Lmst

Back on the topic of auditing code examples in our documentation, part 5 of my 8-part series dives deep into the data model we used to store metadata and the code examples.

I look at the types of queries this model enables, performance implications, and future iteration. Data modeling is one of my favorite exercises, and I'm happy with the outcome for this project!

#TechnicalWriting #DeveloperDocs

https://dacharycarey.com/2025/04/27/audit-model-code-example-metadata/

When we started talking about performing a code example audit, I wondered if we could use an LLM to help us categorize our code examples. I built a quick PoC on a Friday, and we ended up iterating on it to perform the audit.

Spoiler alert: LLM categorization accuracy plunged from 90% to 36.7% using real-world data.

This is part four in an eight-part series about auditing the code examples in our documentation.

https://dacharycarey.com/2025/03/23/audit-ai-assisted-classification/

#TechnicalWriting
#Documentation
#DeveloperDocs

When we built tooling to audit the code examples in our documentation, we had to decide how to access the data. We had a rough idea of what content we wanted to audit, but implementation required more than a rough idea. We also had to decide how to work with this content and what to do with the resulting data.

This is part three in an eight-part series about auditing the code examples in our documentation.

https://dacharycarey.com/2025/03/16/audit-access-data/

#TechnicalWriting
#Documentation
#DeveloperDocs

The next piece of auditing code examples in our documentation was deciding what information we actually want to track. Code example category was definitely one, but there were many ways to slice the data. What should we track?

This is part two in an eight-part series about auditing the code examples in our documentation.

https://dacharycarey.com/2025/03/10/audit-what-to-track/

#TechnicalWriting
#Documentation
#DeveloperDocs