I've written retrospectively about pristine-tar before, when I stopped maintaining it. So, I'll quote part of that here:
[...] a little bit about the reason I wrote pristine-tar in the
first place. There were two reasons:
1. I was once in a talk where someone mentioned that Ubuntu had/was
developing something that involved regenerating orig tarballs
from version control.
I asked the obvious question: How could that possibly be done
technically?
The (slightly hung over) presenter did not have a satesfactory
response, so my curiosity was piqued to find a way to do it.
(I later heard that Ubuntu has been using pristine-tar..)
2. Sometimes code can be subversive. It can change people's perspective
on a topic, nudging discourse in a different direction. It can even
point out absurdities in the way things are done. I may or may not
have accomplished the subversive part of my goals with pristine-tar.
Code can also escape its original intention. Many current uses of
pristine-tar fall into that category. So it seems likely that some
people will want it to continue to work even if it's met the two goals
above already.
For me, the best part of building pristine-tar was finding an answer to the question "How could that possibly be done technically?" It was also pretty cool to be able to use every tarball in Debian as the test suite for pristine-tar.
I'm afraid I kind of left Debian in the lurch when I stopped maintaining pristine-tar.
"Debian has probably hundreds, if not thousands of git repositories using pristine-tar. We all rely now on an unmaintained, orphaned, and buggy piece of software." -- Norbert Preining
So I was relieved when it finally got a new maintainer just recently.
Still, I don't expect I'll ever use pristine-tar again. It's the only software I've built in the past ten years that I can say that about.