Originally Posted by HarryT
PostScript was originally developed as a printer-independent page description language, to allow applications to output highly-complex page images to any printer. The idea was that you'd have software called a "Raster Image Processor" in the printer's firmware, to interpret the PostScript instructions and convert them into pixels for that printer. Used in that way, it was initially popular for 3D graphics applications and desktop publishing. It was only later that it caught on as a format for exchanging
documents with other people.
What happened was people would spool postcript print jobs into .ps files for archival purposes so all that was needed was to copy the file to the postcript printer port. (We're talking command-line interface era, here. ca 1985-90)
Adobe took note and in 1993 introduced Acrobat as an archival tool so the Postcript jobs could be viewed on screens and printed from a GUI environment. People who had been zipping and emailing .ps files slowly moved to pdf.
During the first ebook era some people and publishers took to using pdf as a distribution format for replica ebooks despite the size and usability issues because it was a one-way format; you couldn't easily extract meaningful data from a pdf. (Even today it's a non-trivial effort that often fails. At its core, pdf is still a print spool job with enhancements.)
PDF didn't take off for commercial ebooks because the files were almost useless on PDAs and none of the early ebook readers could afford the cpu power (or licensing) to render them. And because it didn't allow the end-user display options that even the first-gen ebook formats (pdb, prc, rb, imp, lit) allowed.
For years Adobe tried to extend PDFs to try to sell them as ebooks but it never really caught on and by the time the Kindle and, later, ePub readers, came out it became a hopeless cause. So they moved their ebook DRM scheme from pdf to epub and proudly proclaimed it a "standard" and started collecting DRM fees from lots of would-be players. A nice and timely hijack effort.
That has pretty much failed, so far.
Long-term? Who knows.