Use the feed for discovery and the canonical page for choice
A feed item should remain useful in a minimal reader that displays only title, date, description, and link. The canonical HTTPS page then explains the edition, rights, formats, file sizes, accessibility information, and current availability. This two-step route works across mobile operating systems, desktop browsers, in-app browsers, and e-reading applications without assuming one installed tool.
Match representation to reading task
| Representation | Best experience | Design note |
|---|---|---|
| RSS | Following updates and opening concise summaries | Keep HTML conservative and make the item understandable without images |
| Responsive HTML | Universal context, excerpts, OCR transcripts, and accessible fallback | Use semantic headings, stable anchors, keyboard access, and print styles |
| EPUB | Reflowable long-form reading in compatible systems | Validate navigation and use the correct MIME type and stable filename |
| Fixed-layout pages, facsimiles, citation, maps, and complex visual material | Label file size and preserve page references | |
| OPDS | Browsing and acquiring publications inside compatible clients | Offer a human-readable catalog page alongside the machine-readable endpoint |
iOS and Android
Use standard HTTPS links and correct media types so the operating system can offer installed applications. Avoid depending on one proprietary deep link. A labeled web fallback should always remain available. Test the complete path: open the feed item, identify the edition, save or open the file, return to the canonical record, and understand any login or purchase requirement.
Touch targets, text scaling, orientation, dark mode, slow networks, and in-app browsers all matter. Do not place a long excerpt inside a fixed-height card. Avoid pop-ups and third-party cookie requirements. State large download sizes before the user activates the link.
Safari, Chrome, and browser reading
Modern browser workflows often rely on feed readers, extensions, or web services rather than displaying a polished RSS interface directly. The site should therefore expose RSS autodiscovery and a visible feed link, while every item opens to a responsive page. Browser pages can provide accessible HTML, table-of-contents anchors, images, revision notes, and authorized downloads.
Stable anchors matter. Do not derive permanent links from a heading’s position or temporary script state. Preserve aliases when chapters move. A saved feed link may be revisited years later.
E-readers and OPDS
EPUB-capable devices and reading applications vary in how they import files and support OPDS. Publish standards-compliant files, meaningful metadata, and clear acquisition relationships. Do not promise universal support. Test with representative clients used by the audience and keep a web catalog available for devices that cannot consume OPDS.
Accessibility across representations
- Use descriptive link labels such as “Download EPUB” or “Open PDF facsimile,” not repeated “Read now” buttons.
- Publish known accessibility features only when supported by source metadata or review.
- Provide alt text, captions, semantic headings, visible focus, sufficient contrast, and keyboard navigation on the web page.
- Offer reflowable alternatives when a fixed-layout format creates barriers and the rights allow it.
- Keep OCR review status and page-image references visible.
Cross-device QA matrix
Task: identify edition → open canonical page → choose format → return to record
Devices: iOS phone, Android phone, tablet, desktop browser, screen reader, e-reader app
Failures: no compatible app, expired access, offline connection, withdrawn file, missing imageTest tasks rather than screenshots. A route is successful when the user understands what the item is, why it appeared, what access is required, and which format will open. The full cross-device Book RSS guide expands the test matrix and handoff rules.