The cover uploaded, but body images are still wrong: checking website images with an asset list

Successful cover uploading establishes one operation in the cover field, not correctness of all body images. Old references, swapped positions, and replaced images with old alternative text can remain.

Connect purpose, file, upload URL, and placement for each article image. Define each task and inspect actual saved pages to reveal issues folders or editor thumbnails miss.

Give each image a purpose and identity

List usage positions first: cover, section illustration, or paragraph workflow. "First image" or "Attractive image" becomes ambiguous after reorder or reupload. Combine article identity with image-purpose numbering and related content.

Covers identify articles and preview in lists; body images explain passages. Similar style does not make them interchangeable. Ratios, subject scales, and context differ, and a cover may not answer a section's question.

For one cover and three body images, assign four separate files, descriptions, and placements. This hypothetical count does not require four images in every article.

JVDS Design Studio's confirmed article 597 entry used a cover and three body images with saved-content checks. Individual matching and reopening are useful methods; this does not establish equivalent verification of other articles.

Purpose placeholders can remain in source drafts until real assets arrive. Replace them with actual images in final articles, without exposing internal numbers. Describe positions by sections and paragraph relationships rather than fixed line numbers that change.

Distinguish versions too. Regenerated or cropped images need updated records so staff do not mistake old files for current ones. "Final" in a filename alone is insufficient.

Maintain correspondence from local files to uploaded URLs

Original, encoded publication file, and upload address are different objects. Originals support adjustments, publication files undergo format or size changes, and websites assign paths. Retain this chain beyond local names.

Uploads may rename files. Do not assume local names survive or place local filesystem paths as public URLs. Record actual upload outcomes against purposes.

For multiple uploads, record returned files individually. Apparent order is not guaranteed, especially with retries and duplicate names; positional guessing can assign wrong addresses.

Compare similar-ratio, similar-style images together for subject differences before uploading. Dark scenes can look alike, and correct numbers may still conceal repeated content. Confirm distinct information tasks to reduce placement confusion early.

Open encoded files too. WebP extensions cannot prove correct content; resizing may crop essentials and export may use an earlier version. Inspect actual pictures beyond file existence.

Do not delete all historical resources before replacing. Others may share files. Identify current references and replacement scope before cleanup so unrelated pages remain intact.

Check placement and alternative text image by image

After insertion, check section sequence, nearby relevance, duplication, omissions, and reading interruptions. Correct URLs in wrong sections are still entry errors.

Alt text should describe current contextual purpose. Informative diagrams explain relationships; decorations follow actual use. Different images should not share keyword strings or retain old descriptions after replacement. See How to Write Effective Image Alt Text for related checks.

If an upload-flow illustration becomes a material-checking desk, update its alternative text too. Continuing to say "Server upload workflow" gives wrong information when images fail or assistive tools read them.

Filenames, captions, and alt text need not match exactly. They support management, article connections, and visual information respectively. Mechanical copying misses distinct tasks and may create repetition.

Include alt text in review lists. Asset owners describe scenes, article owners check passage relationships, and operators enter approved wording. This avoids improvised descriptions and reveals planned-prompt scenes differing from generated results.

Editor sizing and alignment should match real content areas. Check subjects and edges on phones, beyond desktop clarity. Image content and display rules must agree.

Reopen after saving to confirm real references

Normal editor display does not establish retained results. Reopen to check cover fields and body URLs, placements, captions, and alt text. Authorized implementers can inspect actual stored content where needed.

Reload previews and inspect each clear image. Editors may show temporary local previews despite invalid public paths. Actual saved loading is the useful evidence.

Devices may choose different responsive sizes. Check mobile candidates also match current versions. Updated desktop with old mobile files needs records of system-generated version relationships.

Inspect covers in lists or cards and body images with full passages. One module cannot establish another's acceptance; fields and crops may differ.

For old images, compare actual references to the list before file or cache investigation. Reuploading everything or clearing browser cache alone cannot correct old URLs still in body text.

Use authorized draft previews without temporary publication. Resource and publication states can be checked separately; image inspection must not change planned publishing.

Use asset lists for replacements and handover

At minimum, record purpose, section position, original, publication file, actual address, version, and alt text. Authorization, prompts, and crop rules depend on source and maintenance, while asset origin should remain explainable.

Originals enable exports and publication files reproduce effects. URLs alone can leave migration without source files; originals alone leave later staff unsure which version is public. See Corporate Website Deliverables: Designs, Assets, and Code for related checks.

Update the matching purpose and recheck references during replacement. Upload success does not change an old body URL, and list updates do not save articles. Keep lists and actual records synchronized.

Ask a new maintainer to locate one original, current publication file, and body position from the list. Needing the old operator reveals missing identity or version information. This checks practical maintenance without complex systems for ordinary files.

Bulk imports need article identities in naming and records. Image numbers may repeat across articles, but combined identities should be unique, preventing one article's third image entering another.

Success means correct files and actual URLs at every agreed position, usable saved loading, contextual alt text, and discoverable source and publication versions for handover. Verification completes a relationship, beyond an upload.

Frequently asked questions

Why can body images stay old when the cover is correct?

They may occupy separate fields. New covers do not automatically change body URLs. Check every body reference, placement, and version against the list.

Is server-side renaming an error?

Not necessarily. Record actual returned addresses and pair them with local files and purposes rather than guess final paths from original names.

Does replacement require rewriting alt text?

Check accuracy. Changed content or purpose needs updates; unchanged re-encoding may retain descriptions. Avoid copied keyword lists.

Can unused old images be deleted immediately?

First check other-page references. Cleanup follows real relationships and maintenance rules; replacement in one article does not establish a shared file is unused.

How can draft images be checked without publication?

Use existing saved-content checks and authorized previews for actual URLs and loading. Preview availability depends on the system and does not justify unauthorized status changes.