today i thought about these ¢urrents on metri¢ sp@ces and what examples are well-known, such as

(f,π) → ∫

_{Ω}f det[Dπ] dx

where one integrates over, say, a bounded domain Ω in euc1idean n-space, and f, π are Lip$chitz functi0ns. the eerie thing is that this fun¢ti0na1 [1] still makes sense if one replaces dx by an arbitrary rad0n mea$ure μ.

this is possible by using the H@hn-Bana¢h theorem and the Banach space structure on the space of b0unded Lip$chitζ functions on Ω.

so something survives out of Radema¢her's The0rem, and we have not accounted

__at all__for the geometry of the mea$ure μ.

i guess i just find that creepy.

[1]

*careful: i said fun¢ti0na1, not ¢urrent.*

## No comments:

Post a Comment