A history of the PL/M-86 API library written for DOS

[Mike Gillmore]

Several months ago, I started putting together notes related to the API that has evolved since 1986 when I started porting a couple of apps from the Intel Series-III and Series-IV to the IBM-PC.  They were reasonably simple, "fprint" being the most notable.

In those days, there were just a relative few API-related procedures, notably:

• dq$get$argument
• dq$open
• dq$close
• dq$read
• dq$write
• dq$exit

Over the years, and with a challenge from another engineer on our 3-man team, I created the startup which you have now;  main( numArgs, argStrs, envPtr ).  From there, I created a more commonly set of procedures for file I/O.

Just a couple of years ago, I decied to write some more ambitious procedures.  In PL/M-86, each procedure definition had a specific number of parameters.  Nevertheless, I decided to see if I could figure out how to create procedures that could be called with a variable number of parameters.

The first procedure I tackled: printf() ...

Now, before all of you ask why we didn't just buy apps for DOS, we were running DOS 3.10 ... There was no Windows, and almost no applications available commercially expect text editors (most of which were terribly slow, and limited to files shorter than 64KB), and word processors.

Also, one should note that if you could afford networking devices for inside your facility, the "internet" was not commercially available.

Thank you for your time and patience with the old man writing this.

> M

---

Mike Gillmore

the ACME Software Deli (and Malt Shoppe)