So many things that people want to place online as “publications” still come as PDFs. Books, manuals, research papers, you name it. Nothing wrong with that, right? Wrong. PDFs are no longer useful for most of their applications, and you should use a better format that stores information at the right layer of abstraction.
PDF was (and is) intended for desktop publishing. Its intended output is the pages you see on your screen or printed onto paper. As someone who receives a PDF, you get pages rather than text. This makes perfect sense for desktop publishing. You get pixel-perfect output on your screen, and you can get the same thing on paper. In that sense PDF is not really obsolete when used properly. It’s great for storing a rendered output, ready for printing.
But what sorts of documents are often published as PDFs today? Almost any document that the sender wants to give as a single file, regardless of whether or not the intended application is desktop publishing. Let’s take one example: scientific papers. These are almost universally distributed as PDFs today. It’s true that some sort of rich document format is needed to store and format the text, headers, images, etc. But the exact form that the document is rendered in is not nearly as important as the semantics being displayed.
Let’s say that I find a scientific paper – distributed in PDF – that I want to read. I can read it on my computer okay, but what about reading it on my Kindle? Many PDFs do not work well on it due to the different screen size and such. It’s much more difficult to read it on a Kindle than it would be if it was in an e-book format. The irony is that most scientific papers are originally written in LaTeX – a text markup format – and then rendered as PDF. LaTeX can be rendered as an e-book easily enough, but they just aren’t, and the PDF often loses information that makes the conversion impossible.
The problems with PDFs are most apparent with things like page sizes. Let’s say someone generates a PDF with a page size of US letter, and someone in Europe wants to read it on paper. But the standard paper size there is A4, not letter. The PDF either has to be scaled or (possibly imperfectly) converted to letter. The desktop publishing abstraction creates entirely unnecessary problems here.
E-book formats like epub are far superior for most applications where PDFs are currently used, because they encode the actual document semantics rather than one possible rendering output. They can be rendered well on computers, but in any screen size, and can be easily read on e-readers or paper as well.
A similar problem used to occur – but has been mostly solved – in web design. For many years professional websites were created from fixed-pixel layout templates. The site would be designed to be exactly 800px in width perhaps. This let the designer attempt to control every pixel, but ran into problems from the start since different computers had different resolutions. The problem became insurmountable with the advent of smartphones and tablets, and modern web design is intended so that the same page can be seen beautifully at many different screen sizes and resolutions.
Not coincidentally, the epub format is effectively HTML and images packed into a single file. This lets the reading device do the page layout rather than the author. This is much better for both parties. The author does not need to worry about things like typography and formatting that they often don’t know much about, and the reader can read the text in a situation that’s best for them. If they want to use an e-reader, they can. If they’re dyslexic, they can read the text in a special font. If they’re blind, they can have the text read by voice synthesis. All of these things are much more difficult or impossible with PDFs. And with epubs the author can still control document semantics, for example making document headings, making text emphasized, setting off page breaks, and so on.
I don’t expect the world to switch overnight, but if you’re distributing something in PDF, please also try to provide an epub. At least some of your readers will thank you.