One-Normal Telescope Dual Obstructs Perfect Omega-q
Result
The corrected D03 dualizing complex of Entry 366 does not admit a bounded finite-projective compression over the unlocalized Entry-352 structure sheaf. Consequently neither does
The obstruction already occurs on one ordinary short-normal incidence; it does not depend on the exceptional divisor or on Entry 176.
Projective-generator witness
Fix one short normal (u=u_0). In the target let
and in the corrected carrier let
This is the unique degree-zero source point over (a), so (U_y={y}). For the source representable projective (P_y), direct image is the costandard target module
Indeed its stalk at (t) is (A[u^{-1}]) exactly when (t\le a), and is zero otherwise. There is no higher section cohomology because the source open is a singleton.
By adjunction, the stalk of the relative dualizing complex at (y) is
Surviving telescope-dual sector
Resolve (C_a(A[u^{-1}])) by the standard target representables and dualize. After all spectator parameters are made units, filter by whether the terminal target state has inverted (u). The non-inverted associated sector has the common coefficient
Its finite incidence carrier is the target-chain complex whose first point lies below (a) and whose terminal point does not invert (u). The exact enumeration gives chain ranks
boundary ranks modulo (101)
and homology
Its Euler characteristic is (1). Vanishing in the other degrees modulo (101) bounds the corresponding rational homology by zero there; the Euler characteristic therefore forces a free rank-one rational class in degree two. Thus one copy of (D_u) survives the incidence cancellation.
The complementary associated sector is a finite complex of (A[u^{-1}])-modules. Any later filtered differential from that sector has finitely generated image, so it cannot exhaust the non-finitely-generated part of the surviving (D_u) class.
Algebraic non-finiteness
The standard telescope resolution
shows that (D_u) has a nonzero degree-one cohomology module. After localizing at the height-one prime ((u)), it is the familiar completion quotient
which is not finitely generated. Hence the stalk ((\omega_{\widetilde b})_y) is not a perfect (A[u^{-1}])-pulled-back incidence object. A bounded complex of finite projective sheaves would have perfect stalks, giving the required contradiction.
The occurrence costandard factor preserves and reflects perfectness by Entry 363. Therefore
Entry-176 consequence
Entry 176's exceptional cap is a finite rank-one local relative chain map. It cannot be identified with the actual (omega_q): besides the previously proved type mismatch, (omega_q) contains the non-perfect ordinary-normal telescope-dual sector above, whereas a bounded rank-one exceptional cap is perfect. No support functor can turn a perfect rank-one object into this global dualizing complex by an isomorphism.
Entry 176 may still define a local exceptional summand or a map after a normal/nonresonant localization that kills the telescope obstruction. That is a different statement from identification with the unlocalized global (omega_q).
Evidence boundary
research/voevodsky/check_d03_ringed_carrier_typing.rs verifies the unique
source witness, enumerates the entire associated completion-sector chain
complex, checks its modular ranks and free Euler class, and retains the full
dualizing-term census. The non-finiteness of the localization dual follows
from the displayed telescope resolution. No exceptional geometry,
orientation choice, or fitted coefficient enters the obstruction.