Lecture 4.md (4988B)
1 +++ 2 title = "Lecture 4" 3 +++ 4 # Lecture 4 5 6 People started playing with computers to appropriate them. 7 8 Appropriating — using it for something other than its purpose (e.g. sound) 9 10 Real time computing: 11 12 - it was never an obvious thing 13 - Whirlwind/SAGE 14 - project Whirlwind flight simulator (1944) 15 - for Jay Forrester from MIT 16 - wanted for military training, as universal flight trainer 17 - it didn’t work out, took too long 18 - SAGE for the same thing 19 - IBM, Burroughs, Bell labs 20 - learning about core memories, printed circuits, mass storage, programming 21 - operaional in 1963, cost $8b 22 - but by that time ICBMs were operational 23 - at the end was useful for regulating plane traffic, due to cameras 24 - others: Hewlett SABRE, ATM, UPC 25 - general ideal of “cashless society” (like Diners club) 26 - Barclays cash dispenser in London 1967 — robot cashier 27 - credit card system, ATM and VISA 28 - universal product code (1973) — barcode 29 - this was in the US 30 - of course Europe had to make their own, Intl. Product Code (CIPC) in 1974 31 - Barclays & Burroughs for Decimal Day (Feb 15, 1971) 32 - could’ve been IBM, but Burroughs was more British 33 - building a B8500 to connect to TC500 terminals 34 - took forever, but they managed to sell a nonexistent computer for like 4 million. 35 - lots of problems. they were late with delivery, the Burroughs B8500 programmer left... 36 - in the end, Barclays went with IBM 37 38 Agendas: 39 40 - selling machines — for IBM and Burroughs 41 - academic discipline — Dijkstra wanted it mathematical. cybernetics, logic, sharing, calc. 42 - thinking machines 43 - programming: first wires/tapes/punch cards, then FORTRAN/COBOL 44 45 Programming: 46 47 - language ALGOL60 48 - elegant, universal, satisfies European sense of clarity & order 49 - but from US point of view, it was too academic, inflexible, and hard to learn. better in theory than in practice. but this was mostly ‘regular’ people in user groups, academics liked it. 50 - two issues: make old programs run on new machines, ease of programming 51 - multiple working groups: IEEE, SHARE, ... 52 - SHARE & IBM decided not to go with it. IFIR did. 53 - ACM set it as standard for publication of scientific algorithms 54 - FORTRAN 55 - John Backus, researcher from IBM, produced Formula Translator (FORTRAN) in 1953-1954 56 - one statement would produce many machine instructions, giving programmer more power and making shit easier 57 - his main point was economic. half of cost of running computer center was salaries for programmers, and "programmign and debugging accounted for as much as three-quarters of the cost of operating a computer” 58 - this is why IBM gave him support for developing FORTRAN for the new model 704 59 - main aim was efficiency, elegance of language came second. 60 - used mathematical formula syntax 61 - trying to make a system that could write programs as well as human programmers could 62 - it soon became the most widely used programming language, a ‘standard’ for scientific applications 63 - it spread organically, “by accident”, and universities and colleges eagerly started teaching and using it 64 - COBOL 65 - Common Business Oriented Language (COBOL) was created as standard by US government. 66 - every time the government changed their computers, all programs had to be rewritten, which was expensive and took up time. 67 - in 1959, the government sponsored Committee on Data Systems and Languages (CODASYL) to create a new standard language for data processing 68 - syntax was very similar to english, so non-programmers (managers, administrators) could still feel like they can understand the programs 69 - manufacturers didn’t want to accept it because they liked to be different from others. but then the government decided that it would not lease or buy any new computers without COBOL compilers unless the companies could prove it was useless, and of course nobody did, so everybody started adding COBOL compilers 70 - COBOL was taught on blackboards 71 - companies started providing computer services at start of 60s: maintenance, building/tuning, batch processing 72 - programming started becoming a job, not just spare time hobby 73 - software started becoming an economic commodity 74 75 ## Rise of software industry 76 77 in the 60s, a company couldn’t sustain itself just with software, they also needed maintenance, batch processing, building 78 79 1968 IBM “Unbundling” — software became separate from hardware 80 81 Software crisis as a result: 82 83 - IBM/360 was late af, adding programmers simply couldn’t scale it up 84 - “Hardware developed faster than software developers” is an EU/US agenda, academic, theoretical. 85 - “Nobody knew ho to write proper code” (important in NL, Dijkstra tried to solve this) 86 - there is also the “there wasn’t a crisis” POV 87 - agenda was set by academics 88 - programmer became a profession, informatics a science