I just heard of a good use of the XML Structured Document Complexity Metric to help negotiating a contract variation, for an XSLT stylesheet project.

As I understand it, the project was quite a large one involving up to forty people, for publishing information into a dozen or so documents with strong deadline. At one stage the deadlines were slipping, and it was found that most of the extra work was coming from a couple of the publications which had a stream of formatting problems. The client thought format and document processing was not exceptional while the developers said the format and processing were more tricky than contracted for, and wanted a contract variation.

The client had initially estimated that all the publications were of similar scope; The project manager of the developers said “Get me some numbers” and the programmers ran the Structured Document Complexity Metric against the stylesheets which showed that the troublesome documents were, together, about twice as complex as the rest put together.

How significant the actual numbers are is a point for debate; however it is the ballpark and trend that is important. I am told that with the metrics as objective evidence, it really helped contract variation discussions and development efforts keep focussed. Deadlines are a time when it is easy for relations to sour with suspicion and recrimination; however metrics can provide a great escape towards professionalism and civility and risk-management: with metrics it becomes easier to figure out “Oops, my bad!” or “Doh, your bad!” and move on fast. The lack of objective evidence can breed irrational suspicions, impossible timetables, wrong cause-and-effect-inferences, victim mentalities or stalled negotiations due to entrenched subjective positions.