Comparison with official F2 reader + proposal for prefix/remainder split API

[Wildorchid]

Summary

I've been using text-vide in a product and compared it carefully against the official reader's default F2 setting on a fixed calibration paragraph (travel/itinerary-style prose).

Finding: fixationPoint 1–5 changes how much is bolded, but none of the presets closely match the official F2 output for:

• word-length fixation rules (especially 7+ character words)
• visual tracking, which depends on both bold prefixes and de-emphasized suffixes — not <b> alone

We ended up replacing text-vide with a small custom engine that:

1. Uses a length-based fixation table (closer to F2 visually)
2. Returns { prefix, remainder } per word for separate styling
3. Preserves whitespace/punctuation exactly

Happy to share comparison screenshots and our length table if useful.

Comparison text used

Reading Assist Test
I am at the old hotel near Oxford. We begin our journey through London and continue toward Birmingham. ...
Short words: a I an as at be by do go he if in is it me my no of on or so to up us we.
Long words: accommodation international transportation misunderstanding responsibility extraordinarily organisationally.

Proposal (optional enhancement)

• Add a preset mode (e.g. mode: 'f2') with a documented length table
• Optionally return structured splits instead of only HTML strings:

getFixationSplit("international") // { prefix: "interna", remainder: "tional" }