diff c7a3d7d..254e60f daugavet_l1_exact_moduli_v2.txt
diff --git a/daugavet_l1_exact_moduli_v2.txt b/daugavet_l1_exact_moduli_v2.txt
index 8ad6a03..bec6b36 100644
--- a/daugavet_l1_exact_moduli_v2.txt
+++ b/daugavet_l1_exact_moduli_v2.txt
@@ -23,10 +23,11 @@ norm N is asymptotically uniformly convex (AUC) if
bar-delta_N(t)>0 for every t>0.
The averaged increment H_N and the one-sided increment D_N have
-different convexity characters. Uniform convexity of Clarkson asks that
-both ||x+ty|| and ||x-ty|| exceed one, on average, uniformly over
-|x|=|y|=1. Milman obtained a linear-topology weakening, AUC, around
-three decades later. The averaged version appeared only recently:
+different convexity characters. Uniform convexity of Clarkson is a
+stronger condition on unit vectors x, y : the midpoint (x+y)/2
+stays bounded away from one whenever ||x-y|| stays bounded away
+from zero. Milman obtained a linear-topology weakening, AUC, three
+decades later. The averaged version appeared only recently:
Dilworth, Kutzarova, Randrianarivony, Revalski and Zhivkov introduced
asymptotic midpoint uniform convexity (AMUC) and asked whether AMUC and
AUC are equivalent up to renorming.
@@ -55,7 +56,7 @@ answered the question negatively.
admit an equivalent norm that is asymptotically uniformly convex."
Baudier proved only a linear lower bound for the maximum-based modulus
-bar-delta_{X_KW}(t)>=ct. The aim is to compute both moduli
+m_{X_KW}(t)>=ct. The aim is to compute both moduli
exactly, for every unit center and every positive radius. Baudier's
argument passes through the L^1 splitting identity of Brezis and Lieb
along almost-everywhere convergent subsequences; the lower bound in
@@ -396,7 +397,7 @@ this order,
Such x_0 exists since R=sup_{u in S_E} N(u). The definition of
bar-delta_N(t_0) at x_0/N(x_0) provides F in cof(E) with
- N(x_0+z) >= (1+gamma) N(x_0) whenever z in F and N(z)>=t_0 N(x_0). (*)
+ N(x_0+z) >= (1+gamma) N(x_0) whenever z in F, N(z)>=t_0 N(x_0). (*)
To extend (*) to a larger scale, use the convexity of
r -> N(x_0/N(x_0)+r w) for N(w)=1: for every r>=t_0,
@@ -626,9 +627,9 @@ inclusion G_b contained in G_{b+1} is strict, and E is
infinite-dimensional.
Total boundedness in measure. The union over b of B(G_b) is norm dense
-in B(E): if x_j in the algebraic union tends in norm to x in B(E),
-then x_j/max{1,||x_j||} also tends to x. Repeated use of (c) gives,
-for a>b and g in B(G_a),
+in B(E): a norm-approximant x_j of x in B(E), chosen in the algebraic
+union and normalized to x_j/max{1,||x_j||}, also tends to x. Repeated
+use of (c) gives, for a>b and g in B(G_a),
inf_{psi in B(G_b)} d(g,psi) <= sum_{l=b}^{a-1} h_l <= 2^{1-b}.