Anticlue

Non coding uses of subversion

In a recent project, we went through a major refactoring of the billing interfaces. This was a large endeavor, first we reviewed the current architecture of the billing interface, and one of the most alarming items was teh amount of knowledge loss that had occurred over the years. We uncovered several issues, one of which there was no clear concise source or owner of data.

The next step was to establish the authoritative source and data owners for all of the interfaced data elements. The authoritative source is the correct source of the data, like for instance admit date/time and discharge date/time authoritative source is the registration system. The data owner is the business department responsible for the accuracy and integrity of that data element. We had a couple of instances where we have to interface information to the authoritative source system, from the system that the data owner's utilize.

Next we went into a major specification review. We had a registration, billing, interface expert, and coding expert create interface specifications using standard hl7 as a template. The interesting item is with legacy systems for registration and billing, some of these systems where created before HL7 so it is more of a guideline than a rule.

The key lesson learned after 3 overwrites and the receiving of manic pages from the specifiers for lost information, it that we need to use a control system for revisions to these documents. The other item was that we had team decisions made, all parties need to be included.

Something that can improve our life in the future has been pointed out by Dru Lavigne's Setting up a Secure Subversion Server. Imagine collaborating on documents with safe and secure version control.