[UI] Revert #994: "Update scoring graph descriptions to clarify use of WidgetMatcher"

[bergos]

This reverts all four commits from #994 (squash-merged as 560951b):

• bd88282 — Update scoring graph descriptions to clarify WidgetMatcher usage
• 0323476 — Address review: scoring graph in matcher, best parameter for select
• 31ae34d — Replace the select function with scoring graph preparation
• 8c1cba6 — Remove the best parameter from the score function

Why

The changes in #994 are substantive and not editorial: they replace the Select function with a new Scoring Graph Preparation step, remove the Best parameter from the Score function, and introduce the shui:WidgetMatcher/shui:WidgetAcceptMatcher terminology throughout the scoring algorithm section. Changes of this kind to the scoring system are the sort of design decision the Working Group would normally expect to discuss before merging. #994 was merged without such discussion, during a period when @bergos — who has a longstanding, active interest in this area of the spec — was known to be unavailable to review or weigh in. This revert restores the pre-#994 text so the scoring system design can be discussed on its merits, with the relevant participants present, before any of these changes are reintroduced.

• See this document rendered online here

[bergos]

@smessie , in the call you stated that #1163 covers the changes made in #994. That is not accurate. You introduced the section "Scoring Graph Preparation." The previous version only described the algorithm of the functions. After your changes, it now requires preprocessing the scoring graph beforehand. That requirement is not necessary, and it's unrelated to the outcome of #1163, which doesn't mention it at all.

Let's revert the changes and update the specification through the formal review process once we have an outcome on #1163.

@smessie , it would have been nice if you had mentioned this PR. After the incident with the other pull request that was merged fast-track, in the same time frame when I was absent, transparency is important here. The PR could have been referenced in issue #1163. Unlike the other PR, this one is hard to spot as containing a major change, since you didn't update the PR description after making it. The change was added late, in a commit on Friday, July 31st, and the PR was merged on August 5th. I only noticed it because of a follow-up PR.

@nicholascar @PapoutsoglouE