THE STORY BEHIND APEX AND XPL0


It all started with a group of hackers in 1973. This group, focused at
the Colorado School of Mines and organized by T. Wayne Wall, developed
an interest in the then new 6502 microprocessor, primarily in the form
of the Digital Group kit -- although homebrews, Kims, Tims and even
Jolts existed in the group.

Everyone agreed that software was needed badly, both an assembler and a
higher language. A cross assembler that ran on the Mine's PDP-10 arose
from the cosmic mud puddle, so the focus was on a higher language. It
was during this early experimental phase that Peter Boyle wrote the
original VED editor for use in an editing terminal. It was also around
this time that Jim Lauletta started wheeling and dealing around the
group as Apparat, and contact was made with a similar group in Reno
which later became the 6502 Program Exchange.

A split developed immediately. One group, spearheaded by Wayne talked
about writing an interpreter in the BASIC family. The other, spearheaded
by Peter wanted a block-structured, compiled language.

This resulted in two products: Wayne's FOCAL and Peter's XPL0. Group
enthusiasm and support helped both of these languages rapidly develop to
a moderate degree of maturity. They polarized in application. XPL0
focused on being a C-like, lower-level language with full user control
and very little assistance for the novice. FOCAL became a fully devel-
oped interpreter for numerical work.

At this point Wayne was free to spend large amounts of time on the 6502,
and he soon produced a bewildering assortment of code, including a
native assembler and a speed enhanced version of the I2L P-code inter-
preter behind XPL0. All of this code was marketable, so the Reno group
set about doing so. A small but steady income began to flow to the group
and the authors involved. Larry Fish wrote an XPL manual at this point.
XPL was converted for use on several other machines including the
PDP-10, IBM-360, 8080 and 6800 at around this time. These versions have
since become obsolete.

Then came the floppy disk, and Peter and Larry kludged them into their
systems. It became clear at that point that an operating system was
needed so Peter and Larry began a dialogue on the subject. The rest of
the group were off working on audio and digital cassette drives so Larry
and Peter were left alone at first. Peter, egged on by Larry, wrote a
rudimentary system called FFS in XPL. XPL and the VED editor were
advanced to keep pace.

Then the Apple came upon the scene. XPL, VED and the assembler were
quickly adapted for it. The resulting sales proved that the Apple was
making a market. What the Apple lacked was a general purpose operating
system so Peter rewrote and improved FFS to become Apex V1.0. The idea
was that Apparrat and the Reno group would market it. They did so with
minor success.

Apple's BASIC swamped FOCAL, but there was still no useful compiled
language for the 6502 so XPL was expanded by Loren Blaney to include
floating point. Many others from the group and elsewhere, notably Wayne,
Larry, Loren, Ted Dunning, and Roger Nace as well as the Marshs in Reno,
and Penny Globus and Kevin Norman in Berkeley made contributions to
documentation, utilities, and support code for both Apex and XPL. It was
at this point that Penny and Kevin repackaged all of this code under the
name Computer Sight.

In 1984 Loren, realizing that XPL could be made to run much faster on the
68000, started beating on Rich Ottosen. By August Rich had a board
running in an Apple slot that used the 68008 chip. Loren modified the XPL
compiler to produce native 68000 code, and suddenly XPL programs were
running 20 times faster. The lust for speed continued. Eric Smith
modified Rich's board to run a 68020/68881 combination. The compiler was
modified, and the floating point calculations used by a Mandelbrot fractal
generator were running hundreds of times faster.

A year later DFM Engineering hired Loren to replace the operating system
used in their 68000-based Stride 440. Version 1.8 Apex was created. Its
fast code and RAM disks now made using the Apple seem painfully slow.

By the spring of '87 Loren had XPL running under Apex on the Amiga, and
Peter had XPL running under Atari DOS on the 1040ST.

In 1987 Apple finally came out with an acceptable Macintosh. It had the
68020 with a 68881, and it had slots. What it needed was to have the
rodent in its operating system exterminated. DFM hired Loren once again
to provide Apex so that it could run their telescope control system using
this new platform. At first the task seemed simple: write a few device
drivers for the Mac II. But Apple (having learned from IBM's mistake)
wasn't about to have the market flooded with affordable clones, and made
the details of its hardware very strange and very secret. Using an
extreme amount of perseverance, experimentation and disassembly, Loren
was finally able to divine enough secrets to come up with the necessary
handlers.

By this time everyone seemed to own a PC clone, so Larry translated the
XPL0 interpreter that was running on the 6502 to run on the 8088. He ran
it under MS-DOS, which made it compatible with the entire PC family. It
was about two to three times faster than the Apple II when using the
standard 4.77 MHz XT, and if you needed more speed you could run it on a
386-based machine.

Apex and XPL0 continue to survive in spite of massive support behind
other operating systems and languages. The reason is that they provide
a relatively simple yet powerful programming environment that can be
readily modified to suit the problems at hand.
provide
a relative