Building a Reliable X68000 Game Disk Image Catalog
The Sharp X68000 deserves a game archive that is as carefully documented as its hardware. Thousands of titles were released across floppy disks, hard disks, CD-ROMs, magazine supplements, dealer promotions, and user-made compilations. Many survive only as scattered disk images, private collections, screenshots, or brief references in Japanese magazines. A useful catalog must bring those fragments together without treating every filename as an authoritative record.
Creating a Disk Image Catalog of All Known X68000 Games is therefore a preservation project as much as a database exercise. The objective is not simply to collect files, but to identify releases, distinguish revisions, record provenance, verify image contents, and make the results useful to people using original computers, emulators, and modern storage hardware.
A strong catalog can also expose gaps in current knowledge. It may reveal that a supposedly unique title has several regional releases, that two differently named files are identical dumps, or that a game listed in a magazine never received a confirmed public release. Clear uncertainty is valuable information when it is recorded consistently.
Define What Belongs in the Catalog
The first decision is the scope of the collection. A practical catalog can include commercial games, freeware, shareware, doujin software, conversions, educational titles with game-like content, and major game-related utilities. It should state these categories clearly rather than quietly mixing them together. A “known games” database becomes difficult to search when demos, benchmark disks, operating system utilities, and full releases are all treated as equivalent entries.
Each catalog record should represent a recognizable release or software identity, while each disk image should be recorded as an individual preservation object. This distinction matters when one game has separate boot disks, an installation disk, a soundtrack disk, a later bug-fixed revision, or a compilation appearance. The title record describes the game; the image record describes the specific dump.
Names require special care because X68000 software is often listed in Japanese, romanized Japanese, translated English, and abbreviated filename form. Preserve the original Japanese title whenever possible, then add a consistent romanization and searchable alternate titles. Do not replace the original with an unofficial translation. Publisher, developer, release year, genre, supported models, and media format should each have separate fields.
A catalog also needs explicit status values. “Confirmed release,” “advertised,” “unreleased,” “lost,” “image found,” and “image verified” describe different situations. A magazine advertisement can prove that a title existed, but it does not prove that a particular disk image is genuine. Separating bibliographic evidence from digital evidence prevents assumptions from becoming permanent database errors.
Gather Evidence From More Than Filenames
The most visible source is usually an existing archive of floppy disk images. It is useful, but it should be treated as a lead rather than a final authority. Filenames may be truncated, renamed by uploaders, converted between formats, or copied from another archive without preserving the original disk label. Download descriptions can also repeat mistakes for years.
Japanese computer magazines, software catalogs, retailer advertisements, game manuals, and contemporary review pages provide valuable cross-checks. Scans can establish alternate spellings, release windows, publisher identities, disk counts, prices, and compatibility notes. A magazine index may identify a title that has no surviving image, while a manual scan may clarify that a disk is an expansion or scenario rather than a standalone game.
Community reports are equally important when handled with provenance. A collector may know that a disk boots only on a particular system model, that a copy contains a cracked executable, or that a commercial release uses unusual copy protection. Record who supplied the information, when it was received, and whether it was independently confirmed. A short source note is much more useful than an unexplained confidence score.
Use stable identifiers that survive filename changes. A possible identifier can combine normalized title, publisher, release year, media type, and release sequence, while cryptographic hashes identify the actual files. Keep the original filename and archive path as historical metadata, but do not make either one the primary identity. Two files with different names may have the same content, and two files with the same name may represent entirely different disks.
Capture Disks Without Losing Their Character
Creating a trustworthy image begins with the physical disk and the hardware used to read it. Record the disk label, visible condition, write-protect state, jacket markings, and any handwritten notes before attempting a dump. Photographing both sides can preserve information that will disappear if a label peels away or a disk is later discarded.
A raw or preservation-oriented image should be retained alongside any emulator-ready conversion. Convenient formats are valuable for ordinary use, but they may omit weak sectors, unusual track layouts, intentionally missing sectors, or copy-protection patterns. Keep the original capture unchanged, then generate derivative formats for emulators and real hardware. Never overwrite the source image during repair or conversion.
The drive, controller, and capture software should be part of the record. Note the drive model, alignment adjustments, interface, operating system, imaging tool, read retries, and whether the disk was read successfully in multiple passes. If a disk produces inconsistent results, store the separate captures rather than selecting the most attractive one without explanation.
Modern storage makes it easier to test a large collection on original machines. For example, an enthusiast working with an X68000 can use a CompactFlash IDE adapter to keep a working catalog of hard-disk installations and diagnostic tools alongside floppy-based testing. This does not replace authentic floppy capture, but it creates a practical environment for checking installers, save systems, and multi-disk workflows.
Copy protection deserves a dedicated field. A disk that boots in an emulator may still fail on original hardware if the image format does not represent its protection method accurately. Record whether the software was tested on a real X68000, which model and ROM version were used, and whether a patched or cracked executable was involved. Preserved originals and compatibility patches should be linked, not silently merged.
Build Metadata That Explains Each Release
A catalog becomes useful when its metadata answers practical questions. At minimum, record the title, original script, romanization, alternate names, developer, publisher, release date or year, genre, language, media type, disk count, required memory, supported machine series, and known peripherals. Add relationships for sequels, expansions, remakes, ports, compilations, and magazine distributions.
Image-specific fields should include filename, file format, byte size, hash values, dump source, dump date, physical disk identifier, read quality, boot result, and verification status. If the image is a conversion from another format, preserve the original format and conversion tool. A derivative made for emulation should never obscure the preservation image from which it came.
A useful status vocabulary might distinguish “cataloged,” “source reported,” “dump available,” “hash verified,” “content inspected,” “real hardware tested,” and “needs review.” These values can be combined rather than forced into a single vague label. A release may have a verified disk image but no successful real-hardware boot, or it may be confirmed by physical evidence while no image is currently available.
The following fields provide a compact foundation for a spreadsheet, relational database, or static website. They are deliberately separated so that new evidence can be added without rewriting established records.
| Field | Purpose | Example value |
|---|---|---|
| Canonical title | Primary searchable name | Gradius |
| Original title | Historical Japanese presentation | グラディウス |
| Alternate titles | Romanization, translations, abbreviations | Gradius X68K |
| Publisher and developer | Credits and edition matching | Konami |
| Release identity | Version, region, or compilation relationship | Retail release, revision 1 |
| Media | Physical distribution format | 5.25-inch floppy, 2 disks |
| Image filename | Original supplied name | GRADIUS_D1.XDF |
| Image format | Technical representation | DIM |
| SHA-256 | Exact file identity | Recorded in manifest |
| Source and provenance | Where and how it was obtained | Private disk dump |
| Verification | Current confidence level | Real hardware tested |
| Notes | Protection, bugs, scans, or uncertainty | Requires 2 MB RAM |
Hashes should be generated for every preserved and derivative file, with the algorithm named clearly. SHA-256 is a sensible baseline, while legacy checksums can be retained for compatibility with existing archives. A manifest file makes bulk checking easy and helps detect silent corruption after migration to new storage.
Resolve Duplicates and Version Conflicts
Duplicate detection should begin with exact hashes, but it cannot end there. Two disk images may differ in metadata blocks while containing the same software, or they may share most sectors while representing different revisions. Compare disk geometry, boot sectors, directory structures, executable timestamps, version strings, and visible title screens before declaring two images identical.
Keep multiple legitimate dumps when they document different physical sources. A clean image from a retail disk and a slightly damaged image from another collector may have the same bytes, but their provenance is different. If two captures are byte-identical, link them as equivalent objects while preserving their source records. If one image has unreadable sectors that another fills correctly, retain both and explain the relationship.
Version conflicts are common in games distributed through magazines or later compilations. A title may have altered copy protection, corrected text, a different save-game method, or a shortened loading sequence. Do not collapse these variations into a single “best” image. Assign release or revision labels conservatively, using evidence from manuals, executable strings, disk labels, and contemporary documentation.
Uncertainty should be visible rather than hidden. A record can say “probably retail release,” “title identified from directory name,” or “image appears complete but physical source unknown.” Each claim should have a source note. This approach makes later corrections straightforward and protects the catalog from presenting guesses as historical facts.
Preserve the Collection and Publish Responsibly
A disk image catalog should be reproducible. Store metadata in a plain, documented format such as CSV, JSON, or SQL, and keep field definitions under version control. A static web interface can make the records accessible, while downloadable manifests allow researchers to verify that their local files match the catalog.
Use a three-part storage strategy: an untouched master archive, a working area for inspection and conversion, and a public distribution area containing approved derivatives. Give each file a stable path and avoid renaming it without recording the previous name. Automated scripts can calculate hashes, validate required fields, identify duplicate files, and generate pages from the source metadata.
Legal and ethical boundaries should be part of the project design. A catalog can document that an image exists without distributing the image itself. Where rights holders or community rules require restricted access, publish metadata, screenshots, manuals, hashes, and provenance while limiting the disk data. Credit dumpers, scan contributors, translators, testers, and archive maintainers. Preservation is stronger when contributors can see exactly how their work is represented.
Testing should continue after publication. Invite reports about broken links, inaccurate title mappings, incorrect machine requirements, and failed real-hardware boots, but retain an edit history so changes remain accountable. Every correction should identify what changed and why. Over time, the catalog becomes an evolving research reference rather than a static download directory.
Make the Workflow Repeatable
A small set of operating rules will keep the project manageable as the collection grows:
- Assign a unique release record before adding an image, even when the title is not fully identified.
- Preserve original filenames, source notes, physical labels, and untouched master files.
- Generate SHA-256 hashes and maintain a machine-readable manifest for every image.
- Separate commercial releases, freeware, demos, utilities, compilations, and uncertain entries with explicit categories.
- Mark real-hardware testing, emulator testing, content inspection, and source verification as separate statuses.
Process new material in batches rather than editing records casually. First ingest the files into quarantine storage, then calculate hashes, inspect directory contents, compare against known records, and attach provenance. Only after review should an image move into the public collection. This simple separation reduces accidental overwrites and makes mistakes easier to trace.
A regular backup schedule is essential because disk images are large but metadata is irreplaceable. Keep at least one offline copy of the master archive and test restores periodically. When storage media or file formats change, migrate by copying and verifying hashes rather than by manually dragging files between folders.
The best public interface should support both casual browsing and technical research. Let visitors search Japanese and romanized titles, filter by media and verification status, inspect related releases, and download manifests even when an image is unavailable. A catalog that explains what is missing can be more useful than one that merely displays a long list of filenames.
A carefully maintained X68000 software index can connect collectors, preservationists, emulator developers, and owners of original machines. Begin with the games and disks that can be documented confidently, publish their provenance and hashes, and expand through measured contributions. Each verified record makes the platform’s surviving history easier to study, test, and keep running.
Nereid-X Expansion Board
A personally-produced LAN+USB+Memory expansion board for Sharp X68000 series computers. Multiple production runs were offered, including a final batch and a later revival reproduction run.
Power Supply Repair
X68 power supply repair and modification services were offered by the site owner, with documentation shared through diary entries spanning 2001–2006.
Server & Networking
Notes on FreeBSD administration, ISP changes, server migration, and networking topics. The site itself ran on FreeBSD with the hns diary system and Namazu search integration.
Get in touch
X68K.NET connects Sharp X68000 enthusiasts through community links and shared projects. Reach out with questions about the Nereid project or X68 resources.