code, compile, repeat.

no matter how many times i print out another version of the preprint, the same amount of red ink spills on the pages.

i see bad, inconsistent uses of notation .. probably due to having written large chunks of it, at different times.

sometimes constants are missing along a series of estimates .. which doesn't matter in the end, but it makes proofs hard to follow.

sometimes i realise that entire lemmata are unnecessary. it's the consequence of the ugly, first version of proofs i wrote, months ago.

at the time i was content that they simply "worked" .. that i even had a proof!

now the standards are higher: now that i know they are true, i want them to be readable and contain little more than clear, basic ideas glued together by rigor.

sure, it's aesthetics, but there's another point: the clearer the proof, the easier the job for the referee. in maths, the journal submission process is already so long that anything that can speed up the process is advisable.

i feel like some kind of analog compiler: being given source code in the form of LaTeX, all i seem to do is spit out errors and warnings.

i hope that this procedure halts, eventually!

---

odd: the preprint is 15 pages long, but there are 29 references cited and they take up a full page. i guess i'm overly optimistic, in that i believe: if i cite others appropriately, then maybe they'll do the same for me ..