February update

Posted: 6 months ago

Howdy from snowy Concord, Massachusetts.

We’re getting towards the end of Beta One, so I wanted to give a quick update on what I’ve been doing.

My main goals in Beta One are to:

  • Get a feel for how customers build mental models of what’s happening, and how I can adapt the interface to match them.

  • Stress-test the crawl pipelines, from discovery to crawl to taggings and scoring. I’ve increased the total number of documents by about 20x. This is great, but different things break at different scales. So, I’ve been solidifying as I go.

  • Work on tuning the relevance engine. Relevance is (for me) a function of recency, reputation (how many people are talking about a given document) and relevance (how relevant a given document is to a given tag). This is a relatively complex set of equations, so I’ve been tuning based on real-life data going through the systems.

So, what does that mean now? I have one more major infrastructure shift, which I’m hoping will be done by the end of next week. Once that’s stable, I’ll be able to declare Beta Two and move on to the next set of challenges.