(28-10-2026, 12:00 AM)quimqu wrote: You are not allowed to view links. Register or Login to view.Thanks @Mauro. About the horseshoe: yes, it’s a bit like ordering books by height. The arc is real, but it shows how the hands and sections differ, not the order the pages were written in. My text has no time order at all and still makes the arc.
About the LSA: you actually had it right. If you reorder the bifolios to make neighbouring pages as similar as possible, the score goes up a lot, and it goes up even more in my text than in the real one. If you try every possible order and keep the best, it always looks better than what you started with. So that gain doesn’t show the order was wrong.
About the 1/720: I tried it. I put the bifolios in order with Lisa’s method and with npcompl33t’s measure, on quires 1–7, 13 and 20. The two methods don’t agree with each other, neither on the real text nor on mine. So your odds look fair, but then I can’t reproduce the match either. @npcompl33t, how did you get your order for Q13 and Q20, and which of Lisa’s orders are you comparing with? With both sequences I could test it on my text.
Do you reproduce the ordering of Lisa's LSA on Q13 and Q20, the only ones on which they tried LSA sequencing? Their paper is here: You are not allowed to view links. Register or Login to view., they discuss sequencing ("Bifolia reordering experiments") from paragraph 50 (it's funnily not numbered by pages). Your work surely counts as a repetition of their experiment: the expectation is you should find similar results, if you don't there's something going on which I'd like to come to know (maybe hidden assumptions in the software implementations)? Same goes for @ncompl33t's results, so I hope you two will sort it out!