The front panel contained no lights and only one
switch, which booted the machine via a standard ROM boot loader card [M847] that
was implemented with individual diodes for each bit. The ROM was
"programmed" by removing some of the diodes with wire cutters.
The systems ran a slightly modified version of FOCAL as the high-level
language (interpreter) and operating system. Out of the 4K of memory, about
1.2K was available for FOCAL program code. FOCAL is a small language
(invented by Digital) that is a cross between FORTRAN and BASIC.
These systems were installed in a short rack, with contrasting orange front
and side panels. They were very state of the art for the real estate
industry in their day, and were often prominently displayed in the real
estate office just inside the front door.
Most of the customers of this real estate system were located within a few
hundred miles of Baltimore, Maryland, where the manufacturer was located. A
few customers were scattered across the US, such as Albuquerque, New Mexico.
The customers were usually the larger real estate companies in the larger
The computer system ran a program to provide detailed estimates of
closing costs, summarize the closing costs for each of 11 different
financing plans, and calculate the maximum price home one could afford given
the person's income and expenses. This was very advanced information for
The other major function of the system was to maintain a database of houses
for sale. This database could be queried based on selection criteria such as
number of bedrooms, number of bathrooms, asking price, etc.
In the mid
1970s, the "home office" in Baltimore had an inventory of about 7,000 active
homes, which was stored on two Cartrifile tape cartridges. The length of
one record on tape was relatively small, which caused the Cartrifile to make
a "clackety" sound when a search was run. This also created a lot of wear
on the tape drives, as the tape was stopped and started for each record (no
vacumm columns here!)
Each computer program (settlement programs were written in FOCAL, the search program was
written in assembly language) was stored on one of the 4 Cartrifile tape cartridges.
The first program on the tape was loaded into memory by the boot loader and
was a "menu" program. When a program was selected from the menu, that
program was loaded into memory and overlayed the program area in memory.
The selected program would then run. When it finished, it would then
overlay the menu program from tape.
I joined the company in 1975. At this time the predecessor sytem was still
in use. It was a modifed card sorter. This system only provided database
"searches". Each house that was for sale (a "listing") was coded on a 80
column puch card. The prices, number of bedrooms, number of bathrooms, etc.
was puched as "data". The "listing card" was xerographically copied onto the
back of the punch card. This was before the advent of modern copy machines,
so this was done using cameras and xerography equipment. The search was
executed by pressing special buttons on the card sorter control panel for
price, bedrooms, etc. These buttons were the result of the extensive
customizations of the card sorter. When the sorter was started, only the
matching cards ended up in the output hopper. To get the information about
the matching houses, one merely had to turn the card over and look at the
listing information on the back.
Before my employment, they had started experimenting with using PDP-8
computers. The first system used a DECtape tape unit. This never made it
into production. The Cartrifile system was complete and in production when
I arrived. Some of the units also used the PDP-8/L computer, but it was not
as user-friendly due to the lack of a ROM boot loader. Re-booting required
toggling in a short program from the front pannel. It was difficult to
teach secretaries how to do this. Several times I would have to drive in to
work in the evening after a power failure to reboot the PDP-8/L
During my 5-year tenure, the 8/M was replaced with a PDP-11/05 running RT-11 and
Multi-User BASIC with RK05 disks. This never made it into production, but a
result was that all of the programs were re-written in BASIC.
The next upgrade was a PDP-11/34 running RSTS/E and a RK05F (dual RK05).
This version was very succesful and allowed penetration into new markets in
Pennsylvania, California and Texas. I left the company in 1980 and joined
Digital Equipment Corporation.
Many thanks to the donor for this information!