Score aggregation on SEO and readability analysis could be simplified and/or improved.

[hansjovis]

How the overall readability score is computed can be improved. It now works like this:

1. Translate each assessment score to a category:
   either "error", "feedback", "good", "ok" or "bad".
2. Give each assessment a penalty score, based on the category.
   • "good", "error" and "feedback" get no penalty.
   • "ok" scores get a penalty of 2.
   • "bad" scores get a penalty of 3 when the user's language is fully supported, or 4 if not.
3. Base the overall analysis score based on the sum of the individual assessment penalty scores.
   • The available scores are either "needs improvement" (30), "ok" (60) or "good" (90).
   • The thresholds change based on whether the language is supported.
     • If it is fully supported, the thresholds are more lenient.
     • If not, the thresholds are more stringent, giving lower scores at lower penalty thresholds.

It can be worthwhile to simplify this a bit, or provide some explanation on why this is computed this way.

[igorschoester]

I would like to broaden the scope of this issue to include the SEO scores too.

For a previous scaling idea, see the linked #401 issue.