Webpage

[shaunlebron]

I think a fun, human-centric webpage describing BTrDB would do well to help market ourselves to clients, investors, and talent for hiring.

There's a time and place for minutiae, but I want this page to stand firm as a thing that prioritizes concrete examples over ultra-precise statements about general applications. For example, introducing the database as:

A better time-series database built for the power grid.

I'll update this issue with plan for the outline. But my goal is insert the visualizations into a narrative. Perhaps scroll-driven like visual intros to waveforms or machine learning.

[shaunlebron]

Outline:

• Hero header
  • BTrDB (title)
  • Berkeley Tree Database (inset)
  • pronounced "Better" DB (inset)
  • A better time-series database built for the power grid. (large)
  • created by Michael Andersen
  • visualized by PingThings
  • Non-looped animation that unrolls a full tree in its extended plot state (right side)
• Context (ripping from here)
  • problem statement
  • baseline requirements
  • show image of power lines (maybe this)
• Offer (ripped from here)
  • goals
  • features
• Internals (Reading from the Tree)
  • Tree: start with the root node showing square cells
  • Plot: transition to cells with plot inside (explaining min/mean/max)
  • Branch: show one depth below
  • Mid-resolution: show how the level below can be summarized in different resolutions
  • Depth: expand to a full tree, showing raw points at nanosecond level
  • Correlate: show full zoomable plotter with relevant cells highlighted in the full tree as breadcrumbs
• Internals (Writing to the Tree)
  • Show how a node is a really bucket of <=1024 points, and after filling partitions into 64 cells
  • Show how versions are created to share nodes (we can use compositing here)