calculating new reference points using information from glitch detection

[claude]

In this post I hinted at a method to calculate new reference points using the information available at glitch detection time:
http://www.fractalforums.com/announcements-and-news/pertubation-theory-glitches-improvement/msg84154/#msg84154

I got around to writing a demonstration program (attached) with lots of inline comments (but if you have questions, I can try to answer).

It's deliberately minimal, it doesn't even calculate a good primary reference point by itself, let alone series approximation...

But hopefully it's useful to see how it works, and I'd like if the method were implemented in other programs and evaluated more extensively - maybe it's not as good as I think it is