AI summary by TwinMind
Why is Japan So Weak in Software?
Asianometry
19m 37s
Japan's Digital Trade Deficit Crisis
- 2023: Japan ran digital trade deficit of 5.5 trillion yen ($37 billion)
- Imported trillions more in software licenses, cloud services than exported
- Deficit doubled from 2015 levels
- Hardware companies iconic (Sony, Fujitsu, Hitachi, Panasonic, Canon, Sharp, Toshiba) but no software giants like Microsoft, Oracle, Adobe, SAP
Early IBM Dominance and Government Response (1925-1971)
- IBM entered Japan 1925 selling tabulating machines
- 1937: IBM Japan subsidiary established, assets confiscated during WWII
- 1949: IBM returned, 1950 law mandated foreign companies enter joint ventures
- RCA with Hitachi, GE with Toshiba, Honeywell with NEC
- IBM owned 100% of IBM Japan, refused joint ventures
- MITI-IBM clashes over market access, government pushed "buy local" policies
- Customers favored IBM due to technical superiority
- 1964: IBM System 360 alarmed MITI, triggered Super High Performance Computer Project (1966-1971)
Japan Software Company Failure (1966-1972)
- October 1966: MITI sponsored Japan Software Company
- Joint venture: Hitachi, NEC, Fujitsu, Industrial Bank of Japan
- Goal: common language software, universal OS for Japanese computers
- Only 25% of 10 billion yen budget allocated to software
- Failed due to ballooning requirements, no outside sales, lost worker confidence
- Closed 1972 after government subsidies ended
IBM Unbundling and Japanese Response (1969-1980s)
- 1969: IBM unbundled software/hardware due to antitrust threat
- Created US software industry, enabled small companies to compete
- Japan: Different outcome - computer makers created software subsidiaries
- 1971: Fujitsu began selling IBM-compatible mainframes (faster, cheaper)
- Collaboration with Gene Amdahl (System 360 architect)
- Japanese companies reverse-engineered IBM software
- Modified for Japanese kanji symbols
- Customized for business practices to lock in customers
- Japanese companies became "little IBMs"
Software Theft and Legal Consequences (1982-1985)
- Reverse engineering was theft and illegal
- IBM initially didn't act due to:
- No definitive copyright protections until 1976/1978
- Defending against massive antitrust case (ended 1982)
- Losing PC market share to clones
- 1982: IBM confronted Fujitsu, Hitachi with theft allegations
- IBM spy case: 18 Japanese employees arrested in elaborate sting
- Settlement: Fujitsu paid lump sum + annual licensing fees up to $60M/year
- 1985: MITI proposed reducing software copyright from 50 to 15 years (failed)
Software Crisis and Programmer Shortage
- Japanese government recognized serious software crisis
- Major programmer shortage due to:
- Universities lacked qualified computer science teachers
- Teachers trained in other disciplines (Math), lagged on trends
- Lack of English/overseas experience
- Software engineering not attractive career
- Developers not customer-facing, poor career progression
- Low pay, unfulfilling work (customization/maintenance)
- 2010: 80% of Japan's software spend on maintaining existing systems vs 60% in West
Lack of Software Startups
- 1980s US: Founded Adobe, Synopsys, Cadence, Lotus, Electronic Arts, Autodesk
- Microsoft, Oracle hit stride in 1980s
- Japan failed to produce independent software giants due to:
- Lack of customer acceptance
- Insufficient venture capital
- Lifetime employment system
- Cultural lack of initiative/risk-taking
- No widely accepted platform/OS independent of hardware
- Japanese engineers unproductive without common development environment
TRON Operating System Project (1984-1989)
- 1984: Dr. Ken Sakamura (University of Tokyo) initiated TRON
- Goal: "open" OS to break Microsoft/Intel grip
- Produced interfaces/design guidelines, not source code
- Business TRON (BTRON): PC makers like Matsushita adopted standards
- April 1989: US Trade Representative listed BTRON as potential trade barrier
- Japanese microprocessor companies pulled out, BTRON collapsed
- TRON still exists: Nintendo Switch Joy-Con controllers, Hayabusa 2 space probe
SIGMA Project (1985-1990)
- MITI-sponsored consortium, 25 billion yen funding
- Goal: improve Japanese programmer productivity
- Developed standards for networked Unix workstations
- Caused Fujitsu, Hitachi, NEC transition from IBM platforms to Unix
- Heavily customized versions: Fujitsu UXP, NEC Super-UX, Hitachi HI-UX
- Success in supercomputers (challenged Cray Research) but remained niche
Economic Bubble Burst and Decline (1990s)
- 1990: Real estate bubble popped, long recession began
- Customers couldn't afford expensive software modifications
- Japanese mainframe makers struggled with cost-conscious environment
- US pulled away based on software advantages
- MITI's reverse-engineering playbook failed for software
- Software success depends on standards, networks, familiarity - not replication economics
PC Market Collapse
- NEC dominated domestic PC market with PC-98 (MS-DOS with Japanese support)
- Early 1990s: IBM/Microsoft/Intel cracked NEC's monopoly
- New DOS version with Japanese support opened market to foreign products
- MITI decided on globalized approach, let market play out
- NEC's overpriced, simple products crumbled
- Cemented Microsoft's Japanese market domination
Mobile Internet Missed Opportunity
- NTT DoCoMo created i-mode (mobile 2G web browser, 1999)
- 40+ million subscribers by 2003
- Failed to export standard abroad due to lack of handset partner adoption
- iPhone and rich app ecosystem closed window of opportunity
Current State and Conclusion
- Japan has large software industry but dominated by:
- Custom packages with no sales potential
- Imported products made abroad
- Brightest spot: gaming industry
- Japanese programmers considered skilled without American-scale salaries
- Ruby language created by Japanese (Rails framework by Danish developer)
- Continued lag across multiple technological eras
- Few coherent policy solutions identified
