[Quoth]

Historically scanners and fax servers used Multipage Tiff. Developed by Aldus in 1987. Obviously it was a pure raster format with lossless compression. Adobe bought Aldus. The last change to TIFF was in 1992.
PDF was a development of Postscript and can have text, fonts, postscript programming, bit map images including TIFF, vectors and layers. One idea was so that the scanner or Fax could put a raster layer and an invisible layer of text could be added for search via OCR.
Aldus Pagemaker used Postscript on the Mac from 1985 to give true WYSIWG, which MS Word has never done, because it uses the local printer metrics.
By 1993 Aldus was owned by Adobe. The first PDF format was 1993!
All systems went toward PDF for files to a printer.
Quark Express and Ventura ate into Adobe Pagemaker share, so in 1999 Adobe released InDesign. It's primarily for paper so is orientated to production of PDFs. PDF output from Word and LO Writer has got so good that now only big publishers, people doing non-book publishing, specialist text books or more money than sense use InDesign. It was fudged for Azw3 and epub2, because it's for fixed layout paper, so it only does print replica and fixed layout epub3 well. Its roots are postscript in Aldus Pagemaker for the Mac for paper print. Postscript is practically the native content of a PDF, but a pdf is an envelope with multiple layers of anything in it, which is why Nebo or Xodo or Okular or Acrobat can add layers with drawing/handwritten/text layers as annotation. Editing what is already there is very hard to impossible.

Why didn't comics / Manga use multpage TIFF? Because those may be a vector or PDF source with maybe bitmaps as well as text. A TIFF is raster based. Compression historically is lossless and per line. Conversion of the comic/manga source to a jpeg or png per page gives adequate quality and when put in a zip gives better compression. Some comics/manga are distributed as mixed bitmap and text which allows TTS of the speech bubbles or zooming them.
Scanners, Fax and CRTs are raster (line structured). Modern screens and arbitrary images (from PDF or other) are arrays, bit maps.

Amazon is the odd man out on PDFs with the Scribe. Also they should never have marketed Print Replica as "Kindle".