Networks of Tinkerers a Model of OpenSource Innovation

Networks of Tinkerers: a Model of Open-Source Innovation Peter B. Meyer BLS brown bag seminar, March 15, 2006 1

Introduction Hobbyists have developed important technology. q open source software n q personal computers n q Linux; email processing; Web servers/browsers Homebrew Club of hobbyists, circa 1975 airplanes – a clearly documented case Goal here: try some assumptions about the hobbyists / tinkerers show they would share information in networks explore assumptions a bit 2

Aircraft developments From the 1870 s to the 1900 s there is a lot of “open-source” innovation in aircraft q q 1800 – George Cayley and many others try aeronautical experiments starting 1870 s, several aeronautical journals 1893 Octave Chanute’s Progress in Flying Machines 1903 – Wrights fly 3

Assumptions for micro model n Assume there are motivated tinkerers q n Assume they have a way to make “progress” q n We do observe this defining progress carefully Assume total technological uncertainty q q No market is identifiable so no R&D / competition The tinkerers would share information 4

The Tinkerer has activity/hobby A. (for “aircraft” or “activity”) Tinkerer receives positive utility from A of at period. • a 0 is known • later choices and rules determine at β is a discount factor between zero and one (assume. 95) applied to future period utility. Net present expected utility: 5

Tinkering rules n n Tinkerer may invest in ("tinker with") A The agent thinks that tinkering this period will raise all future period payoffs at by p units each time period. q p stands for a rate of progress, which is subjectively experienced by the agent q q We assume p is fixed and known to the agent Example: . 07 6

Tinkering decision n Tinkerer weights estimated costs and benefits n Benefits from one effort to tinker equal p in each subsequent period. n The present value of those payoffs is: Tinkerer compares those gross benefits to the cost which is 1 utility unit 7

Rates of Progress must meet the criterion above for tinkering to be worth it Progress is subjective There are not many tinkerers working on this activity who can make this much progress. More issues later. 8

Payoffs from endless tinkering Payoffs period 0 period 1 period 2 period 3 period 4 … -1 p p (pβ/[1 -β] ) -1 p p p β() -1 p p β β () -1 p β β β() -1 . . . a 0 -1+4 p a 0 -1+tp a 0 -1+p a 0 -1+2 p a 0 -1+3 p Present value of all that at time zero has a closed form: 9

A network of two tinkerers n Case of two tinkerers with identical utility functions q p 1 and p 2 – subjective rate of progress q n Their innovations are useful to one another Tinkerers form a network Present value of expected utility: 10

Subgroups of occasional tinkerers n n n A group of slow-progress tinkerers might agree to work together to generate progress rate p. Then the group acts like a single “tinkerer” in terms of its output and also in its incentive to join other groups There are something like economies of scale here; it’s a positive sum game. So Wilbur and Orville Wright could be one tinkerer maybe also: q q q Boston-area group All readers of a certain journal Kite people, together, as distinguished from balloon people 11

“Progress” is subjective 12

Standardization and Specialization n n Only the fraction f (between 0 and 1) of experiments by player two are usable to player one Suppose for a cost cs player one can adjust his project to look more like the other tinkerer’s project And, that this would raise the usable findings to f 2 That’s standardization Present value of utility from standardizing is: 13

Standardization and Specialization Key comparison is above Player one benefits more from standardizing if, ceteris paribus: n n q q q n n the other tinkerers are producing a large flow of innovations p 2; the cost of standardizing cs is small; the gain in useful innovations from the others (f 2 -f) is large. Same logic supports specialization These are technology phenomena, not requiring market processes 14

Searching and Joining n Suppose there is a cost to joining the network q costs of subscribing, paying attention q it’s worth the cost to a tinkerer if n the cost (cj) is low n he values future outcomes a lot n the others are producing a lot of progress (p) n their progress is useful to him – f is high enough n Suppose there is also a cost to searching for new members q Chanute wrote book q others published journals q Then the search costs affect innovative output (Web has effect) q There is a role for a special effort to expand the network 15

Entrepreneurial Exits n At a few points there was tension: Ader “drops out” in 1891 q Langley keeps secret wing design after 1901. (Chanute shares it anyway. ) q Wrights stop sharing as much in late 1902 After some perceived of breakthrough Jobs and Wozniak start Apple q n n q n they hire Homebrew club people as employees Red Hat becomes a company 16

Example exit: Clement Ader’s Eole n n n It traveled 50 meters in uncontrolled flight in 1891 French military thought it would be useful. Ader didn’t patent outside France because it would expose details. Chanute criticized this choice. Ader “drops out” from prior communication links. 17

Entrepreneurial exits from network If a tinkerer has an insight into how to make a profitable product it may be worth leaving the network n conducts directed R&D n becomes an entrepreneur n enters economic statistics 18

Conclusion n This process may be important q q n n I do not know of other models of it Key assumptions: q q n n explaining the rise of industrial countries a long time ago with open source software, now technological uncertainty (no clear product and market) motivated tinkerers some way to make progress some way to network Search and matching costs take some more thinking An industry can spring out of this, not well modeled yet 19

Experimenters and Octave Chanute takes interest in flying machines n Wealthy former engineer in Chicago n Ran experiments of his own on gliders n Described previous work in 1894 book Progress in Flying Machines. q discusses a hundred individuals, from many countries, professions q and many experiments, devices, theories q helps define “flying machines” work, focused on kites q book supports network of information and interested people Chanute corresponded actively with many experimenters. Chanute preferred that everyone’s findings be open. 20

What are they making? Aeronautical journals appear in 1870 s and 1880 s. Experimenters make diverse choices. Available metaphors: n Balloons are light, ascends without power q n n Meteorological balloons, hot-air, helium-filled balloons Rockets are high-powered, rigid, hard to control Kites and gliders (light; fixed wings generating lift) q q q For lift (upward force), requires speed. Propulsion? n Flapping wings? Birds are light and have big wings n Propellers? n Jets? Power? muscles, steam engines, internal combustion engines, in models, wound up rubber bands 21 Hard to control

Lilienthal’s Wings and Gliders n n n German engineer Otto Lilienthal studied birds and lift shapes in wind 20 years of experiments, often with brother Gustav Wrote book Birdflight as the Basis of Aviation. Made hang gliders Flew 2000+ times Became famous and an inspirational figure 22

Timelines Red Hat company 1971 Intel 1975 micropro Altair kit cessor 1976 Apple I Novell & Su. SE combine IBM commits to Linux Real businesses: Apple, 1976, many others from Homebrew club; Microsoft 1977, IBM PC 1981 Open-source software … industries? Personal computer and software industry Homebrew Computer club 1804 Cayley 1804 Cayley 1893 Chanute book Progress in Flying Machines Aircraft industry Aeronautical journals Chanute networks (by mail, travels) 23

24

Hargrave’s Wings, Kites, and Engines n n Retired in Sydney Ran many creative diverse experiments starting in 1884 q q q n n n Drawn to flapping-wings designs Also made innovative engines Box kites showed layered wings were stable and had lift Often made small models or designs without building. Devices often did not work right the first time but he moved on to new inspirations. Did not patent. Published hundreds of findings Chanute: “If there be one man. . who deserves to succeed in flying through the air” – it is Hargrave. 25

Samuel Langley's technology choices Professor in Pittsburgh, then Director of Smithsonian Institution in DC His 1896 powered gliders went over half a mile Decides that for safety: • aircraft must be intrinsically stable, and • pilot must sit up craft must be rigid and strong innovatively, makes strong frame from steel tubing much heavier than a glider; needs strong engine for lift So he gets the best engine made, to that time, for its weight. (Balzer-Manly engine) 26

Langley’s aerodrome n Resulting aircraft is heavy, expensive, housed with difficulty q q n n n n Steel materials Large wings Powerful engine Cost ~$50, 000 Hard landings; lands on water => can't try twice easily Operator is not too useful, like rocket, unlike glider Langley's demonstrations are big, sometimes public In key demonstrations in Oct & Dec 1903 it crashes early Editorials attack Embarrassed trustees asked him to stop research But it was designed like a modern passenger jet 27

Wilbur and Orville Wright n n n n No college degrees No interest in engineering/academic careers Ran bicycle shop in Dayton, Ohio, US Starting in 1899 read from Langley and Chanute Corresponded actively with Chanute Good tool makers and users. Have a workshop. Generally crafted each piece. Collaborated intensely. 28

Wrights' technology choices n n n n Focused on wing shape, propellers, and control mechanism Built craft as kites, then gliders Did not attach an engine until 1903. Materials light & cheap, wood & canvas pilot lays flat less drag intrinsically unstable, like a bicycle Pilot controlled that by hip movements which pulled wires to warp (twist) wing tips to turn glider This invented piloting skill had no future 29

Peter, do you need this formula? 30

31

Wrights’ wings and propellers n n n Wrights’ wind tunnel carefully tested to make air flow smooth Their balance device measured lift precisely They tested many wings systematically and came to an ideal design for their craft. What’s a propeller for an aircraft? Standard idea: like a water propeller, it would pushes air back. Having studied wings, Wrights’ experiment with propellers that have a cross section like a wing, with lift in forward direction 32

This evidence is selected n n n n Many other experimenters and publishers would be worth mentioning if time permitted: Alphonse Penaud Horatio Phillips Hiram Maxim James Means Alberto Santos-Dumont Richard Pearse Many others Paper has the beginnings of a list of what was available in the public domain. 33

34

35

Conclusion Airplane case makes plain certain aspects of these individuals and networks. It seems relevant to n personal computer hobbyists n open source software projects A model of this kind could be useful to describe or account for n engineering “skunkworks” in organizations n scientific advances n differences between societies in speed of technology development 36

Conclusion n n n n Why would individuals do this? Start manufacturing company Get revenues from patent Get hired as engineer Lerner and Tirole (2002, and repeatedly) Research funding (Langley, from War dept and Smithsonian) Prestige of accomplishment in contributing To grapple with interesting problems. Or, the concept is so cool! They want the problem solved -- that is, they want to live in a world in which they can fly through the air (that is, to change their world, not their place in it) "Our experiments have been conducted entirely at our own expense. At the beginning we had no thought of recovering what we were expending, which was not great. . . " Wrights, How We Invented the Airplane, [1953] p. 87 "I am an enthusiast, but not a crank in the sense that I have some pet theories as to the construction of a flying machine. I wish to avail myself of all that is already known and then if possible add my mite to help on the future worker who will attain final success. " -- Wilbur Wright, 1899 letter to Smithsonian Institution Other airplane; computer; open source people express this thought. Tentative formal assumption: Relevant individuals ("players") have utility functions that support this activity. - tentatively treat motivation of innovators as exogenous - testable implications of different utility functions? psychic joy of experimenting; or research salary; or imagined future payoff. 37

n n n Secrecy? Not usually Books by Lilienthal (1889) and Chanute (1894) Journal periodicals in France, Britain, US Wrights collected info from Smithsonian and Weather Bureau (location) Chanute actively corresponds with experimenters, researchers q n n n n technology moderator Wilbur’s speech to Society of Western Engineers, 1901 Journal publications in 1901 in England Germany Scientific American article about them in 1902. Visit of Spratt and Herring on tip back problem Langley gets secretive about his wing design Wrights get secretive starting late 1902 Modeling ideas: Sharing institution exists already Innovator chooses sharing vs. secrecy Players may be open (prestige; joy of sharing; desire for progress) Public pool of information is productive But if their device approaches some threshold (technical success or profitability), they close their connections to the network. (Homebrew and Apple example) This creates an industry. Then competition stimulates progress. 38

History founded Aeronautical Society of Great Britain Crouch, p. 30; Anderson p. 4 1866 F. H. Wenham highlight superiority of long narrow wings over short wide ones in generating lift (though this is sometimes forgotten, later) Anderson, p. 45 1868 Britain publication of Annual Reports of Aeronautical Society starts Crouch, p. 31 1868 Moy scale effects in aerodynamics established Anderson, p. 46 1869 Paris beginning of publication of L'Aeronaut Crouch, p. 31 1866 British engineers F. H. Wenham 1870 and John 1871 Browning developed wind tunnel Crouch, p. 31 39

1871 Wenham found that the center of pressure (center of lift) tended to be near the leading edge of a wing -- a fact sometimes forgotten, later Anderson, p. ? ? Anderson, pp. 35 -37 1871 Alphonse Penaud upward sloping tail, for stability; center of pressure. . ; understood it; had theory, created standard. 1871 du Temple powered hop in France 1875 Octave Chanute discovers, on trip to Europe, that European engineers treat airplane as possible Crouch, p. 26 Penaud cambered wing 1876. dihedral angle 2 degrees. Was on track to further success, but committed suicide P 37 Anderson 1883 Osborne Reynolds analysis of "laminar" (smooth) versus turbulent air flows Anderson, p. 44 1884 Horatio Phillips Analysis of wing shapes 1876 40

1884 Mozhaiski powered hop. Russia. 1888 France beginning of publication of the Revue de l'Aeronautique 1889 Lilienthal 1890 Clement Ader 1894 Jan Chanute 1894 Hiram Maxim 1896 Chanute & Herring Crouch, p. 31 published Birdflight as the Basis of Aviation Piloted, steam-engine-powered airplane, the Anderson, Eole; no controls; wings moved like a bat's p. 51 Publication of Progress in Flying Machines Stoff, p. iv Flying machine Anderson, p. 4 adapted Pratt truss to gliders Stoff, , p. iv 41

1896 Chanute/Her ring two-surface, double decker wings Jakab, 47; 5 4 -58 1896 Samuel Langley steam-powered unpiloted one minute flight over Potomac Anderson, p. 5 1897 Arnot / Herring Indiana gliders Crouch, p. 210 1898 Langley and others internal combustion gasoline engine determined to be superior to steam enginesfor lightweight power Anderson, p. 143 1899 Wrights wing warping for control of rolling motion Jakab, p. 54 1900 Mar Wrights wilbur wright contacts chanute; ww's already studied Progress Stoff, p. vi Wrights calculation of smeaton coefficientir wind tunnel, and wing tests, and lilienthal calculations Jakab, circa p. 130 1901 Oct 42

Motivation of the Experimenters: Why Would Individuals Do This? n n n To start manufacturing company To get revenues from patented technology To establish oneself professionally q n n (Lerner and Tirole, 2002) To earn research funding (Langley, from War and Smithsonian) To earn respect for their contribution To win a competition To grapple with interesting problems or solve them 43

Conclusion Collective Invention Model: Individuals are motivated by utility functions q. Sometimes unknown reasons for joining the network q. Discoveries are random q Key choice – share their findings or not? Octave Chanute and Samuel Langley – co-inventors of the Wright airplane or not? How much of the invention X is due to its inventor? 44

Secrecy: When Does it Start? n n Books by Lilienthal (1889) and Chanute (1894) Journal periodicals in France, Britain, US Wrights collected info from Smithsonian and Weather Bureau (location) Chanute actively corresponds with experimenters, researchers n Wilbur’s speech to Society of Western Engineers, 1901 n Publications in 1901 Visit of Spratt and Herring on tip back problem Langley gets secretive about his wing design Wrights get secretive starting late 1902 n 45

1901 Balzer & Manly high powered light engine (not known or available to Wrights) Anderson, p 144 1903 May Wrights blade-element propeller (50% more efficient than contemporaries; apparently highest recorded to that time). Anderson, p. 141 1903 Dec Langley / Manly Public demonstration of aerodrome; crashes before full flight 1903 Dec Wrights self-powered sustained flight; takeoff and landing at same level 1904 Wrights testing grounds on Huffman Prarie 46

End of Information Sharing n n If the activity succeeds, it becomes an industry – competitive “commercial production and sale of goods” The network loses importance, shrinks, breaks up Examples: q q n n Wrights in late 1902 clamp down; disagree with Chanute. Langley's wings Later: Apple computer Model assumption: Network will self-destruct if there is enough success Then industry players have private intellectual capital and don't share R&D. 47

48

Wrights’ Propeller n Propeller: q q “a mechanical device that rotates to push against air or water A machine for propelling an aircraft or boat, consisting of a power-driven shaft with radiating blades that are placed so as to thrust air or water in a desired direction when spinning. ” Wrights invented propellers that delivered 50% more pulling power from a given engine! 49