CURATOR Planning Record
Planned Changes for Next Release
Updated On:
What We're Working On
These are the improvements currently planned for the next CURATOR release. Development is still underway, so the final list may change before release.
CURATOR v2.5.0-beta.1 remains the current public release.
Adding
- Expanded Atari and Sega MAME software-list support includes physical ROM requirements, shared identical ROM filenames, and sets containing both ROMs and disk images. ROMs publish as native ZIPs; declared disks publish as separate CHDs, and every required ROM and CHD must be present for completeness. Original files remain unchanged. Emulation loading instructions remain in the retained DATs and do not rearrange or byte-swap game files. Existing saved catalogues and prepared plans are not automatically rewritten.
- Platform associations cover the 32 added Atari and Sega software lists, including Atari Portfolio, Sega AI Computer, Beena, Car Beena, Yamaha/Sega Copera, Mega-Tech, SF-7000, SC-3000, TeraDrive and ST-V.
- DAT Discovery accepts optional No-Intro schema v4 terms-of-service headers while retaining their exact source metadata. Unknown structures and duplicate scalar fields still fail strict mapping checks.
- Eligible legacy v4 Dreamcast CHDs can recover their disc layout after normal DAT matching finishes. Recovery uses the loaded DATs, stages recovered tracks in the configured temporary folder and transfers them into Workspace. Every recovered track and GDI descriptor must match the DAT by size, CRC32 and SHA-1 before acceptance. Originals and source history are preserved. Unsupported or unverified layouts retain normal review handling; a failed recovery does not stop other images. Testing so far has confirmed that the recovered Dreamcast CHDs tested have worked without any perceived issues. This result applies to the files and emulator setups tested; identification alone does not guarantee compatibility with every emulator. Rebuild Stages 4–9 to apply this handling to an existing prepared plan.
Changing
- Documentation includes five PDF guides: Quick Start for a first successful run, End User for everyday operation, Administrator for administration, Read Me First for consolidated warnings, and the Core API handbook. They cover current No-Intro inputs, DAT completeness and dated layouts, conflict-folder placement, archive progress visibility and authority-source provenance. The operator guides reference Read Me First for shared warnings.
- Organizer's final console summary shows Duplicate copies avoided and Existing outputs reused in a separate Duplicates section below the output-category totals.
- Incomplete Sets identifies incomplete games when the loaded DATs define media as separate entries. Game assessments use the title and recognized Disc, Disk, File, Part, Side and Tape positions within each DAT catalogue; different declared disk totals remain separate layouts. Disk or Disc numbering takes precedence over Part labels. Each DAT-defined set retains its own required members and packaging. Unresolved or conflicting declarations stay in Needs Attention. Rebuild Stages 4–9 before Organizer to apply this handling to an existing prepared plan.
- On platforms classified as computers, crack and pirate tags alone do not make a retail game unofficial. Other applicable classifications, including trainers, bad dumps and language restrictions, still apply. Unresolved platform classifications remain unresolved.
- Destination conflicts are preserved directly under `needs_attention\Conflicts`. Numbered subfolders are added only when another conflict has the same filename. Organizer shows a separate final conflict count; ordinary review-file numbering is excluded.
- Preserved CHDs in Needs Attention retain a platform association when matched tracks within that same CHD establish one unambiguous platform. Review folders show the recorded member mismatch reasons. Matching requirements and the separate Incomplete Sets category remain unchanged; rebuild Stages 4–9 to apply this routing to an existing prepared plan.
- Incomplete Sets is a separate output category from Needs Attention in planning, saved records, Organizer and Reporting. Completion totals show Official, Unofficial, Needs Attention and Incomplete Sets without format breakdowns or internal counters. Existing prepared plans containing the old incomplete-set classification require rebuilding Stages 4–9; retained output is preserved.
- Identical final payloads may share storage within the same output category and destination folder while retaining original sources and DAT set memberships. Incomplete Sets also retains each allocated DAT filename. Detailed logs distinguish duplicate occurrences, copies consolidated and existing outputs reused. Append verifies retained output before reuse. Existing duplicate collections block publication; plans prepared before this change require rebuilding Stages 4–9 before Organizer. A compatible database does not require a reset.
- Platform Preparation and DAT Discovery archive accepted inputs in dated batch ZIPs under `platforms/History` and `DATs/History`. Identical reintroduced contents reuse retained archives without recompression, even when their names or locations change. Empty intake is supported when the required saved catalogue is ready. Interrupted archiving retains recovery evidence; DAT Discovery removes empty intake subfolders after successful archiving while preserving nonempty folders and History.
- New platform and DAT intake batches use Fastest ZIP compression to reduce archive creation time, with a modest increase in archive size. Live progress identifies the batch and current file.
- Gateway displays missing platform and DAT setup requirements before the menu. While setup is incomplete, Restart / Recheck setup is the available action; supported workflows appear after the required inputs or saved catalogues are ready.
- Changes to enabled output-routing and collector-routing policies are included when Platform Preparation determines whether its saved catalogue needs updating.
- Progress displays show current activity and elapsed time during long stages, including database work and deferred disc processing. Archive extraction distinguishes overall and current-batch counts, including resumed work. Organizer identifies the destination ZIP during creation and presents aligned output totals with separate failure and duplicate counts. Detailed activity remains in the logs.
- Scan Preparation retains successfully processed nested archives alongside their extracted contents so Matching can identify archive files required by multi-file sets. Archive depth and safety limits still apply.
- Scan Preparation prepares original and eligible headerless identities for Matching while preserving source bytes and original-match priority. Missing or stale preparation evidence requires resuming Scan Preparation before Matching. Header normalization applies only within the configured whole-file size limits; larger files keep their original identity.
- Supported CDI images defer writing decoded tracks until output needs them. Images with external CD-TEXT retain their original container instead of publishing converted tracks. Successful decoding does not establish DAT identity or playback compatibility.
- Nuclear cleanup includes the contents of `platforms/History` and `DATs/History`. It preserves those folders and intake files outside History. Retained intake archives are therefore removed by this operation; review the cleanup scope before confirming it. Temporary database sharing conflicts are retried up to five times with two seconds between checks. If database files remain in use, cleanup stops before deleting any items.
- DAT, media and region policy files now reside in `config/policies` and must remain with the application. Prepare DAR accepts their rules into the collection database for subsequent collection runs.
- Stage information and Gateway outcomes are retained in the collection database. Stage 1–8 intermediate JSON files and their output templates are retired; integrations that used those files must be updated. Published collection manifests and Reporting exports remain supported.
- Incompatible older database layouts require a complete clean restart: move the previous database and generated runtime state out of active folders, preserve original inputs and retained collection output, then rebuild from the inputs. Compatible current-layout databases receive the supported additive extensions. Database schema remains 26; general migration of older layouts is not supported.
- The guides explain current setup, intake reuse, recovery and output requirements, including category/folder-scoped payload storage, independent DAT set completeness and retained-output verification. They distinguish Stage 5 CHD identity preparation from Stage 6 legacy Dreamcast recovery and describe DAT requirements, configured staging, original preservation and emulator compatibility limits. They explain responsibility for current DAT/XML catalogues, incorrect-declaration risks and conflict review within the corresponding Needs Attention folder structure.
Removing
- The fourteen bundled Atari format-routing and Libretro-reference XML stubs are removed from `platforms` and future packages. Their routing and role policies remain available.
- The obsolete root `templates` folder is removed. Console and documentation text templates remain in `config`.