Genre browse stays route-reachable (deprioritized, not retired).
Session/Mix single-track is a hard upload constraint.
/albums redirects to /cuts when CUTS lands.
Four-wave plan for ReleaseMedium discriminator (Cut/Session/Mix),
medium-specific metadata tables, CMS Release Archive tab, and public
ARCHIVE nav + CUTS/SESSIONS/MIXES browse + detail surfaces.
Resolve Phase 8 open question 0.3 — TrackDto gets a nested Release
(ReleaseDto); flat release fields removed, all consumers updated as
part of §8.0 (flat read-model rejected). Add §0.6 implementation
sequencing: five mergeable waves with Waves 1+2 as a single deployment
unit and Waves 3+4 parallelizable. Update PLAN.md §8.0 Shape to match.
Add §8.0 TrackEntity → Release/Track normalization as a breaking
pre-requisite before Phase 8 UI. Fold in review decisions: Waveform tab
removed (in-grid status column + per-row/page-level generate),
ViewModel is DI-scoped (TracksViewModel pattern), BatchEdit confirmed as
a new page sharing extracted sub-components. Dissolve the AlbumSummaryDto
widening question (Release table supplies the fields directly).
Resolve two open questions in the ParallaxImage spec: TS toolchain
co-located in DeepDrftShared.Client (Interop/parallax -> wwwroot/js), and
parallax direction exposed as the InvertDirection parameter. Update PLAN.md
7.1 constraint to reflect no remaining blockers.
Renumber CMS Enhancements section to Phase 6 (6.1-6.3). Resolve three
6.3 open questions: one album per batch (all release fields shared in
header), persistent track ordinals via new TrackNumber field, and artist
as a release-level header field. Drag-and-drop reorder remains the only
open question.
Track the two EXTENSIBLE WAV sub-cases scoped out of the
WAVE_FORMAT_EXTENSIBLE PCM fix: non-PCM (IEEE Float) SubFormats and
padded 24-in-32 containers.
Record the design-gate decision for HTTP Range support: Range headers in
the JS fetch retaining the AudioBuffer decoder, rejecting MediaElement
(loses early-playback) and synthesized-header-over-Range (breaks caching
invariant). Add per-file shape, acceptance criteria, and the file-absolute
offset constraint. Tighten 4.2 — disk-streaming already done on the
default path; only the legacy offset branch remains.
design/TWO-APP-SPLIT.md proposes splitting the entangled Blazor host into
independent public and CMS apps to escape the layout/cascading-auth deadlock.
Ten open questions at §10. PLAN.md gains an in-flight pointer.
daniel-c-harvey