NEWS_04.tar is a 96 MB Nintendo NEWS workstation backup that preserves a large amount of graphics-side production material rather than source code. Where NEWS_05 captures the Star Fox 2 3D toolchain, NEWS_04 captures the more traditional 2D side of console production: character banks, palettes, screen layouts, object definitions, maps, and a huge number of backup revisions.

The archive is especially useful because it is not a clean, single-project handoff. It is a live multi-user workstation snapshot with three home directories, several different projects, and visible evidence of iterative art work.

---

At a Glance

NEWS_04 is best understood as a mixed graphics workstation backup. It preserves:

• 5,309 archive entries under three user homes: arimoto, sugiyama, and kakui
• 2,297 .BAK files, showing heavy iteration and local backup habits
• 991 .SCR files, making screen and scene layout one of the dominant data types
• 876 .CGX files and 431 .COL files, pointing to SNES/GB graphics-bank and palette work
• 266 .OBJ files and 108 .MAP files, showing object/layout and map-side asset organization
• One especially important late Star Fox 2 art workspace under home/arimoto/SF2
• Older Zelda and Game Boy Zelda art workspaces under home/arimoto/zelda and home/arimoto/GB-zelda
• A second artist-style workspace under home/sugiyama with fly, flyman, CAR, SIM, MARIO, and FX2

Unlike NEWS_05, this archive contains almost no conventional program source. Its value comes from file naming, layout formats, revision backups, and the way several projects coexist on one machine.

---

Glossary of Key Terms

If you are new to Nintendo workstation graphics formats, this glossary will make the rest of the page much easier to follow.

• CGX - Character graphics or tile graphics data. In these archives, .CGX files look like graphics-bank or sprite/tile resources rather than source code.

• COL - Palette data. These files usually sit beside .CGX and .SCR assets and define the color sets used to display them correctly.

• SCR - Screen layout data. This typically represents how tiles or graphics are arranged into a scene, menu, background, or composed screen.

• OBJ - Object-side asset data. In NEWS_04 this seems to refer to 2D object/sprite-side resources or layout groupings, not the 3D CAD object pipeline seen in NEWS_05.

• OBX - A related object-side format that appears beside .OBJ in some Star Fox 2 folders. It likely represents a companion state or variation format, but the exact structure still needs deeper reverse-engineering.

• MAP - Map or level-layout data. These files appear most strongly in the Zelda-related folders.

• PNL - Panel or tile-layout resource. These often look like intermediate layout assets used with map and screen files.

• BAK - Backup copy. The huge number of .BAK files is one of the strongest clues that this archive preserves active production work rather than a final handoff.

• CBM - A less common asset format present mainly in sugiyama’s workspace. Its exact meaning is unclear here, but it appears among other authored graphics-side files rather than code.

• MD7 - Very likely Mode 7-related data. Its appearance inside CAR is notable because racing and pseudo-3D background work often depended on Mode 7 transformations.

• NEWS workstation - Sony NEWS Unix workstation hardware used in Japanese game development environments. Nintendo preserved several such workstation-side snapshots inside the Gigaleak.

---

What NEWS_04 Actually Is

The top-level structure is simple but revealing. The archive is almost entirely a home backup:

At a high level, the archive breaks down like this:

Workspace	Files	Dominant types	Date range	Why it matters
home/arimoto	about 3,278	.BAK, .CGX, .SCR, .COL, .OBJ, .MAP	1991-05-23 to 1995-09-19	Main late-production art workspace, especially for SF2
home/sugiyama	about 1,900	.SCR, .CGX, .COL, .BAK, .CBM, .OBJ	1989-10-13 to 1994-03-18	Older multi-project graphics workspace with several prototype or pre-SF2 strands
home/kakui	26	Mostly shell/profile files	n/a	Personal workstation setup only

That date spread is important. NEWS_04 is not a single synchronized snapshot from one project phase. It is a personal workstation backup carrying several years of older project residue plus one clearly later Star Fox 2 branch.

---

File-Type Profile

The overall extension spread explains why NEWS_04 feels so different from NEWS_05. It is dominated by authored graphics assets and revision copies rather than code or CAD data.

Extension	Count	What it suggests
.BAK	2297	Heavy manual backup and revision churn
.SCR	991	Screen composition and layout were central tasks on this machine
.CGX	876	Graphics-bank and character/tile art production
.COL	431	Palette pairing was a routine part of the workflow
.OBJ	266	Object-side or sprite-side grouping data
.MAP	108	Map or room layout work, especially in Zelda folders
.PNL	37	Panel/layout intermediates
.CBM	31	Less common graphics-side authored resources
.DAT	19	General data sidecars or tool outputs
.SFX	13	Small sound-related resources
.OBX	12	Object-side companion files, mainly in SF2
.MD7	3	Probable Mode 7-related data in CAR

This tells us two things immediately:

• the archive was used heavily for asset preparation and screen assembly
• the preserved machine sat much closer to the art/layout pipeline than to the source-code or build-tool chain

---

The Main Story: Arimoto’s Workspace

Arimoto’s home directory is the most important part of the archive. It combines one late and unusually dense SF2 workspace with several older Zelda-related branches.

Arimoto at a Glance

Project	Files	Dominant types	Date range	Reading
SF2	1236	.BAK, .CGX, .SCR, .COL, .OBJ, .OBX	1993-07-01 to 1995-09-19	The late, most active branch and the real centerpiece of NEWS_04
GB-zelda	824	.BAK, .OBJ, .CGX, .MAP, .SCR, .PNL	1991-11-27 to 1994-08-02	Game Boy Zelda visual and layout work, with stronger map/object emphasis
zelda	545	.BAK, .CGX, .SCR, .COL, .MAP, .PNL	1991-05-23 to 1994-07-25	Earlier Zelda screen/map art branch
DELDA	213	.BAK, .CGX, .SCR, .COL	1991-05-23 to 1991-10-24	Small early Zelda-related branch or internal variant

The important split is between:

• older Zelda/GB-Zelda work concentrated in 1991-1994
• late Star Fox 2 work concentrated in 1995

That makes Arimoto’s home feel like a long-lived artist workstation where old materials were retained rather than cleaned out between projects.

---

The Late Star Fox 2 Art Branch

Arimoto’s SF2 directory is the strongest and latest branch in the whole archive. Its newest sampled files reach 19 September 1995, which is later than the material we saw concentrated in NEWS_05.

This is also a very different side of Star Fox 2 from the CAD-heavy 3D workflow. Instead of .cad, .anm, and .nca, NEWS_04 keeps 2D graphics banks, palettes, screen layouts, and object-side resources. That suggests the archive captures the presentation and interface side of the project rather than the polygon authoring pipeline.

SF2 Folder Profile

arimoto/SF2 contains about 1236 files:

Extension	Count	Interpretation
.BAK	686	Very heavy iteration with many preserved prior states
.CGX	189	Graphics banks and character/tile art
.SCR	140	Screen/layout assemblies
.COL	132	Palette sets for those graphics/layouts
.OBJ	83	Object-side resources
.OBX	6	Object-side companion variants

The internal subfolders are much more informative once you add date ranges and extension balance:

Subfolder	Files	Date range	Dominant types	Reading
t	325	1993-07-09 to 1995-06-03	.CGX, .COL, .SCR, many .BAK	Large numbered screen/layout batch
s	172	1993-08-27 to 1995-08-31	.COL, .SCR, .CGX	Structured a* scene bank with regular triplets
m	129	1993-10-15 to 1995-08-18	.SCR, .COL, .CGX	Structured b* scene bank
ma	47	1995-07-03 to 1995-08-30	.SCR, .CGX, .COL	Explicit late alternate branch for m
o	164	1993-11-19 to 1995-09-19	.OBJ, .CGX, .OBX, small .SCR/.COL set	Main object/enemy-side branch and latest-edited area
oa	25	1995-07-07 to 1995-09-19	mostly .CGX	Explicit late alternate branch for o
obj	96	1994-02-04 to 1995-06-26	mostly .OBJ	Object library / object-store bucket
watanabe	8	1994-03-16 to 1994-10-24	obj-*, color-*, spt_2	Tiny handoff or shared sample subset
mt	1	1994-11-21	lone .OBX	Residual single-file bucket

That split is important. t, s, and m look like long-running banked layout groups that started in 1993. ma and oa appear much later, only in mid-to-late 1995, which strongly suggests explicit alternate or revised sub-branches created near the end of work.

A Clearer SF2 Taxonomy

The naming prefixes are repetitive enough that they start to form a real internal taxonomy rather than a loose pile of files.

1. t - Numeric layout banks

The t folder is dominated by numbered families such as 0-*, 1-*, 2-*, 6-*, 7-*, 15-*, and 16-*. Typical triplets include:

• 0-2.CGX, 0-2.COL, 0-2.SCR
• 0-6.CGX, 0-6.COL, 0-6.SCR
• 1-1.CGX, 1-1.COL

This looks like a broad numbered scene or bank repository. It is graphics-heavy and layout-heavy, with only .CGX, .COL, .SCR, and backups. So t is best read as a large screen/tile bank rather than an object store.

2. s - a* scene group

The s folder is unusually consistent. Its top prefixes are a0, a1, a7, a10, a11, a14, a15, a16, a18, a27, and a28. Representative file groups include:

• a0.CGX, a0.COL, a0.SCR
• a10.CGX, a10.COL, a10.SCR
• a15.CGX, a15.COL, a15.SCR

This is one of the cleanest sections of the archive. It looks like a scene-set or stage-set bank with a stable triplet workflow of graphics, palette, and composed screen files.

3. m - b* scene group

m behaves very similarly to s, but its naming family is b* rather than a*. Its heaviest prefixes are b7, b1, b8, b9, b10, b14, b15, and b16. Representative file groups include:

• b1.CGX, b1.COL, b1.SCR
• b10.CGX, b10.COL, b10.SCR
• b14.CGX, b14.COL, b14.SCR

This strongly suggests s and m are parallel production buckets inside the same broad graphics system. They may separate different screen families, gameplay contexts, or region/build groupings.

4. ma - late alternate b* branch

ma is much smaller and much later. Its files are concentrated in 1995-07 to 1995-08, and almost everything carries a -a suffix:

• b0-a.CGX, b0-a.COL, b0-a.SCR
• b7-a.CGX, b7-a.COL, b7-a.SCR
• b16-a.CGX, b16-a.SCR

That makes ma look like an alternate or adjusted branch of m, not a separate independent system. The naming is too close to be coincidence.

5. o and obj - object-side branch

o and obj are where the archive becomes more object-heavy. Representative filenames include:

• e0-0.OBJ, e0-1.OBJ, e1-3.OBJ
• d0-1.OBJ, d1-1.OBX, d3-3.OBX
• cm.CGX, w0.CGX, pm.CGX

The dominant prefixes inside o are e3, w0, mm, e0, w2, e4, e1, d3, and d0. Inside obj, the heaviest families are d3, d2, d4, w0, and d0.

That split suggests a two-layer system:

• obj as a denser object library bucket with many .OBJ definitions
• o as a broader working object branch where those objects are paired with graphics, a few palettes, and occasional .OBX companions

The presence of .OBX beside .OBJ suggests paired object-state, alternate composition, or behavior-related companion data. Whatever the exact format, this is clearly different from the pure scene-bank logic of t, s, and m.

6. oa - late alternate object branch

oa looks like an alternate object-side graphics branch. The repeated -a suffixes imply variants or adjusted revisions:

• e0-a.CGX
• e3-a.CGX
• pm-a.CGX
• w0-a.CGX

The date range matches late 1995, and the naming mirrors o too closely to read any other way. oa is best understood as an explicit late alternate graphics branch for object families already present in o.

7. watanabe - tiny handoff subset

The watanabe folder contains only eight files:

• color-date.CGX, color-date.COL, color-date.SCR
• obj-0.CGX, obj-0.COL, obj-1.CGX
• p_col.COL
• spt_2.cgx

This is too small to be a real working branch. It reads more like a shared sample, handoff, or imported subset tied to Watanabe’s side of the Star Fox 2 workflow.

How the Buckets Relate to Each Other

Taken together, the strongest interpretation is:

• t, s, and m are structured screen/layout banks
• ma is a late alternate revision layer for m
• obj is an object-definition store
• o is the active object/enemy working branch
• oa is a late alternate revision layer for o
• watanabe is a tiny shared subset or handoff residue

That is much more specific than simply saying the folder contains “art assets”. It suggests a real internal organization where scene banks and object banks were kept separate, then selectively forked into -a revision branches during late 1995 cleanup or adjustment work.

What the Date Ranges Add

The date spread strengthens that reading:

• t, s, m, and o all begin in 1993
• obj only starts showing up in 1994
• ma and oa only appear in mid-1995, right near the latest visible Star Fox 2 edits
• o and oa carry the latest timestamps, both reaching 19 September 1995

So the most plausible sequence is:

1. long-running scene and object banks are built from 1993 onward
2. an explicit object library (obj) stabilizes during 1994
3. late 1995 creates focused alternate branches (ma, oa) for final adjustments
4. the object-side branch remains active slightly later than the scene-bank side

That last point matters because it hints that late visible work was not broad world-building anymore. It looks more like targeted object, enemy, presentation, and polish changes.

Inferred Workflow Inside SF2

The repeated file groupings imply a local workflow that looks something like this:

flowchart LR
  A["<b>Scene or object family</b><br>a*, b*, d*, e*, w* naming"] --> B["<b>Graphics bank</b><br>CGX"]
  B --> C["<b>Palette pairing</b><br>COL"]
  C --> D["<b>Composed layout</b><br>SCR or object-side OBJ/OBX"]
  D --> E["<b>Backup before change</b><br>BAK copy"]
  E --> F["<b>Late alternate branch</b><br>-a variants in ma / oa"]

That is exactly the kind of detail that a clean source archive would normally erase. NEWS_04 preserves it because the machine was backed up in the middle of active production use.

SF2 Filenames That Stand Out

A few filenames make the branch easier to interpret:

• open-logo.CGX and open-logo-5.CGX - very likely title or opening-logo work
• logo.SCR - direct evidence of composed logo layout
• character-L.CGX, character-La1.CGX - character-bank or portrait/state art
• e9-96.CGX, e9-97.CGX - late numbered revisions in September 1995
• w0.CGX, w2.COL - compact numbered assets tied to object-side folders

The timestamps are the most important part. The densest and newest files cluster in July-September 1995, which makes this one of the latest visible Star Fox 2 graphics-side workspaces in the NEWS tape set.

---

Zelda and Game Boy Zelda Material

Beyond Star Fox 2, Arimoto’s home preserves three distinct Zelda-related workspaces. They are not all the same game or the same moment. Reading them in chronological order tells a story about how a Nintendo workstation accumulated Zelda history over a nearly three-year window.

Three Zelda Projects in One Machine

Project	Files	Date range	Dominant types	Reading
DELDA	213	1991-05-23 → 1991-10-24	BAK, CGX, SCR, COL	Five-month early prototype; likely the first Zelda work on this machine
zelda	545	1991-05-23 → 1994-07-25	BAK, CGX, SCR, COL, MAP	Long-running SNES-side Zelda branch spanning three-plus years
GB-zelda	824	1991-11-27 → 1994-08-02	BAK, OBJ, CGX, MAP, SCR	Game Boy Zelda branch; largest of the three; strong localization evidence

The dates tell the first part of the story: DELDA and zelda both open on the same day — 1991-05-23 — suggesting the workstation was set up or first used for this project in late May 1991. GB-zelda follows six months later in November 1991. DELDA closes in October 1991. zelda and GB-zelda both continue into mid-1994.

---

DELDA: The Earliest Prototype

DELDA is the smallest and shortest-lived of the three Zelda folders. Its five-month window — May to October 1991 — makes it a snapshot of the very first phase of Zelda work on this machine.

Subdir	Files	Date range	Dominant types	Reading
soto	66	1991-05-23 only	BAK, COL, SCR, MAP	Outdoor / field tile work (soto = outside/exterior in Japanese)
m	51	1991-05-23 → 1991-10-24	BAK, CGX	Tile art; runs the full five-month span
d	23	1991-10-21 → 1991-10-23	CGX, BAK	Dungeon-side tiles; concentrated in the final two days of the folder’s life
i	1	1991-05-23 only	BAK	Single residual file

The soto subfolder’s name (外 = outside) points to outdoor/overworld tile sets. The m subfolder runs the whole five months, suggesting it was the active tile art store. The d subfolder has almost all its files dated to October 21-23, which looks like a concentrated push — possibly porting or adapting dungeon tiles right before the project reorganised into the main zelda tree.

DELDA does not contain .OBJ files and has almost no .MAP data. That makes it feel like a pure tile-painting phase: the team was building raw graphics assets before the broader map and object infrastructure was established.

---

zelda: The Long-Running SNES Branch

The zelda folder runs for over three years and is the richest Zelda workspace on the machine.

Subdir	Files	Date range	Dominant types	Reading
d	170	1991-05-23 → 1994-07-25	BAK, CGX, COL, SCR	Dungeon and overworld area tile banks; latest files in the whole zelda tree
khn	72	1991-05-23 → 1991-11-07	BAK, SCR	Room/screen layouts; active in the first six months
m	89	1991-05-23 → 1992-04-28	MAP, CGX, COL, SCR	Map-side data with a special tmp sub-snapshot
spl	10	1991-05-23 only	SCR, BAK	Early sprites and panels (kabe = wall, osr = enemy/contact)
atari	8	1991-06-11 → 1991-08-09	BAK, CGX	Collision data (atari = hit/contact in Japanese game dev — not the company)
i	14	1991-05-23 only	BAK, SCR	Interior room screens (h-1 through h-a2)
w	9	1991-05-23 only	SCR, BAK, COL	Simple early screens (1, 2, 3, 4)
obj	1	1991-05-23 only	CGX	Single early object tile

The d Subfolder: Dungeon Areas Named

The d subfolder is by far the most interesting part of the zelda tree. Its files are named by dungeon or overworld area — and the area names are directly readable:

Name	Japanese reading	Likely area
0-osiro	お城 — castle	Hyrule Castle or Light World castle zone
1-tika	地下 — underground	Underground passage tiles
10-nukemichi	抜け道 — shortcut passage	Hidden passage or cave area
11-kajiya	鍛冶屋 — blacksmith	Blacksmith or village tile set
12-sabaku	砂漠 — desert	Desert region tiles
13-pyramid	pyramid	Pyramid dungeon exterior

These names map well to the dungeon and overworld vocabulary of A Link to the Past. The numbered prefix (0-, 1-, 10-, 11-, 12-, 13-) is almost certainly an area ID scheme that matches the game’s internal map numbering.

The d subfolder also contains the newest files in the entire zelda tree — as late as 1994-07-25. That is nearly three years after the game shipped in Japan (November 1991). This strongly suggests zelda/d was actively used for follow-up work — potentially the eventual SNES remake, regional port, or a successor project — long after the original title was complete.

The m/tmp Sub-Snapshot

The m subfolder contains a nested tmp folder dated entirely to 1992-12-01. Its 34 files carry named tile types rather than numbered IDs:

• hokora — wayside shrine (祠)
• kumo — cloud (雲)
• kyokai — border / church (境界 or 教会)
• machi — town (町)
• mori — forest (森)
• caddata — likely a raw CAD tool export
• op-ed — opening / ending sequence tiles
• op-ed-nes — NES variant of the opening/ending tiles

This tmp snapshot reads like a world-map tile palette capture: terrain types (forest, town, shrine, cloud) alongside an opening/ending sequence set. The op-ed-nes file is especially notable because it preserves a NES-format tile set alongside the SNES assets — suggesting the team was still referencing the NES original in late 1992 when designing or revising world-map terrain art.

---

GB-zelda: The Game Boy Zelda Branch

GB-zelda is the largest of the three folders and structurally the most distinct. Its extension profile is strikingly different from zelda:

Extension	zelda	GB-zelda	Implication
.BAK	268	469	More revision history in the GB branch
.OBJ	2	147	Object focus is far heavier on the Game Boy side
.CGX	84	86	Similar tile bank count
.MAP	35	62	More map data in GB-zelda
.SCR	77	45	Fewer raw screen layouts; objects replace them
.COL	63	3	Almost no palette files — GB palette system is simpler

The steep drop in .COL makes sense for Game Boy hardware where the palette space is trivially small. The surge in .OBJ reflects a more object-composed screen architecture.

Localization Evidence

The s subfolder contains screen layout files with explicitly localised names:

• gameover
• gameover1
• gameover-France
• gameover-Germany
• gameover-usa
• fue-neiro — likely 「フエ音色」, a flute/ocarina sound timbre reference

Three named regional variants of the game-over screen (France, Germany, USA) are strong evidence that this branch was being prepared for a multi-region release. Link’s Awakening shipped in Japan in June 1993, in North America in August 1993, and in Europe in December 1993. The GB-zelda branch date range (1991-11-27 → 1994-08-02) brackets the entire localisation window exactly.

Object Taxonomy

The o folder and its sub-trees form a clear object classification system:

Sub-tree	Files	Date range	Reading
o (top level)	84	1991-11-29 → 1993-07-01	General objects; clear, numbered variants
o/cbos	21	1992-09-18 → 1993-02-27	C-type boss objects (numbered 1-11)
o/dbos	47	1992-08-07 → 1993-04-12	Dungeon boss objects (numbered 1-6, with 6-1 and 6-2 variants)
o/f	28	1992-04-15 → 1993-03-16	Field objects (f1-f12 + f-gomi = discarded field data)
o/h	33	1992-12-03 → 1993-03-17	House/hero objects (h1-h13)
o/d	57	1992-04-09 → 1993-02-25	Dungeon objects (d1-d13 + d-gomi = discarded dungeon data)
o/s	2	1992-07-07 → 1992-11-04	Small: op-2, s1
o/y	1	1993-02-12 only	Single residual: ygomi (discarded)

The -gomi suffix (ゴミ = garbage/junk) marks explicitly discarded or superseded files. f-gomi and d-gomi are named junk heaps — earlier object definitions that were replaced but not deleted, which is consistent with the cautious personal-backup pattern seen across the whole archive.

The cbos / dbos naming (c-type boss / dungeon boss) sets up a clear enemy taxonomy. Boss objects were separated from general field and dungeon objects and given their own numbered sequences.

The Panel and Transition System

The p subfolder and its sub-trees handle room transitions:

Sub-tree	Files	Date range	Reading
p (top level)	26	1992-03-04 → 1993-03-04	Staircase/step transitions (dan = stairs; dan-irekae = stair-swap variants)
p/f	38	1992-09-10 → 1993-04-08	Teleport/warp transitions (irekae-tenso = transfer/warp swap sequences)
p/d	10	1992-11-26 → 1993-02-09	Dungeon-specific transition data
p/y	4	1992-12-15 → 1993-03-16	Y-type transitions

Key vocabulary:

• dan — 段 (steps / stairs)
• irekae — 入れ替え (swap or replacement)
• tenso — 転送 (transfer or teleport)

So irekae-tenso = “swap-teleport” = warp-point transition screen. This is a named internal convention for the room-transition pipeline, not just a label someone chose at random.

---

How the Three Branches Relate

The simplest reading of the timeline:

1. May 1991: DELDA and zelda/d both start the same day — the workstation is set up for Zelda work
2. Oct 1991: DELDA closes; tile work has moved fully into the main zelda tree
3. Nov 1991: GB-zelda opens — a parallel Game Boy-targeted branch begins
4. Nov 1991: A Link to the Past ships in Japan; SNES work in zelda continues
5. 1992–1993: Both SNES and GB branches active; m/tmp world-tile snapshot dated Dec 1992; GB localization screens appear 1992-mid-1993
6. Jun 1993: Link’s Awakening ships in Japan; GB-zelda continues through mid-1994 (localization and maintenance)
7. Jul 1994: Last file in zelda/d — the SNES branch is still being touched three years after the original ship date, suggesting active follow-up or reuse

DELDA predates the platform split. zelda is almost certainly the SNES Zelda workspace — its dungeon area names, NES-reference tile sets, and multi-year span all fit a project that started as A Link to the Past development and continued into follow-up work. GB-zelda is the most likely Link’s Awakening workspace, given the regional game-over screens, object taxonomy depth, and date bracket.

---

Sugiyama’s Mixed Graphics Workspace

Sugiyama’s home is the second major component of NEWS_04. It is older than Arimoto’s, broader in scope, and reads like a workstation that served multiple productions over five years rather than one focused project.

Sugiyama at a Glance

Project	Files	Dominant types	Date range	Likely game
flyman	429	SCR 286, BAK 139	1989-10-13 → 1991-05-07	Pilotwings-era prototype (unconfirmed)
fly	388	BAK 137, CGX 102, SCR 78, COL 69	1989-10-13 → 1994-03-18	Pilotwings-era prototype (art side; 1994 date = tape restore)
CAR	415	BAK 198, SCR 148, CGX 38, MD7 3	1991-04-05 → 1994-03-18	Super Mario Kart ✓
SIM	165	SCR, OBJ, CGX, SFX	1990-11-27 → 1993-01-22	SimCity SNES (probable)
MARIO	77	CGX, SCR, COL, BAK	1993-04-08 → 1993-06-21	Mario title TBD
FX2	41	CGX, SCR, COL, BAK	1993-07-06 → 1993-12-08	Wild Trax / Stunt Race FX ✓

The “status at backup” column matters. Of the six Sugiyama projects, only fly and CAR were still receiving changes in March 1994 — the same date. flyman was last touched in May 1991 — nearly three years dormant. Everything else falls between.

---

fly and flyman: An Early Flight Game — Possibly a Pilotwings Prototype

fly and flyman start on the same date — 1989-10-13 — and share enough naming vocabulary to be two directories for the same project. The layout side (flyman) closed in May 1991. The art side (fly) continued to 1994-03-18 — the backup date.

Folder	File count	Dominant type	Role
fly	388	CGX 102, SCR 78, COL 69	Art production: tile banks, palettes, sprites
flyman	429	SCR 286, BAK 139	Screen assembly: stage layouts from the art in fly

The Pilotwings Hypothesis

The fly/flyman vocabulary divides into two distinct groups that sit in tension with each other.

Group A — flight disciplines (matching Pilotwings):

File	Reading
SKYDIVE.CGX/COL	Skydiving — Pilotwings discipline
HANG.CGX/COL, HANG-L.CGX	Hang gliding — Pilotwings discipline
PARA.CGX/COL/SCR, PARA-L.CGX	Parachuting — Pilotwings skydiving phase
ROCKET.CGX/COL	Rocket Belt — Pilotwings discipline
PLANE.CGX/COL	Light Plane — Pilotwings discipline
HELI.CGX/COL, HELI-L.CGX	Helicopter/Gyrocopter — Pilotwings later missions
ROKETMAN.CGX	Rocket Belt character sprite
MYSHIP.CGX, MY.SHIP.CGX	Player vehicle / craft

These six disciplines match Pilotwings (SNES, released November 1990 in Japan) almost exactly.

Group B — combat elements (NOT in Pilotwings as shipped):

File	Reading
BG-FORTRESS.CGX, BG-ENEMYSHIP.CGX, BG-BASESHIP.CGX	Enemy base/ship/fortress stage backgrounds
BOSS.CGX, BOSS-1/2/3.CGX/COL	Boss sprites — three numbered bosses
CHIKABOSS-01.OBJ, CHIKABOSS-02.OBJ	Underground boss (two-part object)
UFO-0/1/2.CGX	UFO enemy type (three variants)
CAMEL.CGX, CAMEL2.CGX, CAMEL3.CGX	Camel-type enemy (three variants)
CORE-1/2/3/4.CGX/COL	Destructible core enemies
OBJ-BOMB.CGX, OBJ-BOMB-2.CGX	Bomb objects

Pilotwings has no enemies, no bosses, and no combat.

English localization files confirm Western release intent across both groups: LICENSE-ENG.CGX, ROGO-ENG.CGX, CON-ENG.CGX, PAUSE-ENG.CGX, LINE-ENG.CGX.

The Most Likely Reading

The simplest explanation that fits all the data:

fly/flyman is a prototype of what eventually became Pilotwings, at an earlier development stage when combat/enemy elements were still part of the design.

Pilotwings was in development from approximately 1989 — exactly when this folder opened. The project’s layout side (flyman) was closed six months after Pilotwings shipped (November 1990 + ~6 months = May 1991). If flyman represented the shipping branch, the six-month tail is consistent with post-ship polish, documentation, or internal archival.

Alternatively, this could be a parallel unreleased game using the same Mode 7 flight framework but with a combat layer that was never completed — a direction Nintendo tried alongside Pilotwings and then abandoned.

The fly art folder remaining active until 1994 (the backup date) is most likely explained by the tape restore itself refreshing file timestamps.

Mode 7 Landscape Architecture in fly

The 47 M7-* files form the most detailed Mode 7 art system in the archive:

File group	Reading
M7-BG-L.CGX, M7-BG-L-NIGHT.CGX	Mode 7 landscape tile banks (standard and night variants)
M7-BG-RACE.CGX, M7-BG-JUMP.CGX	Race-stage and jump-stage Mode 7 tiles
M7-BG-HELI.CGX, M7-BG-DESERT.CGX	Helicopter-stage and desert-stage tiles
M7-BG-BONUS.CGX	Bonus-stage Mode 7 tile bank
M7-BG-C0.CGX, M7-BG-C00.CGX, M7-BG-C01.CGX	Course-specific tile banks (3 courses)
M7-L-FINE/GRASS/ISLAND/NIGHT/RACE/RAIN/SNOW/SUNSET/DESERT.COL	Nine weather/terrain palette variants for the Mode 7 ground
M7-CHIKA/CHIKA-B/C/FORTRESS.COL	Underground and fortress-stage Mode 7 palettes
M7-METER.SCR, M7-METER-B.SCR	HUD meter drawn against the Mode 7 plane

Nine named weather-palette variants (Fine, Grass, Island, Night, Race, Rain, Snow, Sunset, Desert) is a strong hallmark of a Pilotwings-style flight sim, where the ground appearance changes with conditions. No SNES shooter or action game from this era used nine weather-palette variants for a Mode 7 ground plane.

The flyman Stage Structure

flyman is dominated by MAP-prefixed screen files:

Stage group	Files	Reading
MAP1 through MAP8	4–64 per group	Eight numbered lesson/mission stages
CHIKA-*	~20	Underground stage screens (BG-CHIKA-A/B/C variants)
BONUS / BONUS1-4	5	Bonus stage layouts
POOL1 / POOL2 / POOL3	5 each	Water stage screens
DESERT1	~5	Desert stage layouts
JUMP1 / JUMP2 / JUMP3	7 each	Jump-sequence screens
RACE1	4	Race-course screens
BGBG-HA, BGBG-HELI, BGBG-PA, BGBG-PL, BGBG-RO	5	Double-layer background composites

MAP6 holds 39 files and MAP7 holds 64 — suggesting those were the most complex or iterated missions. The BGBG- (background-over-background) composites are a layering technique not seen in the other projects in this archive. The suffix abbreviations HA (hang?), HELI, PA (para?), PL (plane?), RO (rocket?) match the flight disciplines named in fly.

---

CAR: Super Mario Kart

CAR’s identification is not ambiguous. Five files make the project unmistakable:

File	Why it matters
MARIO-CAR.CGX, MARIO-CAR.OBJ	Literally the Mario Kart vehicle sprite sheet and object
JUGEM.CGX, JUGEM.OBJ	Jugem (ジュゲム) is Lakitu — the cloud-riding character who lifts fallen racers and waves the start flag in Super Mario Kart
DOKAN.CGX, DOKAN.COL, DOKAN.SCR	Dokan (土管) = pipes; warp-pipe obstacles that line the Mario Kart courses
POLE.CGX, POLE.OBJ	The finish-line goal pole
SLOT.CGX	The item-box roulette drum

No other Nintendo game combines JUGEM, DOKAN, POLE, and SLOT as distinct race-side objects. This is Super Mario Kart (released Japan August 1992, North America September 1992, Europe January 1993).

Mario Kart UI Screens Preserved

The non-track content is equally direct:

File	Reading
CAR-SELECT.SCR, CAR-SELECT2.SCR, CAR-SELECT3.SCR	Kart / character selection screens (three iterations)
CAR-SELECT2-ENG.SCR	English localisation of the car-select screen
MAP-SELECT.SCR, MAP-SELECT2.SCR	Course/cup selection screens
MAP-SELECT-ENG.COL	English colour set for the map-select screen
RESULT.SCR, RESULT2.SCR	Race result screens
RESULT-ENG.SCR, RESULT2-ENG.SCR	English race result screens
D-POINT.SCR, D-POINT-ENG.SCR	Death/danger-point screen (when you fall off)
REGI.SCR	High-score / ranking register screen
SROT-DRAM.SCR	Sort drum — the animated item-roulette scroll sequence
DEMO-MOJI.CGX	Demo-mode text graphics
END-MOJI.CGX	Ending-sequence text
HATA.CGX, HATA.SCR	Racing flag art
BG-ITEM.CGX	Item-box background tile sheet

Track Structure

The track coding system uses single-letter family prefixes, each with four segment views (#-0 through #-3):

Family	Tracks	Special files	Most likely reading
STAR	1	STAR-B.CGX, STAR.CGX/COL	Star Cup — named explicitly
B	B1–B6	—	Six B-course prototypes (most fully realised family)
C	C1–C4	C1-B.CGX, C1.CGX/COL, C.MD7, CCC1.MD7	C-courses + two 32 KB Mode 7 map tables
D	D1–D3	D1-B.CGX, D1.CGX/COL	D-courses
G	G1–G3	G1-B.CGX, G1.CGX/COL	G-courses
H	H1–H3	H1-B.CGX, H1.CGX/COL, H1-Z.COL	H-courses; Z variant = special palette
K	K1–K3	K1-B.CGX, K1.CGX/COL	K-courses
S	S1–S2	S1-B.CGX, S1.CGX/COL, S1-Z.COL, S.MD7	S-courses + one 32 KB Mode 7 map table
W	W1–W2	W1-B.CGX, W1.CGX/COL, W1-B.SCR	W-courses

The three 32,768-byte .MD7 files (C.MD7, CCC1.MD7, S.MD7) are raw SNES Mode 7 background maps: 128 × 128 entries × 2 bytes = exactly 32 KB. Each one is a complete race-track ground-plane bitmap. The companion C.DAT and CCC1.DAT files (117 bytes each) are likely track parameter tables (lap count, speed, scroll indices).

B1–B6 is the most fully developed family with six courses, each having four segment views. The final Super Mario Kart has 20 tracks (16 race + 4 battle) across four cups — the prototype in CAR has more distinct family codes than the final game, consistent with a development phase where courses were still being created and culled.

English Localization Evidence

Super Mario Kart shipped in all three major markets within its first year. CAR shows at least two rounds of English title art:

• TITLE-ENG.CGX/SCR — English title screen (first revision)
• TITLE-ENG2.CGX/SCR, TITLE2-ENG.SCR — English title screen (second revision)
• SELECT-ENG.CGX, OBJ-MOJI-ENG.OBJ, MOJI-ENG.CGX — English text art for menus and object labels

CAR, MARIO, and FX2 as a Production Timeline

With the identification confirmed, the three Sugiyama racing/Nintendo projects form a coherent timeline:

Project	Date range	Game	Notes
CAR	1991-04-05 → 1992+	Super Mario Kart	Mode 7 track maps; development started ~16 months before Japanese release
MARIO	1993-04-08 → 1993-06-21	Mario title TBD	Game-select screen art; short engagement
FX2	1993-07-06 → 1993-12-08	Wild Trax / Stunt Race FX	Player/cup select art; the last Sugiyama project before the backup

---

MARIO and SIM: Two More Identifiable Fragments

The MARIO directory (77 files, ~10 weeks mid-1993) is a short engagement:

• GAMESELECT.CGX, GAMESELECT.COL, GAMESELECT.SCR — game-select screen art
• GAMESELECT-N.*, GAMESELECT-P.* — N (Nintendo?) and P (Player?) variants
• MA-ROGO-OBJ.CGX — Mario logo object tile sheet
• 2PR-S1.* — two-player layout variant

This is game-select and mode-select artwork for a Mario-branded title. The 2PR (2-player) flag and N/P variants suggest a versus or multiplayer mode game, but the small scope and short timeline make it hard to identify conclusively.

The SIM directory (165 files, November 1990 → January 1993) has a different character altogether:

• SELECT-SCENARIO.SCR, SELECT-SCENARIO-2.SCR, SCENARIO.OBJ — scenario selection screens
• MAP-SELECT.SCR, MAP-SELECT2.SCR — map selection screens
• TOWN.SCR, TOWN2.SCR — town view screens
• LEVEL1.SCR, LEVEL2.SCR — level screens
• INPUT.OBJ, INPUT-BG.CGX, INPUT1.SCR, INPUT2-KEY.SCR — data-entry / controller-input screens
• icon_p.CGX, icon_p-F/G.CGX, iconpd.CGX, optishd.CGX — UI icons (F/G = flag/green variants?)
• .SFX files paired with each main screen (a screen-effects or screen-state format)

The date range (Nov 1990 → Jan 1993) brackets the Japanese release of SimCity for SNES (August 1991, published by Nintendo). The vocabulary — SELECT-SCENARIO, TOWN, MAP-SELECT, LEVEL — matches SimCity’s UI structure almost exactly. The .SFX format is distinctive and not seen elsewhere in this archive, suggesting a different toolset used specifically for the SimCity UI layer.

FX2: Wild Trax / Stunt Race FX

FX2 (41 files, July–December 1993) confirms Sugiyama’s involvement in the Wild Trax / Stunt Race FX project:

• cpt-1.CGX, cpt-1.COL, cpt-1.SCR — cup / captain art (note lower-case naming vs. all other Sugiyama projects)
• cpt-2.CGX, cpt-2A.SCR
• p-select.CGX, p-select.COL, p-select.SCR — player-select screen
• test-1.OBJ

The lower-case naming is unusual and may reflect a later tool version or a different artist’s convention applied at handoff. Wild Trax shipped July 1994 — exactly seven months after FX2 went quiet on this machine.

---

Sugiyama’s Project Timeline at a Glance

With identifications in place, the complete Sugiyama timeline becomes readable:

Period	Project	Status
1989-10-13	fly + flyman open together	Early SNES flight game begins (Pilotwings-era prototype?)
1990-11	Pilotwings ships	Layout side (flyman) winds down toward May 1991
1990-11-27	SIM opens	Likely SimCity SNES UI/art work begins
1991-04-05	CAR opens	Super Mario Kart development begins
1991-05-07	flyman last modified	Layout side of early flight game finalised
1991-08	SimCity SNES ships	SIM continues until Jan 1993 (post-ship polish or localization)
1992-08	Super Mario Kart ships (Japan)	CAR continues — localization and follow-up work
1993-01-22	SIM last modified	SimCity work ends
1993-04-08	MARIO opens	Brief Mario title UI engagement begins
1993-06-21	MARIO last modified	Mario UI work ends
1993-07-06	FX2 opens	Wild Trax / Stunt Race FX art begins
1993-12-08	FX2 last modified	Wild Trax art ends on Sugiyama’s machine
1994-03-18	Tape backup	fly and CAR show this date — almost certainly the restore timestamp, not active edits

fly and CAR sharing the exact same March 18, 1994 timestamp is most simply explained by the tape-restore process itself refreshing modification times for files that were open or had been accessed during the archive restore. The project timelines — fly/flyman centred on 1989–1991 and CAR on 1991–1992 — do not otherwise overlap.

---

A Practical Graphics Workflow Hiding in the File Types

One of the most useful things about NEWS_04 is that it preserves the shape of a 2D asset workflow. The repeated file groupings make the likely pipeline much clearer than any single file would.

flowchart LR
  A["<b>Graphics Bank</b><br>CGX character or tile art"] --> B["<b>Palette</b><br>COL color set"]
  B --> C["<b>Screen Layout</b><br>SCR assembles the scene"]
  C --> D["<b>Object / Map Side</b><br>OBJ, OBX, MAP, PNL"]
  D --> E["<b>Revision Loop</b><br>BAK copy preserved before next edit"]

That pattern appears over and over again:

• a .CGX graphics file
• a matching .COL palette file
• a matching .SCR composed layout
• sometimes an .OBJ, .MAP, or .PNL companion file
• then a .BAK copy of one or more of them

So even without a formal tool manual, the workstation backup shows how artists likely worked in practice:

• create or update a graphics bank
• pair it with the correct palette
• compose it into a screen or layout
• save a manual backup before the next round of changes

That is exactly the kind of process evidence that polished source archives usually erase.

---

Why NEWS_04 Matters Beside the Other NEWS Tapes

Each big NEWS tape now has a different role:

Archive	Main value
NEWS_04	Mixed 2D graphics and layout workstation snapshot across several projects, with a particularly late SF2 branch
NEWS_05	Star Fox 2 3D CAD and toolchain snapshot with source code and animation pipeline
NEWS_09	Yoshi’s Island supplementary art workspace
NEWS_11	Larger and richer late Yoshi’s Island art/archive workspace

That makes NEWS_04 the missing complement to NEWS_05. If NEWS_05 tells us how Nintendo built Star Fox 2’s polygon assets, NEWS_04 helps show how the same broader development environment handled 2D banks, screen composition, UI, and sprite/layout-side resources.

It also matters because it preserves older project residue rather than only one final branch. The mixed zelda, GB-zelda, MARIO, CAR, fly, and FX2 folders make it feel like a real artist workstation that stayed in use across multiple productions.

---

What Seems Most Important to Study Next

The main project layers are now mapped and three of six Sugiyama projects are identified.

1. Identify the fly/flyman Game Definitively

The case for a Pilotwings-era prototype is strong but not proven.
The most useful next step would be:

• Comparing the fly Mode 7 palette set (Fine, Grass, Island, Night, Rain, Snow, Sunset, Desert) against the landscape backgrounds in the Pilotwings ROM
• Checking whether the ROKETMAN, MYSHIP, HANG, PARA sprites match any assets visible in Pilotwings debug or prototype builds
• Looking at whether the flyman MAP1-8 structure matches Pilotwings’ lesson/mission numbering
• Searching other Nintendo Gigaleak archives for a project name matching FLY or the 1989 development start

2. The zelda/d Area ID Scheme

The d subfolder in zelda names areas by number: 0-osiro, 1-tika, 12-sabaku, 13-pyramid. A full catalogue could map directly onto A Link to the Past’s internal area list — or diverge, pointing to a follow-up or prototype.

3. The Super Mario Kart Track Families

CAR has more track families than the final game. Matching B, C, D, G, H, K, S, W to their final SMK equivalents (or confirming which were cut) is tractable: the three 32 KB .MD7 files can be parsed as raw Mode 7 background tables to reconstruct actual track layouts.

4. Exact Toolchain Recovery

The .CGX → .COL → .SCR → .OBJ/.MAP/.PNL pipeline is consistent across every project on this machine. Matching it to known Nintendo SNES dev tools (or other archived tool documentation from the Gigaleak) would make the whole file set far more navigable.

---

Working Conclusion

NEWS_04 is not the most glamorous archive in the NEWS tape set. It does not have the source-code shock value of NEWS_05, and it is messier than the Yoshi tapes. But it is one of the most workshop-like snapshots in the leak — and, on examination, one of the most historically rich.

Its real value is that it preserves how graphics work accumulated on a live Nintendo workstation:

• old project folders kept for years — some untouched since 1991
• hundreds of incremental .BAK revisions showing the iteration rhythm
• tightly grouped graphics/palette/screen triples throughout every project
• three distinct Zelda layers — early prototype, long SNES branch, and localised Game Boy branch — on the same machine
• Super Mario Kart development material including Mode 7 track map tables, JUGEM, DOKAN, POLE, and SLOT files
• an early SNES flight game (possibly a Pilotwings prototype) from October 1989 — among the earliest SNES development work visible in any Nintendo archive
• probable SimCity SNES UI work bracketing its August 1991 Japanese release
• one especially late Star Fox 2 graphics branch from 1995 sitting beside all of the above

So NEWS_04 fills an important gap. It does not tell us how Nintendo wrote the games. It tells us how one developer’s workstation accumulated half a decade of visual asset history — from early SNES hardware exploration in 1989 all the way through to the final Star Fox 2 art push in 1995.