Accessibility of PDF electronic books, case of DGD7
1. The PDF is tagged, this is good. In order for a screen-reader user to interact with a PDF the file needs to be parsed into an accessibility API / virtual buffer. Tagging makes it easier and faster for the file to be parsed.
2. Lack of proper structure. Tables and headings "look" like tables and headings, but they are not semantically represented as such. This makes it difficult to navigate between and within pages.
3. Alternative text for images. The images don't appear to have alternative text, although it does seem (I might be wrong) that some have captions. Captions do help where alternative text isn't available, but they sometimes are not sufficient to communicate the meaning. Alternative text should be written from the perspective of "What text needs to be here if the image isn't?"