It was suggested that this change is good, but perhaps a little bit subtle. Should we mention that Scrabble is a game, or is the description suitably vague to encourage just a little bit of research?

Then, you may also like to change some other instances of the uncapitalised word "scrabble".

exercises/etl/canonical-data.json:        "Transforms a set of legacy scrabble data stored as letters per score",
exercises/etl/description.md:We're going to extract some scrabble scores from a legacy system.
exercises/etl/description.md:The shiny new scrabble system instead stores the score per letter, which
exercises/scrabble-score/description.md:Given a word, compute the scrabble score for that word.
exercises/scrabble-score/metadata.yml:blurb: "Given a word, compute the scrabble score for that word."

of course, choosing not to do so wouldn't be grounds for refusing this PR, it would just be an indication that perhaps there is more work to be done.

Good idea, keeps it in line with the PR, and appropriate.

We should not change the exercise name in the canonical-data.json “exercise” field. Several other exercises have proper name’s in that field (armstrong-numbers and pascal-triangle for example) and there’s no sense settings the precedent for a corporate trademark over a human person’s name. Plus this is used to map to a directory on the file system in at least one script I can think of, so this would end up forcing brittle case comparison and could cause real problems with case-sensitive file systems.

I was thinking it might be "sensitive". Thanks for the confirmation.

I think the changes elsewhere are good. It's a proper noun if nothing else, hence deserving of caps.

Versioning on json is blocking.