weekend (research) relapse.

[from friday, 22 may 2009]

in other news, i´ve relapsed.

i had really intended to give up working on this one idea, but i´m too close. if it´s just a matter of resolving whether a particular w€ak-star 1imit is nonzero or not, then ..

.. i just can´t let it go,
not that easily.

at any rate, i´ve thought of a convoluted argument which might do it. then again, it´s likely to be gibberish which, in the frenzy of working, i thought was rigorous, and in it might contain an obvious error.

it wouldn´t be the first time, so i´m setting it aside for a few hours, maybe a day. later i´ll write it up and see where the weaknesses are.

(it´s not easy to account for one´s obsession, you know.)

.
.
.
[earlier this morning, sunday 24 may 2009]

the argument is still convoluted and there's still a crucial part that i haven't written up [1]. i have my doubts. in fact, it's the trickiest part and i have a lot of doubts.

"tricky" doesn't mean wrong .. not yet, anyway. the argument is just, well, weird. i'm partitioning sets in a measurable way [2] but because of some self-similarity in the method, i can't see the limiting geometry very clearly.

usually that's bad news,
and likely it means that i've gotten something wrong ..

.. but i haven't spotted an error yet.


[1] as for why these notes are not fully written up yet: today is my girlfriend's birthday. i wouldn't have gotten as far as i have, were it not for the fact that she also wanted to work on her own research, this morning!

[2] in fact, i'm following the standard tactic of splitting space into a "good" part and a "bad" part, relative to these integra1 estimates i'm setting up. let's hope that the bad part isn't actually that bad. \-: