Pagination widows, or, why I'm embarrassed about my eBook (2023)

Original Article ↗ Hacker News Discussion ↗

Ebook typography still lags far behind print, with readers trading classic layout quality for reflowable text that often produces widows, orphans, broken headings, and poorly handled figures, tables, and formulas. Commenters describe how weak CSS and paged-media support in browsers and major ebook engines (especially Kindle and ADE) forces publishers into hacks, crude conversions from print tools like InDesign, or fallback to PDFs. Many argue that for technical, mathematical, or design‑sensitive works, high‑quality PDFs or LaTeX‑driven workflows remain superior, while EPUB is acceptable mainly for simple narrative text.

CSS and layout hacks for headings, widows, and orphans

  • Commenters explore how to prevent orphaned headings in reflowable layouts.
  • Suggestions include using large positive spacing on headings with a matching negative margin or padding on the following paragraph so the visual gap disappears but pagination respects a larger “block”.
  • Others point out edge cases (e.g., enough room for the “fake” multi-line heading at the bottom of a page) and note these hacks may cause big gaps or incorrect breaks.
  • Some mention the orphans CSS property, but support is limited; ebook CSS support is compared to HTML email: fragmented, buggy, and inconsistent.

Ebook renderers: outdated and inconsistent

  • Many criticize Kindle and Adobe Digital Editions–based engines as the “IE5/IE6” of ebook rendering: dominant, old, and blocking progress.
  • iBooks and some WebKit-based readers are seen as top tier, but even they have quirks (e.g., unjustified design choices or incomplete CSS).
  • Poor CSS implementation forces ugly workarounds (tables for layout, grouping headings with first paragraphs in a block, etc.).
  • People ask for something like “caniuse” for ebook engines; current reality is trial-and-error across many broken implementations.

PDF vs EPUB and reflow vs fixed layout

  • Strong split in preferences:
    • EPUB is favored for fiction and anything read on phones or small devices; reflow and user control of fonts are appreciated.
    • PDF is preferred for technical, mathematical, and richly laid-out content (figures, tables, code, cookbooks).
  • Complaints that many publishers only ship badly auto-converted EPUBs and withhold PDFs, despite already having high-quality print layouts.
  • Debate over page sizes (A4/Letter vs A5/A6) and how poorly fixed-page PDFs work on small screens; some argue for multiple size-specific PDFs, others see this as unscalable.
  • Some wish PDF itself were “responsive”; others note HTML already had reflow but has been misused into pseudo-fixed layouts.

Math and complex typography

  • Math in ebooks is widely seen as terrible, especially on mainstream readers.
  • Using images for formulas is criticized: baseline misalignment, non-matching fonts, low resolution, poor scaling, and accessibility issues.
  • Vector formats (SVG, PDF fragments) and proper markup (with alt text) are suggested as better but often unsupported or unused.
  • Broader point: high-quality pagination, figure placement, and typographic control is a hard, craft-like problem; current ebook tools and renderers underdeliver.

General state of ebooks and paged media

  • Many feel ebooks have stagnated: DRM, monopolies, and lack of investment trump features and typographic quality.
  • Web/CSS paged media support for printing is also called a “never ending mess,” pushing people to HTML→PDF pipelines or desktop tools for anything serious.
  • Some readers prioritize faithful print-like aesthetics; others want minimal publisher styling and robust semantics so devices can render cleanly to user preference.