Archaeology - Part 2: Metadata and Legacy Understanding

Archaeology—Part 2: Metadata and Legacy Understanding

The existing data is the "horse's mouth" in terms of how the system has been used up to this point. It highlights what people have been doing in fields that have little semantic validation. It doesn't shed any light on what they might have done, or what the system allows them to do. To obtain this information, we look at the systems that are enforcing the current semantics: the metadata and the existing applications.

Metadata Review

Somewhere, there is data about the data model. Maybe it's in a repository, or maybe it's in comments in the data definition level (DDL). Maybe it's in copybooks for the code. Maybe there is some separate documentation. Regardless of where you find it, you will want to review it from two standpoints:

• Does it indicate any understanding you haven't come across already?

• Does it agree with the data? If not, consider the metadata to be the intention. You may wonder why the intention is not currently executing, but that's all you can know at this point.

Legacy Understanding

Finally, much of the semantic validation is done in the existing application code. There are three main ways to gain understanding of this:

• Interview the maintainers of the existing system (this is perhaps the most efficient).

• Read the code (this doesn't scale very well).

• Employ legacy understanding software.

Legacy understanding code reads the source code to the application systems and "slices and dices" it into data that can be interrogated. Legacy understanding software can cross-reference where certain fields are being accessed, and it can also help you find all the hidden types we discussed earlier. This sofware can uncover every place there is an "if" statement and report on what it appears to be distinguishing. Some of the key vendors of legacy understanding software are Relativity, Merant, and SEEC.