Actually, it is not trivial to programmatically extract plain text from a PDF in a consistent format. In many cases it is easy, but the PDF format is visual first and content second, resulting in plenty of opportunities for Captcha-like problems.

Despite that, PDF, especially PDF-A, is the superior format for preserving published content.

Huh, I didn't realize that PDF hasn't been propriety format since 2008. That's awesome.

Reading plain text files is simpler to do across platforms. If an app on a mobile device just pays attention to the text then that's fine for functionality, I guess. But then I have a problem with the extra formatting being there, unused. Kind of like having a problem with unused CSS in a style sheet just sitting there.