Video summary

COBOL: The Language Your Bank Is Afraid to Touch

Main summary

Key takeaways

Educational

Main ideas, concepts, and lessons

  • Cold War pressure drove a need for portable business software

    • After the Soviet Sputnik launch (Oct 4, 1957), the U.S. massively increased spending on science, defense, and computing.
    • The Pentagon needed large volumes of software quickly, often written by people not necessarily trained in mathematics.
    • Major issue: every computer vendor had its own assembly language (IBM, Univac, Honeywell, Burroughs, etc.).
    • Result: programs were machine-specific, forcing the government to rewrite software repeatedly for different systems.
    • DoD concluded it was spending more on translating software than on buying hardware.

  • A government-led effort created a standard business programming language

    • In April 1959, the U.S. Department of Defense convened a meeting in Washington with a “simple on paper” goal:
      • Agree on one common programming language for business data processing
      • The language should run on any machine
    • A committee produced a language, and Grace Hopper (a Navy officer) pushed the standard forward despite resistance from major computer manufacturers.
    • The language was COBOL: Conference on Data Systems Languages.

  • COBOL’s defining philosophy: readable code for non-mathematicians

    • Hopper’s earlier tool, Flow-Matic, influenced COBOL’s syntax by translating English-like statements into machine instructions.
    • The argument (as presented): to enable non-mathematicians to write software, the language must look like something they can read.
    • The 1959 committee used Flow-Matic as the primary source for syntax.

  • COBOL structure: four divisions

    • A COBOL program is divided into:
      • Identification Division: names the program
      • Environment Division: describes the hardware environment
      • Data Division: defines variables, fields, and file structures
      • Procedure Division: contains the program’s actual logic
    • Explicit design goal: a bank manager (non-programmer) should be able to follow a COBOL listing.

  • Rapid adoption and long-term dominance in business systems

    • COBOL spread quickly through the 1960s.
    • By 1970, there were more COBOL lines in production than any other language.
    • Adoption drivers:
      • Federal government requirements
      • Banks, insurance companies, and payroll systems
    • IBM’s System/360 (1964) accelerated adoption:
      • It supported COBOL natively
      • Organizations wrote COBOL apps because the compiler and support were readily available
    • The core reason it persisted:
      • Systems were built to run for years and decades
      • Organizations avoided disruption because the cost/risk of change was high.

  • Scale and ubiquity of COBOL today

    • The video cites estimates such as:
      • ~220 billion lines of COBOL in active production
      • Processing roughly $3 trillion in commerce
    • Examples mentioned:
      • SSA (Social Security Administration) payments
      • IRS tax processing
      • About 95% of ATM transactions touching COBOL
      • Retail card settlement at grocery stores likely using COBOL on the bank side

  • Why banks avoid changing COBOL

    • Fear of breaking systems is central.
    • Many systems reportedly have had:
      • Minimal modification since the 1970s/1980s
      • Edge cases already handled over decades (rare transactions, tax edge cases, leap-year logic, etc.)
    • Replacement attempts can fail:
      • Commonwealth Bank of Australia: ~5 years, ~$750M (2008–2013), abandoned much of it
      • UK TSB Bank: migration (2018) caused customer lockouts, CEO resigned, significant compensation/remediation costs

  • Grace Hopper’s lifelong influence and the “port” metaphor

    • Hopper’s career details are used to reinforce her role:
      • Nobel-era computing work (Mark I/Mark II)
      • “Bug” story from a moth trapped in a relay
      • Built Flow-Matic and pushed COBOL standardization
    • She continued lecturing into her 80s with a “nanosecond” wire and explained satellite communication lag.

    • Quote (as presented) from a 1987 David Letterman interview:

      “A ship in port is safe, but that’s not what ships are for.”

    • The video frames an irony:

      • What she helped create became the industry’s safest “port”—hard to move because of reliability.

  • COBOL’s durability is a feature: backward compatibility

    • COBOL evolved, but the video emphasizes continuity:
      • COBOL 2002: adds object-oriented features
      • COBOL 2014: floating-point improvements and better Unicode support
    • Programs written in 1972 can still compile and run on modern systems.
    • The video claims backward compatibility is carefully maintained so that older programs remain operational.

  • Ongoing demand due to retirements and recurring crises

    • Y2K (1999) made COBOL’s importance visible:
      • Two-digit years like “99” could be read as 1900 instead of 1999
      • Large portions of the $300B global fix were in COBOL
      • Many COBOL programmers were brought back because they were the only ones who understood the code
    • A similar pattern is described for 2020 COVID-19 disruptions:
      • Some state systems failed when unemployment claims spiked
      • New Jersey asked for COBOL programmers urgently
      • Kansas Governor Laura Kelly asked on TV for volunteers with COBOL knowledge
    • Workforce issue:
      • Average experienced COBOL programmer age is estimated over 55
    • Mitigation efforts:
      • IBM training programs with universities (since 2019)
      • Online courses
      • MicroFocus modern tooling (IDE integration, cloud/Kubernetes/container support)

Bullet-point methodology / instruction-like content (if any)

  • Standardization approach (historical process described)

    • Convene a government committee to:
      • Select a single common language for business data processing
      • Ensure it can run on multiple hardware platforms
    • Use an existing compiler/syntax concept as a foundation:
      • Apply ideas from Flow-Matic (English-like statements translated to machine instructions)
    • Define a clear, structured program format:
      • Identification
      • Environment
      • Data
      • Procedure
    • Maintain strict backward compatibility to preserve long-running systems:
      • Keep old COBOL code compiling/running across decades

  • Modernization/continuity approach (as implied by current tooling)

    • Instead of rewriting everything:
      • Use modern COBOL compilers/tooling (e.g., IDE/cloud features)
      • Run COBOL in containers (e.g., Kubernetes) alongside other services (e.g., Python microservices)

Speakers / sources featured (as mentioned in the subtitles)

  • Grace Hopper (Navy officer; central figure credited with pushing COBOL)
  • David Letterman (interviewer referenced for Hopper’s quote)
  • Laura Kelly (Governor of Kansas; asked for COBOL volunteers on television)
  • United States Department of Defense (convened meeting in 1959)
  • Soviet Union (launched Sputnik—trigger for the funding/computing surge)
  • IBM (System/360 support and training programs with universities)
  • MicroFocus (modern COBOL compilers/tooling)
  • JP Morgan Chase (publicly acknowledged COBOL systems)
  • Wells Fargo (made similar disclosures)
  • Bank of America (made similar disclosures)
  • Commonwealth Bank of Australia (described COBOL replacement effort)
  • TSB Bank (described failed migration)
  • Social Security Administration (SSA) (COBOL use)
  • IRS (COBOL use)
  • New Jersey (asked publicly for COBOL programmers)
  • Kansas (governor-led request)

Original video