EPUB 3 is Awesome

Generate PDF

·

The bulletin from Part 1 and Part 2 is an EPUB now. The Research Centre still needs a PDF, because a PDF is what gets archived and what gets printed. In SEED.html's Projects view there is a button for that, Generate PDF. I pressed it. This is what came out.

The contents page

The original bulletin's first page: a purple panel headed Bulletin 39, listing the article titles in plain text with no page numbers.
The original: a list of titles.
The generated PDF's first page: the same list of titles, each one a blue underlined link, with a page number at the right of every line.
The generated PDF: the same list, every title a link, a page number on every line.

The contents page in the EPUB is a list of links, because that is what a contents page is in an ebook. In the PDF the same list has a page number on every line. The numbers came from wherever each article landed when the book was paged, and the links came through unchanged, so a reader of the PDF can click as well as turn.

The events

The original's news page: a dense list, each line starting with a bold numeric date such as 25-27.07.2025 followed by a dash and the event, justified to both margins.
The original: dates as numbers, run into each line.
The generated PDF's news page: the same events in two columns, the date in words on the left, 25 to 27 July, 12 to 14 October, and the event on the right.
The generated PDF: dates in words, in their own column.

The same list that a screen reader read as "12 to 14 October" in Part 2, set for paper. In the original the date is the start of each line; here it has a column of its own, so the eye can run down the dates the way a listener runs down the list. Nothing about the printed page was decided separately. It is the same list, and the rule that spelled the dates out for the listener spelled them out for the page.

An article

The original's page for the soinari article: the section name in italics at top left as ordinary text, a centred title, justified paragraphs with first-line indents, and the photograph set into the text with no caption.
The original: the photograph sits in the text without a caption.
The generated PDF's page for the same article: the section name as a small running head in the margin, a larger centred title, ragged-right paragraphs, and the photograph with its caption, Varden Meparishvili playing a soinari, beneath it.
The generated PDF: the same photograph, with its caption.

The PDF inherits the layout decisions made for the reflowable EPUB's readability; limiting line length while maximizing use of space for images, captions on images, bulletin sections moved to page headers

The one thing that did not come out

The left rail of Chrome's PDF viewer showing the original bulletin: two icons, one for page thumbnails and one for the document outline.
The original, in the browser's PDF viewer: thumbnails, and an outline.
The left rail of Chrome's PDF viewer showing the generated PDF: one icon, for page thumbnails only.
The generated PDF, in the same viewer: thumbnails only.

Open the original in Chrome's PDF viewer and the left rail offers two things: thumbnails and an outline. Open the generated file in the same viewer and it offers one. The generated PDF has thirty-two headings and no bookmarks, because the browser that renders it does not write them. The original has two headings and two bookmarks, because Word does. On this one measure the new file is worse than the old, and the PDF viewer makes it clear.

What I did

I pressed one button. The PDF is a convenience for a Research Centre that still archives one; the book is the EPUB. The headings, the descriptions on the photographs, the language of the text and the links in the contents were all in the EPUB already, because they were decided while it was written. The PDF inherited them. The one thing the EPUB could not hand on, the outline, is the one thing the renderer did not produce.

The PDF is here. The EPUB is here. Both were made with SEED.html.