|The City of San Antonio had hired a collection of engineering firms to analyze
and ultimately upgrade the city's sewer system. This was driven by growth the city
experienced in the 1960's and 1970's as well as new environmental regulations
coming into effect that required the construction of more advanced treatment plants.
The City had a Datapoint 5500 with a 10 Mb fixed/10Mb removable disk drive available
for the project. This computer was not particularly appropriate for engineering
analysis, but it was appropriate for data entry, given that large amounts of
detailed information had to be entered before analysis could proceed.
For capacity estimation purposes, there were 60,000 trunks and 120,000 junctions,
each of which required about 15 columns of data. Some of this was parameterization
of the trunk (inside diameter of the pipes) and some a reflection of the amount
of material introduced during 'dry' and 'wet' weather. Sewer lines typically carry
quite a bit of storm water during rainstorms. This can be caused by outdoor drains,
drainpipe vents, and inflows through manhole covers, among other sources.
Our job was to design the forms for collecting this information and develop an application
for analysis. We used the Datapoint Databus language entering and maintaining the node
tables. We used the Datapoint version of Basic to run the analysis. To keep this in
context, the computer had 48K, ran at about 4Mhz (somewhat equivalent to an IBM PC/AT),
and could run 5 user (ASCII) terminals in a time sharing mode.
One of the discoveries we made, having built a body of test data, was that the ISAM
indexing scheme we were using worked one way with the Databus language and slightly
differently with Basic. This proved to be a major complication, although rebuilding
the index using utilities when we switched between languages helped.
The objective of this system was to identify the trunks or junctions that would have
to be expanded or rebuilt in order to handle the worst case volume. This turned out
to be major portion of the system. We were given about six weeks to develop the system,
which turned out not to be enough, however the surveyors were also behind, and they got
their data into the system about the time we got it to work.