Saturday, October 06, 2012

budgeting topics for a talk is like packing a weirdly-partitioned backpack with ill-fitting objects.

// initially posted on sat, 6 oct 2012

so i have $C = 16$ pages to write (by hand) for next week.  if you think of $C$ as a quantitative constant, then it depends on the following parameters:
$C = 4C'$, where $4$ is the average rate of pages* per hour i can discuss,
$C'=2C_{\rm talk}$, with $C'$ the number of talks i'm giving next week, and
$C_{\rm talk} = 2^{**}$ is the length of a typical talk in a finnish seminar.
the good thing is that i already have both punchlines in mind, so each talk outline essentially writes itself.  i don't think i can fit in all that i want to say, though, which for me is normal.

in my talks, i always run out of time. /sighs/

the constraints are the usual ones for presentations ..
various components of the talks depend on each other, of course.  if i don't add this one part to the exposition, then later on, another part won't make sense at all ..

meanwhile, the point of the main result is that it has this nice application, as advertised by the title .. but to explain it would probably take a precious 11/2 pages ..!

i hate making choices, deciding what is important.  i always have this fear that i'm making the wrong choice.

// added later, the same day.

when i think about it, these are just talks. i don't know why i'm so wound up about them. i've given dozens of them over the years; sure, there have been some bad ones, but there were plenty of good ones, too.

maybe in having no teaching duties this year (and hence, no recent feeling for what is "reasonable" in a lecture) i've gone to over-thinking a lot of it, dwelling on things that i wouldn't usually dwell on.

not more than a year ago, i wrote twice as many pages per week for the lectures i was giving .. and i rarely worried about any of it.

* = this iaccounts for "simple" pages, which are handwritten analogues of slide #1 or #2 (i.e. basic definitions)
** = 1 lecture hour = 45 minutes

No comments: