Turning specifications into estimating imports

Estimators re-keyed architects’ specifications into a legacy system whose import format had no available documentation. Its fixed-column rows, placeholder rules, and silent rollback behavior made valid imports difficult to prepare and check.

Each opening carries door, frame, hardware-set, and finish details, but the importer expects a fixed-width text row with 107 fields. Some invalid files appear to import successfully before the system silently rolls back the data.

The format was decoded by changing one field at a time, exporting, and comparing files. A converter then turns specifications into structured imports, with field and placeholder rules handled in code and output checked through import and re-export.

1. 1

   Decode the file

   Create a small record, change one value, export again, and compare files to identify field meanings.

2. 2

   Read varied specifications

   Extract schedules, hardware sets, and project details while accounting for differing document layouts.

3. 3

   Build and verify rows

   Apply required-field and placeholder rules, then import and re-export a test job to check the result.

Specification details can be prepared as system imports, and the format, placeholder rules, and rollback traps are documented in code. The verification method checks whether generated data survives the system’s import and re-export round trip.

The extraction supports several schedule layouts, including direct listings, merged banner rows, and room-type templates expanded against a door census. One full building has been processed from architect specification to a job opened in the estimating system.