tell the tale of what we bought the computer for and what happened with it. -- Tom Marchant ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
