diff 56bc728..f6bf760 bounded_scalar_ricci_flow_v2.txt
diff --git a/bounded_scalar_ricci_flow_v2.txt b/bounded_scalar_ricci_flow_v2.txt
index 0151be3..81dec4c 100644
--- a/bounded_scalar_ricci_flow_v2.txt
+++ b/bounded_scalar_ricci_flow_v2.txt
@@ -108,12 +108,12 @@ bound is weaker because after rescaling a would-be singularity the scalar
 curvature can tend to zero while full curvature stays bounded away from zero.
 The limit is then Ricci-flat without being flat, so a scalar bound permits a
 whole tree of collapsing Ricci-flat bubbles.
-src{pdf: src-sesum2005.pdf (Curvature tensor under the Ricci flow)}
+src{pdf: ./pdf/src-sesum2005.pdf (Curvature tensor under the Ricci flow)}
 
 Bamler--Zhang proved the heat-kernel and curvature estimates under scalar
 bounds, including the four-dimensional energy and compactness theory used
-below, and obtained smooth extension when H_2(M;mathbb F)=0bbF)=0 for all
-mathbb F,bbF, Corollary~1.10 of the first part. Simon continued through the
+below, and obtained smooth extension when H_2(M;F)=0 for all
+F, Corollary~1.10 of the first part. Simon continued through the
 limiting space by an orbifold flow; Bamler's convergence theory gives smooth
 convergence off a codimension-four set; Liu--Simon treat four-dimensional
 convergence under scalar-integrability hypotheses. All allow
@@ -122,12 +122,12 @@ keeps the given flow on its original manifold. Blow-up is known for
 K"ahler--Ricci flows (Zhang), for global Type~I singularities (Enders,
 M"uller and Topping), and for local Type~I points (Buzano and Di Matteo);
 no rate is imposed.
-src{pdf: src-bamlerzhang2017.pdf (Bamler and Zhang, Heat kernel and curvature bounds in Ricci flows with bounded scalar curvature)}
-src{pdf: src-bamlerzhang2019.pdf (the same authors, part II)}
-src{pdf: src-simon2020.pdf (Simon, Extending four dimensional Ricci flows with bounded scalar curvature)}
-src{pdf: src-bamler2018.pdf (Bamler, Convergence of Ricci flows with bounded scalar curvature)}
-src{pdf: src-zhang2010.pdf (Zhang, Scalar curvature behavior for finite-time singularity of K"ahler--Ricci flow)}
-src{pdf: src-endersmullertopping2011.pdf (Enders, M"uller and Topping, On type-I singularities in Ricci flow)}
+src{pdf: ./pdf/src-bamlerzhang2017.pdf (Bamler and Zhang, Heat kernel and curvature bounds in Ricci flows with bounded scalar curvature)}
+src{pdf: ./pdf/src-bamlerzhang2019.pdf (the same authors, part II)}
+src{pdf: ./pdf/src-simon2020.pdf (Simon, Extending four dimensional Ricci flows with bounded scalar curvature)}
+src{pdf: ./pdf/src-bamler2018.pdf (Bamler, Convergence of Ricci flows with bounded scalar curvature)}
+src{pdf: ./pdf/src-zhang2010.pdf (Zhang, Scalar curvature behavior for finite-time singularity of K"ahler--Ricci flow)}
+src{pdf: ./pdf/src-endersmullertopping2011.pdf (Enders, M"uller and Topping, On type-I singularities in Ricci flow)}
 
 The spatial picture descends from Einstein four-manifold theory: Anderson, Bando (with a correction), Bando--Kasue--Nakajima, and
 Cheeger--Tian. The slices below are not Einstein, so the needed
@@ -155,7 +155,7 @@ E=o(M) on each fixed
 limiting tree (M the weighted, dimensionless Ricci error). The static input is
 recorded below; the argument must verify the hypotheses for flow metrics and compare
 the functional variation with spacetime Ricci energy.
-src{pdf: src-deruelleozuch2025.pdf (Deruelle and Ozuch, A Lojasiewicz inequality for ALE metrics)}
+src{pdf: ./pdf/src-deruelleozuch2025.pdf (Deruelle and Ozuch, A Lojasiewicz inequality for ALE metrics)}
 
 Two estimates drive the argument. On a nearly flat annulus the contracted
 Bianchi identity removes the critical radial harmonic mode of Ric, improving
@@ -167,7 +167,7 @@ differentiates in time on patches without a smoothly varying tree. This rests
 on the Saint-Venant relation of the Calabi complex (Khavkine) and on Korn
 control modulo rigid motions (Ciarlet and Ciarlet), both used in the collar
 construction below.
-src{pdf: src-khavkine2017.pdf (Khavkine, The Calabi complex and Killing sheaf cohomology)}
+src{pdf: ./pdf/src-khavkine2017.pdf (Khavkine, The Calabi complex and Killing sheaf cohomology)}
 
 parmedskipnoindentProof structure and conventions.
 
@@ -200,7 +200,7 @@ needed.
 
 parmedskipnoindentGeometric data and the functional.
 
-Let mathcal TcalT be a finite rooted tree of four-dimensional Ricci-flat spaces
+Let T be a finite rooted tree of four-dimensional Ricci-flat spaces
 (V_v,h_v). A vertex has one exterior end and one punctured end per child.
 Every end carries Cartesian coordinates on a Euclidean annulus modulo a finite
 subgroup of O(4) acting freely on S^3; the group on a joined end is
@@ -344,9 +344,9 @@ Squaring and optimizing in r^2le cs gives spatial persistence
  d_t(x,y)ge c'd_s(x,y)quadif d_s(x,y)le c''l_0.
 
 
-For the effect on the curvature radius fix a coordinate energy e(t,cdot) for
+For the effect on the curvature radius fix a coordinate energy e(t,·) for
 the covariant components of Rm in a fixed initial normal chart, with
-e(0,cdot)=sum_{jle12}|nabla^jRm|^2. For a small fixed beta>0 the
+e(0,·)=sum_{jle12}|nabla^jRm|^2. For a small fixed beta>0 the
 smoothing estimate sup_{B(x,r/2)}e(t)le C_zeta r^{-6}int_{B(x,r)}e(0)
 together with volume and derivative bounds gives e(t)ge|Rm(t)|^2 and
 
@@ -365,7 +365,7 @@ Chern--Gauss--Bonnet int|Rm|^2le C+4int|Ric|^2le C+CF
 orientation cover if needed) and Gronwall bound F; this is (geom:volume-energy).
 
 Bad set and fixed energy. If l(y)ge r and r^2|Rm|(y)geε, then the derivative bound gives |Rm|geε/(2r^2)
-on B(y,cr) and int_{B(y,c'r)}|Rm|^2ge cε^2r^4; volume comparability upgrades this to a definite Ricci energy e_0>0 on a smaller ball. Now fix small rho and suppose {l<rho} needs more than M radius-Crho balls. Take a maximal family of disjoint unit balls centered at points x_jin{Crhole l<rho} chosen by the primitive xmapsto l(x)psi(rho/l(x)); the derivative bound on this primitive gives disjoint definite-energy balls, and the slice energy bound yields Mle Ce_0^{-1}. Hence {l<rho} is covered by boundedly many radius-Crho balls, with the bound independent of rho.
+on B(y,cr) and int_{B(y,c'r)}|Rm|^2ge cε^2r^4; volume comparability upgrades this to a definite Ricci energy e_0>0 on a smaller ball. Now fix small rho and suppose {l<rho} needs more than M radius-Crho balls. Take a maximal family of disjoint unit balls centered at points x_jin{Crhole l<rho} chosen by the primitive x|-> l(x)psi(rho/l(x)); the derivative bound on this primitive gives disjoint definite-energy balls, and the slice energy bound yields Mle Ce_0^{-1}. Hence {l<rho} is covered by boundedly many radius-Crho balls, with the bound independent of rho.
 
 Convergence of distances and volume comparison. For |t-s|le cr^2,
 cover the bad set at the first time by bounded radius-Cr balls, replace visits
@@ -445,7 +445,7 @@ forcing vanish on fixed annuli; interior elliptic estimates yield a nonzero
 limiting Cartesian tensor H, harmonic and satisfying the uncontracted Bianchi
 identity on the flat punctured cover. Spherical harmonics (radial powers j,
 -j-2) and mu<1 leave only H=r^{-2}C for constant C; Bianchi then reads
-x_iC_{jkab}+x_jC_{kiab}+x_kC_{ijab}=0 for all x, forcing C_{cdotcdot ab}=0,
+x_iC_{jkab}+x_jC_{kiab}+x_kC_{ijab}=0 for all x, forcing C_{·· ab}=0,
 contradiction. Derivatives follow by interior elliptic estimates; slowly varying
 boundary coefficients are absorbed by decreasing mu.
 
@@ -497,7 +497,7 @@ persistence and a finite cover, giving fixed charts, bounded derivatives, and
 smooth terminal convergence (local injectivity gives one sheet). If the
 exceptional set is empty, finitely many such charts bound curvature globally and
 the DeTurck equation continues the flow past T (principal part
-g^{ij}∂_i∂_jg after adding mathcal L_Wg);calL_Wg); so only a genuine
+g^{ij}∂_i∂_jg after adding L_Wg); so only a genuine
 concentration point needs work.
 
 Call two nonflat small-scale limits equivalent if their scales have bounded
@@ -520,9 +520,9 @@ parmedskipnoindentThe outermost scale.
 Fix a small c_*>0 and a bounded continuous psige0 with psi=0 on
 [0,c_*], psi>0 on [2c_*,1], and set
 
- a(t)=sup_{yinmathcal U}calUl(y,t),psibig(l(y,t)^2|Rm|(y,t)big).
+ a(t)=sup_{yinU},t),psibig(l(y,t)^2|Rm|(y,t)big).
 
-After shrinking mathcal UcalU the supremum is attained in the interior at every
+After shrinking U the supremum is attained in the interior at every
 late time; a is positive, continuous, and a(t)to0. For every sequence of
 late times, a(t) and any maximizer represent the outermost nonflat entry, and
 every other nonflat entry has scale at most C a.
@@ -698,9 +698,9 @@ Write P for the componentwise spherical average on a lifted quotient annulus
 
 [Radial inequality for the oscillation]
 Suppose PZ=Z, Z compactly supported in time, E=(∂_tau-Delta_delta)Z.
-For 1<ple2, with z(s)=|,|Z(s,cdot,tau)|_{L^2(S^3)}|_{L^p_tau},
+For 1<ple2, with z(s)=|,|Z(s,·,tau)|_{L^2(S^3)}|_{L^p_tau},
 
- -z''-tfrac3sz'+(3)/(s^2)zle|,|E(s,cdot,tau)|_{L^2(S^3)}|_{L^p_tau}.
+ -z''-tfrac3sz'+(3)/(s^2)zle|,|E(s,·,tau)|_{L^2(S^3)}|_{L^p_tau}.
 
 For p=2 this holds for a Hilbert direct sum over charts/windows with the same
 radial variable. Homogeneous comparison powers are s and s^{-3}.
@@ -713,7 +713,7 @@ in time (boundary term vanishes), and use
 int w^{p-1}w_{ss}le z^{p-1}z'' from differentiating
 int w^{p-1}w_s=z^{p-1}z' and Cauchy. H"older bounds the source, division
 proves the inequality; zeros are handled by regularization.
-mathcal L(s^calL(s^beta)=(3-beta(beta+2))s^{beta-2} has zero exponents 1,-3.
+L(s^-beta(beta+2))s^{beta-2} has zero exponents 1,-3.
 
 
 [Control of the mean derivatives]
@@ -729,7 +729,7 @@ Proof.
 The trace-reversal S=T-tfrac12deltatr_delta T=overlineoverline S(s)+PS;S(s)+PS; the
 divergence constraint and flat divergence differ by a term bounded by
 esup|T|_{k+1,s}, and radial differentiation of
-overlineoverline S'_S'_{ij}n_i=mathcal E_j-∂_i(PS)_calE_j-∂_i(PS)_{ij} along coordinate vectors
+overlineoverline S'_S'_{ij}n_i=E_j-∂_i(PS)_calE_j-∂_i(PS)_{ij} along coordinate vectors
 bounds overlineoverline SS and its radial derivatives, hence the Cartesian derivatives.
 Conversion to covariant derivatives adds only metric-error terms.
 
@@ -737,9 +737,9 @@ Conversion to covariant derivatives adds only metric-error terms.
 parmedskipnoindentThe exterior estimate.
 
 For dyadic radii based on a let
-mathcal A_t(r)=calA_t(r)={r/2le d_t(x_t,y)le2r} and
-A(r)=|sup_{mathcal A_t(r)}calA_t(r)|G||_{L^2(I)},
-V(r)=|sup_{mathcal A_t(r)}calA_t(r)sum_{j=1}^mr^j|nabla^jG||_{L^2(I)}, m a fixed
+A_t(r)={r/2le d_t(x_t,y)le2r} and
+A(r)=|sup_{A_t(r)}|G||_{L^2(I)},
+V(r)=|sup_{A_t(r)}{j=1}^mr^j|nabla^jG||_{L^2(I)}, m a fixed
 high order.
 
 [Exterior concentration bound]
@@ -789,7 +789,7 @@ window, so that
 
 using |f|_2le C(K+o_i(1)) from the energy ball and the shell decay
 (eq:exterior-a), and the weak-L^1 bound plus extrapolation
-|mathcal MvcalMv|_q^qle C_qint v^q (v=f^p, q=2/p).
+|Mv|_q^qle C_qint v^q (v=f^p, q=2/p).
 
 [Interior estimate with a neck weight]
 Extract a tree at times with eta(t)ne0. On its true compact and half-neck
@@ -827,22 +827,22 @@ the extension be differentiated in time without choosing a smoothly varying
 tree; the far exterior enters the error estimate only.
 
 All lengths are in collar units; C_1={1le|x|le2}/Gamma,
-mathcal E=calE={|x|ge1}/Gamma, Gammasubset O(4) finite acting freely on
+E={|x|ge1}/Gamma, Gammasubset O(4) finite acting freely on
 S^3. For a fixed-ratio shell use
 |v|_{H^s(r)}^2=sum_{j=0}^sr^{2j-4}int|∂^jv|^2 and
 |v|_{C^s,r}=sum_{j=0}^sr^jsup|∂^jv|, and
-|v|_{H^s_beta(mathcal E)}calE)=sup_r r^beta|v|_{H^s(r)} (filling a ball on
+|v|_{H^s_beta(E)} r^beta|v|_{H^s(r)} (filling a ball on
 filled Euclidean space). Shells range over finitely many neighbors, with width
 chosen once independent of the outer radius R.
 
 [Collar extension]
 There are sigma_*>0, R_*<∞, C<∞ such that for sigmalesigma_*,
-Rge R_* there is a metric barbar hh on mathcal EcalE with
+Rge R_* there is a metric barbar hh on E with
 
  |barbar h-h-delta|_{H^k_q}le C|g-delta|_{H^k(C_1)},
  |g-barbar hh|_{H^k(C_1)}+|Ric(barbar h)h)|_{H^{k-2}_{q+2}}le Cm_1.
 
-The map g|_{C_1}mapstobarbar h-h-delta is real analytic on a fixed neighborhood
+The map g|_{C_1}|->barbar h-h-delta is real analytic on a fixed neighborhood
 of zero in collar H^k, with uniformly bounded first derivative into H^k_q,
 independent of R, the weights, and the exterior. There is also an analytic
 extension hathat hh equal to g near the inner boundary with the same bounds.
@@ -850,24 +850,24 @@ extension hathat hh equal to g near the inner boundary with the same bounds.
 
 parmedskipnoindentLocal linear estimates.
 
-With Sv,mathcal K(v)calK(v) the flat linearizations of Ricci and Riemann curvature
+With Sv,K(v) the flat linearizations of Ricci and Riemann curvature
 and Bv=∂^iv_{ij}-tfrac12∂_jtr_delta v,
 
- Sv=-tfrac12Delta v+tfrac12mathcal L_calL_{(Bv)^sharp}delta, BSv=0,
- Deltamathcal K(v)=calK(v)=mathcal T(∂^2Sv),calT(∂^2Sv),
- Smathcal L_VcalL_Vdelta=mathcal K(calK(mathcal L_VcalL_Vdelta)=0.
+ Sv=-tfrac12Delta v+tfrac12L_{(Bv)^sharp}delta, BSv=0,
+ DeltaK(v)= T(∂^2Sv),calT(∂^2Sv),
+ SL_V K(calK(L_V
 
 
 [Local potential and Korn]
 On a fixed Euclidean annulus/ball/overlap there is bounded linear
-vmapsto Pvin H^{k+1} with
+v|-> Pvin H^{k+1} with
 
- |v-mathcal L_calL_{Pv}delta|_{H^k}le C|mathcal K(v)calK(v)|_{H^{k-2}},
+ |v-L_{Pv}delta|_{H^k}le C|K(v)|_{H^{k-2}},
  |Pv|_{H^{k+1}}le C|v|_{H^k},
 
 and for vector fields,
-inf_{Winmathfrak e(4)}frake(4)|V-W|_{H^{k+1}}le C|mathcal L_VcalL_Vdelta|_{H^k},
-mathfrak e(4)=frake(4)={xmapsto Ax+b:A^t=-A}. Choices are linear, normalized modulo
+inf_{Winmathfrak e(4)}frake(4)|V-W|_{H^{k+1}}le C|L_V|_{H^k},
+mathfrak e(4)=frake(4)={x|-> Ax+b:A^t=-A}. Choices are linear, normalized modulo
 rigid motions, and equivariant; after dilation the first right side is multiplied
 by r^2 and vector norms divided by r.
 
@@ -877,8 +877,8 @@ Regard the linearized connection C^a_{ij}=tfrac12(∂_ip_{ja}+∂_jp_{ia}-∂_ap
 as a one-form in i (fixed a,j) with curl K(v); the Neumann gradient
 projection writes it as dV^a_j plus an H^{k-1} error controlled by K(v);
 symmetry in i,j and a second projection give C(v)-D^2V bounded in H^{k-1}
-by C|mathcal K(v)calK(v)|_{H^{k-2}}; then ∂_i(v_{aj}-(mathcal L_VcalL_Vdelta)_{aj})=C^a_{ij}+C^j_{ia}
-controls D(v-mathcal L_VcalL_Vdelta), and Poincar'e plus an affine field of half a
+by C|K(v)|_{H^{k-2}}; then ∂_i(v_{aj}-(L_V{aj})=C^a_{ij}+C^j_{ia}
+controls D(v-L_V, and Poincar'e plus an affine field of half a
 constant symmetric matrix completes the estimate. For Korn,
 2∂_i∂_jV_a=∂_ie_{ja}+∂_je_{ia}-∂_ae_{ij} with
 e=S V controls second and higher derivatives; Poincar'e controls V modulo an
@@ -912,18 +912,18 @@ maximum principle and local estimate give the shell-uniform bound.
 
 parmedskipnoindentA filled-space inverse and the analytic map.
 
-Let mathcal E_0calE_0 contain the exterior and the filled unit ball with one smooth
+Let E_0 contain the exterior and the filled unit ball with one smooth
 metric piece; if R is the outer radius use larger balls as needed. For
-qge0 let |v|_{H^s_q(mathcal E_0)}calE_0)=sup_{rge1}r^q|v|_{H^s(r)}, and set
+qge0 let |v|_{H^s_q(E_0)}{rge1}r^q|v|_{H^s(r)}, and set
 
- mathcal X=calX={vin H^k_q(mathcal E_0):v=calE_0):v=mathcal L_VcalL_Vdelta near the inner
+ X={vin H^k_q(E_0):v=calE_0):v=L_V near the inner
  boundary for some vector V},
- mathcal Y=calY={(v_1,f):Sv=f on C_1, Bf=0}.
+ Y={(v_1,f):Sv=f on C_1, Bf=0}.
 
-The filling identifies mathcal XcalX with the H^k collar modulo rigid motions,
+The filling identifies X with the H^k collar modulo rigid motions,
 and Bf=0 is imposed in the specified function spaces.
 
-Define A:mathcal XcalXtomathcal Y,calY, vmapsto(v|_{C_1},Sv), and the analytic
+Define A:X, v|->(v|_{C_1},Sv), and the analytic
 piecewise map Q: on |x|<1 take an inward harmonic extension f=Sw and its
 Newton potential; on C_1 interpolate between Sp_1 and f with a fixed cutoff;
 outside use the Newton potential; then replace the vector contribution by the
@@ -932,58 +932,58 @@ correction. Q is real analytic and Q^2=Q, and
 
  ALA=A,
 
-where L:mathcal YcalYtomathcal XcalX is the following bounded linear operator. On
+where L:Y XcalX is the following bounded linear operator. On
 the filled unit ball f^#=S(Ev_1), on C_1
 f^#=chi Sv_1+(1-chi)f, outside f^#=f; set
 u=Delta^{-1}(-2f^#) (Newton, Lemma~collar:newton),
-Ld=u|_{mathcal E}calE+mathcal L_calL_{V_d}delta with V_d from
+Ld=u|_{E}+L_{V_d}delta with V_d from
 Lemma~collar:potential applied to v_1-u|_{C_1} and extended linearly; it is
 bounded, with
 
- |v_1-(Ld)|_{C_1}|_{H^k}le C|mathcal K(v_1-u)calK(v_1-u)|_{H^{k-2}(C_1)},
+ |v_1-(Ld)|_{C_1}|_{H^k}le C|K(v_1-u)calK(v_1-u)|_{H^{k-2}(C_1)},
  S(Ld)=Su.
 
 
 For (eq:generalized-inverse) take d=Av, fill v by the same inward
 extension to V on all of R^4. Then f^#=SV, Bf^#=0, so Bu=0 and
-Su=f^#; mathcal K(V-u)calK(V-u) is entire harmonic and decaying, hence
-mathcal K(v-u)=0calK(v-u)=0 on C_1, giving (Ld)|_{C_1}=v and S(Ld)=Sv. Filling the
+Su=f^#; K(V-u)calK(V-u) is entire harmonic and decaying, hence
+K(v-u)=0calK(v-u)=0 on C_1, giving (Ld)|_{C_1}=v and S(Ld)=Sv. Filling the
 hole is essential (a decaying harmonic field on the punctured exterior need not
 vanish).
 
 Set Q=LA; (eq:generalized-inverse) gives Q^2=Q and AQ=A, and a
-linear operator A':imQtomathcal XcalX with A'Q=1 identifies
-imQ with mathcal X,calX, transporting A and the real analytic Q
+linear operator A':imQtoX with A'Q=1 identifies
+imQ with X, transporting A and the real analytic Q
 to A and a real analytic B with AB=1, imB=ker A=imQ.
-A bounded linear mathcal N:calN:mathcal EcalEtomathcal XcalX with
-mathcal Nv=P_V(calNv=P_V(varphi v) for a fixed compactly supported varphiequiv1 on
+A bounded linear N:calN:E XcalX with
+Nv=P_V( v) for a fixed compactly supported phiequiv1 on
 C_1 and the H^k collar projection makes a small closed metric family
-solvable, real analytic in the data. The metric barbar h=(1-Bh=(1-Bmathcal N)calN)delta
+solvable, real analytic in the data. The metric barbar h=(1-Bh=(1-BN)
 depends analytically on the collar data with the stated bounds, and the extension
 is independent of the exterior realizing the collar.
 
 parmedskipnoindentA comparison gauge and the residual.
 
-For tensors on 1le|x|le R there are linear P_0, mathcal Dz=calDz=mathcal L_calL_{P_0z}delta,
+For tensors on 1le|x|le R there are linear P_0, Dz= L_calL_{P_0z}delta,
 zero on C_1, with
 
- |z-mathcal DzcalDz|_{H^k(r)}le CBig(r^2max_{rhoasymp r}|mathcal K(z)calK(z)|_{H^{k-2}(rho)}
+ |z-Dz|_{H^k(r)}le CBig(r^2max_{rhoasymp r}|K(z)|_{H^{k-2}(rho)}
  +mathbf1_bf1_{rle C_0}|z|_{H^k(C_1)}Big),
 
 C_0 a fixed width. On roomy dyadic overlaps Lemma~collar:potential gives
 local v_j; on overlaps v_{j+1}-v_j=A_jx+w_j with A_j skew, Gamma-commuting,
 |w_j| controlled by the local errors; there is no equivariant translation.
 For a radial partition the strain has only neighboring differences, so P_0z has
-length-independent bounds; mathcal D^2=calD^2=mathcal DcalD and mathcal DcalD is bounded
+length-independent bounds; D^2= DcalD and D is bounded
 for the weight n(z)=sup_rtheta_r^{-1}|z|_{H^k(r)}. One may write
 P_0z=J(x)x+U(x), J^t=-J, Jgamma=gamma J, J=U=0 on C_1, with
 sup|J|le Csigma n(z) and |J|, r^{-1}|U| controlled by theta_rn(z).
 
 Using this gauge (a comparison device only) the metric is close to Euclidean,
-with p:=F^*g-barbar hh satisfying mathcal Dp=0,calDp=0, p|_{C_1}=p_1, and
+with p:=F^*g-barbar hh satisfying Dp=0, p|_{C_1}=p_1, and
 [
  Y:=sup_{rle R}r^{-2}|p|_{H^k(r)}le C(e+X),
- X:=sup_{rle R}|mathcal K(p)calK(p)|_{H^{k-2}(r)}.
+ X:=sup_{rle R}|K(p)|_{H^{k-2}(r)}.
 ]
 The coordinate Ricci formula and the gauge give
 Sp=T-Ric(barbar h)+E_2=T+f+E_2h)+E_2=T+f+E_2 with |E_2|_{H^{k-2}(r)}le Ctheta_rY
@@ -994,25 +994,25 @@ factors O(theta_r), last r^2Y), giving the residual bound
  le Cbig(m_1+sigma N_0+R^{-q-2}big).
 
 
-Run L on d_1=(p_1,f), e=|d_1|_{mathcal Y}calY; ALd_1=0. It remains to show
+Run L on d_1=(p_1,f), e=|d_1|_{Y}; ALd_1=0. It remains to show
 d_1-ALd_1 is small. With u=Delta^{-1}(-2f^#),
 |Bu|_{H^{k-1}_{q+1}}le CD_0 (D_0=m_1+sigma e), and
-Su-f^#=tfrac12mathcal L_calL_{(Bu)^sharp}delta gives
+Su-f^#=tfrac12L_{(Bu)^sharp}delta gives
 
- |Su-f|_{H^{k-2}_{q+2}(mathcal E)}calE)le CD_0.
+ |Su-f|_{H^{k-2}_{q+2}(E)} CD_0.
 
-Fill p by Ep_1 to P on the full ball and let W=mathcal K(P-u);calK(P-u); on the
+Fill p by Ep_1 to P on the full ball and let W=K(P-u);calK(P-u); on the
 filled inner ball SP-f^#=0, on the collar bounded by CD_0, so the shell sum
-of Delta W=mathcal T(∂^2S(P-u))calT(∂^2S(P-u)) is le C(m_1+sigma e), with the
+of Delta W=T(∂^2S(P-u))calT(∂^2S(P-u)) is le C(m_1+sigma e), with the
 filled ball contributing one fixed term. At radius R,
 |W|_{H^{k-2}(R)}le C(m_1+eR^{-q-2}), and the filled-ball Green sum with the
 fixed-scale estimate near C_1 gives
 
- |mathcal K(p_1-u)calK(p_1-u)|_{H^{k-2}(C_1)}le Cbig(m_1+(sigma+R^{-q-2})ebig),
+ |K(p_1-u)calK(p_1-u)|_{H^{k-2}(C_1)}le Cbig(m_1+(sigma+R^{-q-2})ebig),
 
 with no inner boundary term proportional to e. Combining with
 (eq:L-residual) and ALd_1=0,
-|d_1-ALd_1|_{mathcal Y}calYle C(m_1+(sigma+R^{-q-2})e); choosing sigma_*
+|d_1-ALd_1|_{Y} C(m_1+(sigma+R^{-q-2})e); choosing sigma_*
 small and R_* large makes the e-coefficient at most one half, hence
 ele Cm_1 with length-independent constants, proving (eq:collar-extension).
 An interpolation hathat h=h=barbar h+h+zeta(g-barbar h)h) with a fixed radial zeta