Hacker News new | ask | show | jobs
by xcjs 962 days ago
I love projects like this, but I really wish instead of converting text resources to web sites that these projects would produce epub outputs. It's great for distribution, offline reading, and scaling to different display sizes, aspect ratios, and resolutions.
3 comments

I wonder if there will be a time when textbooks will be created in digital-first format, instead of being mere replicas of what print books are. It doesn’t have to be static text and images on A4 pages.
The page links to a pdf version[0], which can easily be converted to epub using Calibre[0], which is free and open-source.

[0] https://www.gutenberg.org/ebooks/33283

[1] https://calibre-ebook.com/

I'm familiar with those tools, but they don't preserve the formatting of the source document (when PDF) and certain types of embedded resources don't transfer properly.

Some PDF documents don't contain the source text as digitized text, either. It's just a bundle of scanned images.

Epubs are horrible for technical documents
Only if they exist as poor conversions. Properly formatted epub documents are worth their weight in effort.

Epub documents _are_ HTML files in a zip archive - I'd argue that if a web site is a proper presentation of the source material, an epub is even better.

Additionally, PDF documents _are_ worse for any kind of document technical or otherwise. PDF documents are primarily Postscript instructions without any relation or hierarchy to the included elements. Epub documents/HTML provide semantic relationships and hints to the content.

I think that what previous poster is pointing is that technical documents as epub are often difficult where they're no textual contents. Images may be the easiest case. Things are a bit more clunky when code source are involved and it's weird with complex formulas.
A competent stylesheet will sidestep all those issues. Admittedly, I do wish MathML were in place to be used online and in epub documents as that would make the problem much more easily solvable.
Epub3 (not previous versions) does support MathML. The main concern is that most e-readers are not as capable as main browsers we use. Very few e-readers have good support of Epub3 features, and maths rendering is limited when available.

For codes, `pre` tag should be enough, stylesheet just adding some enlightenment. What I'm pointing is that many source code are more than 80 columns wide and that often make you scroll like with PDF. Physical readers are best for content that use aspect-ratio of pocket paper books and as far I can remember coding paper books are more (at least twice) wider and may even suffer that problem (they're often hacks I dislike to split lines.)

To short, Epub per se is not the limiting factor. Vendors must update their hard and soft (even then, things will improve only when everybody has upgraded.) Authors also must take more care to technical contents for that media (some of them convert formulas and codes to image, oooch…)

EPUB works fine for documents that mostly consist of easily reflowable text. For documents with lots of equations, diagrams, tables/charts, footnotes, etc. I would much rather have a PDF.