Blocks not laid out optimally

This is a bit of a subjective decision, but I'll try to back things up with as much data as I can:

Currently, the Scratch 3.0 block palettes are laid out in the same way as they are in 2.0 (with the exception of groupings: #752). However, this layout isn't the best. After taking some time to look through the block usage statistics, I found that the blocks placed at the top of the palette aren't always the most used blocks. This is an issue because the further down a block is in the palette, the more likely it is that a user will need to scroll in order to reach it (especially with the larger 3.0-style blocks, which will require more scrolling in general).

Take, for instance, the say and think blocks. They are all placed at the very top of the "looks" palette, and take up a fair amount of room. Yet the block usage statistics (found at the bottom of the stats page) show that the think and say blocks make up a very small portion of all looks blocks used.

In order to aid with the creation of an alternate palette, I've created an ordered list of Scratch blocks, sorted by usage: https://gist.github.com/PullJosh/aa842db1cdfe1fe553b2a1ebe2396646

Based on these statistics, I've created an alternate layout for the Looks and Motion palettes: https://gist.github.com/PullJosh/85275f2b8f05ca255fa1891ec4cabdc1 (If this is something that people are interested in, I'd be happy to arrange the rest of the palettes as well.)

My goal with the new arrangement was to keep blocks grouped logically, while also ordering them based on usage. Every decision was made intentionally (I feel that I could justify every decision made), but things could always be changed if someone disagrees. The hope is that, with a new layout, Scratch could retain its user-friendliness while also saving time for users.

Review the block-usage statistics and the two linked gists describing the proposed ordering for the Looks and Motion palettes. Compare those proposals with the current Scratch 3.0 palette layout; done requires an agreed-upon revised arrangement, but the issue names no implementation files or tests.