3 ms·
Long sympathetic with the Jacob Nielsen / PDF bad camp ... I've had some recent changes of heart. Not a full convert, but PDF is often superior to HTML, especi
by dredmorbius 5y ago
Long sympathetic with the Jacob Nielsen / PDF bad camp ... I've had some recent changes of heart. Not a full convert, but PDF is often superior to HTML, especially for longer-form and complex noninteractive content.
Books are an artefact whose design has evolved over the centuries to accommodate human-scale ergonomics: font size, paper and ink colour, words per line, lines per page, pages per volume, overall weight and dimensions. Standard-sized books are all larger than the largest current mobile phones, with diagonal measures of about 9--12 inches. There are smaller and larger books, but those are compromises either to portability (pocketbooks) or to large-format resolution and detail ("coffee table" books, atlases, and the like). Magazines tend to run even larger (about 13"), broadsheet newspapers larger yet. Most criticisms of PDFs are actually criticisms of the devices and displays used to read them.
Poor resolution, incorrect aspect ratios, and small display sizes (especially mobile devices) are the key problems.
Reading PDFs on a tablet, especially a larger e-ink device, is a game-changer. I now actively avoid HTML, or at least launch it in a browser designed with e-ink in mind (EInkBro: https://github.com/plateaukao/browser https://github.com/plateaukao/browser). Otherwise, my large (13.3") high-DPI (200+) B&W ebook reader is an excellent long-form immersive reading tool.
The key requirement of a mobile phone is that it fit in a pocket, handbag, or purse. They are too small for reading, and aren't designed for that purpose. Current devices feature screen sizes of roughly 5--7 inches (diagonal measurement). At the lower end, that's smaller than a 3x5 index card (6"), and the largest barely the size of a 4x6 card (7").
On desktops, the first display that offered what I felt was a truly comfortable two-pages-up PDF reading experience was the 27" Retina iMac. Its 5K display (itself an oddball size) suits document work well. Even not fully maximised, most books are highly readable (leaving screen space for other tasks), and at full maximisation, details really stand out especially from scans of historical editions. (Such details aren't always relevent, but often are.)
PDF also provides capabilities HTML either cannot or does not by default (and few seem to be persuaded to offer), especially pagination, formulae, and a spatially-persistent layout (if you have a spatial memory, this is very valuable).
PDFs can though often do not include internal navigation (chapters, sections, etc.), search (if full text is included), and most critically, metadata (at a minimum, author, title, date, and publisher, see the full Dublin Core metadata specification for what should be required).
PDFs can also be published directly to device sizes (or to a set of form factors encompassing typical devices), as several others note.
Some of the issues aren't entirely intrinsic, and my feeling is that wider use of PDFs for online content would lead to a proliferation of PDF annoyances to match present-day Web annoyances. In each case, the fundamental problem is that publishers rather than readers have final say over presentation. An alternative, of distributing raw minimum markup and formatting that to user specifications following a set number of templates ... might help.
It's ironic that the article here embodies a number of PDF annoyances:
- The shaded background renders quite poorly on a B&W e-ink reader (though can be eliminated with a watermark-removal setting).
- The filename provides no clues as to contents or provenance, and is likely to collide with other content.
- I'm a fan of serif fonts, not sans serif, for high-DPI reading.
- Internal and external hyperlink support is ... variable. At times utterly missing, at others, inconsistent or inconvenient.
- PDFs are not trivially directly editable, which means both authors and readers can change errors or address issues.
- Many PDFs lack internal structure, even where the document they encompass do. The number of books lacking PDF table-of-contents support is ... large.
- Metadata standards and practices are abysmal. See the Dublin Core standards.
- Naming conventions similarly. "Report", "Resume", "Project", or "0.pdf" are names which should never be used. Describe author, content, and date, as a minimum, if possible.