Review Native HTML Headings and Compare Outlines

Harbor Repair, a fictional community repair workshop, is revising a website guide to replacing a drawer runner. Several volunteers have edited the instructions, leaving an uneven heading sequence. Their review shows why eight extracted headings are not automatically eight useful sections, and why equal heading totals can conceal several exact additions and removals.

Distinguish a native outline from a visual reading

Start with the element names, not the typography. The checker inventories native H1, H2, H3, H4, H5 and H6 elements in the supplied HTML. It does not promote a bold paragraph because the words resemble a section title. Nor does an ARIA heading role become a native H element in this report. Those other structures may matter in a broader review, but they are not additional native headings waiting to be counted.

An H3 styled to look enormous is still recorded at level three. An H2 reduced to a small caption remains level two. This separation is useful when a website template hides its structural choices behind consistent visual styling. Read the outline to understand the supplied markup, then examine the implemented page separately to understand the appearance. A source inventory cannot confirm what is displayed, spoken by assistive technology or available at a particular screen size.

Treat each extracted heading as a claim about the following material. Measure the old runner should introduce measurement instructions; a section about ordering parts needs a different promise. The checker can show the label and level but cannot decide whether the section fulfills that promise. Reading the paragraphs remains necessary. A tidy sequence can still organize inaccurate instructions, while an awkward sequence can sit above genuinely useful explanations.

A useful review outcome is a set of justified choices: retain the page heading, promote a preparation section, rename two confusing labels and remove a leftover empty element. It is not a certificate of accessibility or an SEO ranking score. Do not interpret the absence of findings as approval of a page’s meaning. The inventory is evidence for a person making those decisions, not a substitute for that person’s judgment.

Prepare comparable source snapshots before diagnosing levels

Record what each snapshot includes. Harbor Repair supplies a complete page export for the first review, with the article, navigation and footer. If the next export includes only the article, shared template headings may appear to have been removed even though no one edited them. Compare like scopes when the question is revision. When scopes deliberately differ, describe that difference in the handoff instead of calling every removed record an editorial deletion.

The checker parses supplied HTML as a document without executing its scripts or loading resources. Head, script, style, template and noscript content do not contribute heading records. A heading written inside a script string is not an extracted article heading. A section created only after a website script runs is also not recovered from an absent source element. Obtain the relevant snapshot through your normal tools; entering its address here will not retrieve it.

Both inputs share one chosen mode and explicit base. HTML mode supplies native outlines; text mode provides counts, not inferred headings, even when a line starts with H2 or a Markdown marker. Converting a document to text can discard the levels you wanted to compare. Name article fragments as fragments: an absent H1 may belong to an omitted template. The comparison source receives a full inventory, not its own review checklist; findings concern the primary source only.

Use project names and dated files to distinguish revisions. Keep a short external note explaining who exported each snapshot and whether it came from a draft or a published page. The platform field merely labels the website project; it does not change heading rules or inspect a native social profile. For this repair guide, all platforms is sufficient. A website resource supporting one channel would still use exactly the same native heading inspection.

Inspect the page heading and empty labels in context

Harbor Repair’s supplied sequence is H1 Drawer runner replacement; H3 Before you start; H2 Measure the old runner; H4 Record the mounting pattern; H2 Fit the replacement; H2 Check the movement; another H2 Check the movement; and an empty H3. The eight records include one H1, four H2 elements, two H3 elements and one H4. Write the sequence out before editing so isolated flags do not obscure the overall shape.

The single H1 gives the document an identifiable main label in this snapshot. Other inputs may trigger no H1, empty H1, multiple H1 or first-heading-not-H1 findings. Read those as separate questions. A fragment beginning with H2 can have both no H1 and a first heading other than H1 for the same scope reason. Two flags do not necessarily require two edits. Establish whether the missing top-level material was omitted from the input.

An empty element needs a different investigation from an unclear label. Harbor Repair locates its empty H3 after a removed parts photograph. If the whole section was deleted, removing the leftover heading may be appropriate. If instructions still follow, the editor needs to restore a useful heading instead. The checker does not determine which explanation is true. Trace the source line and surrounding content before choosing between deletion and completion.

Do not pad an empty heading with a keyword merely to clear the finding. The replacement should tell a reader what comes next. Likewise, do not delete a meaningful main heading solely because another H1 exists elsewhere in a supplied full-page export. Review the page’s actual structure and editorial responsibilities. A practical decision describes both the element and its intended role, making it possible for another editor to understand why the change belongs there.

Use level jumps to question the intended hierarchy

A descending jump moves toward a deeper heading level while skipping an intermediate level. The repair draft moves from H1 to H3 at Before you start, then later from H2 to H4 at Record the mounting pattern. These are two separate transitions to review. Returning from H4 to H2 is a move back toward a broader level, not another descending jump merely because the number changed by two.

Ask which section owns the material. Before you start explains the preparation needed for the whole repair, so the editor proposes H2 alongside the other major stages. Record the mounting pattern elaborates on Measure the old runner, so H3 is a plausible subordinate level. These decisions follow the explanation, not a rule that every page must use every available heading level. There is no need to add H4, H5 or H6 when the content has no corresponding depth.

Inspect what happened before changing the markup. A copied block may have retained its old level from a deeper guide. A deleted parent section may have left its child headings behind. Alternatively, the supplied fragment may begin midway through an established hierarchy. Those cases can produce similar transitions but call for different decisions. Adding a vague bridging heading such as More information can hide the numerical jump without improving the structure or answering the reader’s next question.

The report follows parsed source order. CSS arrangements can place blocks differently, and hidden content is not filtered by visual visibility. Review that behavior outside the checker when it affects the published experience. The tool does not evaluate rendered hierarchy or make an accessibility compliance determination. For the handoff, describe the proposed level and relationship: make mounting pattern a subsection of measurement. That is more useful than a bare instruction to reduce jumps to zero.

Separate repeated heading wording from duplicate identifiers

The repair page contains two nonempty H2 labels reading Check the movement. The duplicate-text finding groups repeated normalized wording for review. It does not establish that the sections have identical paragraphs, compete in search results or should be combined. Locate both occurrences. One may discuss free movement with an empty drawer while the other discusses movement after loading it. Different tasks can deserve more specific labels rather than a merged section.

Harbor Repair proposes Test the empty drawer and Test with the usual contents. Those labels distinguish two checks the volunteer actually described. If both sections instead repeat the same instructions, consolidation may be clearer. Do not rename headings merely to manufacture uniqueness while leaving confusing repetition underneath. Read the surrounding material and decide whether there are genuinely two responsibilities. The checker compares normalized text, not meaning, and cannot resolve that editorial question.

Duplicate IDs are a separate source concern. Two headings can have different labels but both declare id="movement"; two identical labels can also have distinct IDs. An ID finding does not become a duplicate-text finding, or vice versa. When revising identifiers, inspect references to their fragments in the website workflow. The checker can inventory a fragment-bearing link, but it does not confirm live navigation behavior or automatically update incoming references elsewhere.

Use ordinals and source lines to distinguish repeated occurrences. A task saying change the second Check the movement is clearer when it also identifies the supplied file and location. After editing, line numbers may shift, so keep the old snapshot when exact provenance matters. For the address side of this task, consult the metadata and link review guide; heading text, identifiers and link destinations are related evidence, not interchangeable measurements.

Calculate an exact heading multiset difference

For comparison, consider a smaller five-heading baseline from the same repair project. It contains H1 Drawer runner replacement once, H2 Measure the runner once, H2 Fit the replacement twice, and H3 Check clearance once. The revised inventory contains the same H1 and measurement heading, H2 Fit the replacement once, H2 Check clearance once, and H2 Select fixings once. Both versions total five headings, but their matching keys differ.

Each key combines the native level with normalized text. The H2 Fit the replacement key falls from two occurrences to one, producing one removal. H3 Check clearance falls from one to zero, producing another removal. H2 Check clearance rises from zero to one, producing one addition, while H2 Select fixings produces another. The arithmetic is five baseline occurrences minus two removed plus two added equals five revised occurrences. An unchanged total has not meant an unchanged outline.

This is a multiset rather than a simple set because occurrence counts matter. Removing one of two identical heading keys does not remove that key from every location. It is also not a semantic diff: the tool does not decide that the H3 and H2 clearance entries represent the same topic at a revised level. The editor can explain that relationship, but the reported evidence remains one removed key and one added key.

Read the comparison’s stated input direction before using additions as a work list. If you reverse which supplied version is the baseline, additions and removals reverse too. Keep each snapshot identified, then reconcile the literal results with the planned edits. Normalized text is the matching value, not an inference about synonyms. A differently worded heading may express the same idea and still appear as a removal plus an addition. That is expected evidence, not a failed understanding of the article.

Read unchanged matches and renamed headings honestly

Moving a matching heading to another position can leave its multiset count unchanged. Suppose Select fixings is moved before Measure the runner while keeping its level and normalized label. That reordering may improve or damage the repair sequence, but it does not necessarily produce an addition or removal. The comparison is not reliable moved-heading detection. Read both ordered outlines when the order of instructions is the actual subject of review.

A rename creates the opposite interpretive trap. Replacing Fit the replacement with Install the new runner can yield a removed key and an added key even if the underlying instructions remain identical. The result alone cannot establish an added topic, deleted advice or a substantive rewrite. Keep an editorial note connecting the old and new wording when that matters. Do not convert every added heading into a claim that the document gained new information.

The outline also stays unchanged when an editor corrects a measurement, changes a product reference or reverses an instruction beneath an existing heading. None of those paragraph changes alters a native level-and-text key. The comparison does not detect stale dates, evaluate factual accuracy or prove that all content remained intact. Read the actual revised instructions, especially where a mistake would make the task hard to complete. Use a separate document comparison when paragraph-level differences are needed.

For Harbor Repair, the final human read follows the intended sequence: identify the runner, record measurements, select compatible fixings, fit the part and test movement. The source inventory helps locate those steps but cannot decide whether the sequence is sufficient for every drawer design. Resolve uncertainties in the article itself. A reassuring zero in an additions or removals column is not a substitute for checking the decisions that the revision was supposed to address.

Package the outline review for a practical handoff

Return to the structure checker and review the current findings against the agreed editing tasks. Marking a finding reviewed records consideration, not a repair. A retained repeated heading can remain flagged with that status. Any source, mode or base change clears review exceptions, so a revised snapshot needs fresh attention. Make the actual edits in your website tools; this checker cannot change heading elements in a CMS.

Save JSON to preserve both raw sources and the compatible primary review state. You can save a blank-source draft, but inspection needs nonempty primary input. There is no autosave. Keep dated files for history and assign one current owner; colleagues’ copies do not merge automatically. Extraction caps apply separately to each source: 200 native headings, 400 links, 200 images and 400 metadata records. More than 500 primary findings causes explicit refusal, not a partial checklist.

CSV retains the extracted inventory for both sources plus primary findings; copy checklist supplies the primary tasks. The PDF contains scope, complete extracted records for both and primary findings with status, not a live rendering or raw HTML. Retain JSON for the raw material. If the PDF exceeds 100 pages or 40 MiB, use JSON or CSV instead of an incomplete substitute. Keep the source labels attached when sorting records so the two outlines do not become mixed evidence.

Finish with a concrete acceptance note: preparation is now a major section, mounting detail is subordinate to measurement, repeated checks have distinct labels, and the empty element was removed in the website draft. Then inspect the implemented page separately. Social promotion is a subsequent workflow, not something triggered by this export. The tools directory offers other focused preparation helpers; SMTasker can organize supported social routines without editing or automatically importing this website review.

Questions about Heading Outline Review

Continue with another content review guide

Inspect your supplied draft in the Content Structure Checker, or choose a focused guide below. More browser-based helpers are in the free tools directory.