"The same should hold for Theorem 1, as will be discussed. We avoid the O(1) notation as far as possible, as it can be severely misleading (e.g. 10^42 = O(1)^O(1)) = O(1)).
This work could be viewed as another O() warning showing, how important factors, and even subdominant additive terms, are."
Right, the critique is evaluating algorithms based solely on asymptotic time / space characteristics is insufficient. Constants can be significant if they are big enough.
with arXiv the TeX source is available for download (under the "Other formats" link), which I think is what arXiv-vanity uses. I imagine it would be really tough to go reliably from arbitrary PDF to HTML without screwing up the math and figures.
I mean firefox' built in pdf-reader is javascript based. I imagine it renders to a canvas instead of to html, but it doesn't seem like an impossibility to render to html in any case.
PDFs can be annoying on a small tablet or phone. It seems like I used to have more display options, but I tried to catch up on some reading on my last flight and I couldn't find an option to reflow the text.
https://en.wikipedia.org/wiki/AIXI
http://people.idsia.ch/~juergen/goedelmachine.html
https://en.wikipedia.org/wiki/G%C3%B6del_machine