2026-09-22 — HTML Research Summary

Frontend Analysis (FA) treats HTML as an independent semantic research domain rather than as a simple source of tags or a conventional DOM tree.

The current HTML research direction is centered on determining which authored and derived facts must remain observable for reliable frontend analysis.

Primary areas of interest include:

• document structure
• element semantics
• attribute semantics
• parsing and recovery behavior
• authored source correspondence
• semantic ownership
• cross-domain boundaries with CSS and ECMAScript

A central research concern is the distinction between authored HTML source and the structure ultimately observed by the platform.

HTML parsing is not necessarily a one-to-one transformation from source text into document structure.

Parsing may involve:

• implicit structure
• normalization
• context-sensitive interpretation
• error recovery
• parser-generated relationships
• transformations that obscure direct source correspondence

Because of this, FA does not assume that a generic DOM representation is sufficient for every downstream analysis consumer.

The research instead asks which semantic facts need to survive and how their provenance should be represented.

Relevant questions include:

• Which source occurrence produced a semantic fact?
• Which component owns that fact?
• Is the fact authored directly or derived by the platform?
• Does document order need to be preserved?
• Are duplicate authored occurrences semantically relevant?
• Can the expected result be observed independently from production implementation?
• Which HTML facts should be exported to CSS or ECMAScript analysis?
• Which interactions require their own cross-domain research?

The intended research workflow remains falsification-first:

HTML authority → bounded semantic question → independent observation or oracle → positive and negative witnesses → attempted falsification → accepted evidence → production representation

No broad HTML representation should be frozen merely because it resembles an existing parser, DOM, or AST architecture.

Production representation should instead follow demonstrated semantic requirements.

The current objective is therefore not "support HTML" as a single feature.

It is to progressively establish small, evidence-backed HTML capabilities whose semantic meaning, source correspondence, ownership, and limitations are explicit.

As research progresses, individual areas such as document structure, attributes, parsing recovery, or source correspondence can be promoted into separate research records when their evidence lineage becomes large enough to justify independent tracking.

Until then, this summary serves as the current HTML research baseline.

#HTML #FrontendAnalysis #WebPlatform #StaticAnalysis #Research