The case for open source can be based entirely on technical and economic factors. The Open Source process of development produces code that is characterized by: - better quality
- higher reliability
- lower costs of production, and,
- provides increased choice
Also, for hardware vendors - Nobody wants to wait six months for bug fixes -- if a user wants to fix a problem and has the ability to do it, let them
- Drivers do not bring in revenue -- let someone else do your work for you; concentrate on innovation
- OS will extend the lifetime of hardware beyond the point where it is economicall viable for the hardware vendor to support it
- Closing source will give the worst of all worlds: your secrets will be exposed (users are smart enough), you won't get free help from users, and you won't leave your competition behind (because they are forced to innovate, they will).
Ultimately, these are the driving factors in any open society with a laissez fair economy. Politics does not have to play any role in the decision making. Use value and Sale value - Use value of software: productivity multiplier (intermediate good)
- Sale value: as a salable commodity
The factory model of software production: - Developer time paid by sale value (cost plus margin)
- Sale value is proportional to development cost.
Coding practice - 95% of code written in-house; only 5% sale value driven
- Difficult to re-use -- customized to user organization
- Maintenance work required, and takes up > 75% of a developer's time
- Has little or no sale value, only use value
- Price a user will pay for software determined by expected future value of vendor service
- Software is a service industry under the delusion it is a manufacturing industry
- Price structures incongruent with actual development costs (because 75% of project cost is maintenance -- debugging and extensions)
- Incentives to produce low quality code
Closed Source - maximize buyers, minimize users --> rewards vendors of shelfware.
- maintenance cost paid for by future customers of software
- Only escape in long run is to have no competitors: winner takes all
Solution - Pricing structure should be based on service contracts, subscription, and a continuing exchange of value between vendor and customer
- Effect of making sofware "free" is to turn it into a service-fee based biz
- Consequently, demand for programmers goes up, not down.
Myth of information wanting to be free - "Tragedy of the Commons
- Inverse commons: Using software increases rather than decreases its value (fixes, code patches, extensions)
- No free-loading because people need solutions NOW
- User benefits by passing on patch rather than hoarding it (future maintenance is assured)
- Having more free-riders costs nothing
Reasons for Closing Source - To sell code to others (but this is only 5% of code ever written)
- Deny use to competitors (competitive advantage) -- but, sunk costs, bad design if code reveals business process rather than application
- To make code secure from hackers -- this is a delusion
Use value funding models - Use value is NOT Threatened by OSing; only Sale Value is
- Use value GROWS with use
- Examples:
- Apache: cost sharing, cooperative funding by webmasters -- alternatives were to buy proprietary software, roll your own, or OS. Netcraft: http://www.netcraft.com/survey
- Cisco: Risk spreading; ensure extended life of product beyond original developers
Why sale value is problematic in OpSo - There are no restrictions on use, redistribution, modification
- No vendor is in a privileged position to extract profits (symmetry; pure market vs. hierarchy)
- Unintended consequences of restrictions -- better to Keep Licensing Simple
- Preservation of gift culture
Indirect sale value models - Loss leader: give one away to sell products (e.g., Internet Exploder)
- Widget frosting: e.g., drivers -- there is no money in it anyway
- Give away recipe, open restaurant: services
- Accessorizing: e.g., sell documentation, add-ons
- Free the future, sell the present: open source it at a fixed point in the future
- Free the SW, sell the brand: e.g., StarOffice
- Free the SW, sell the content: e.g., subscription service, MSN
When to Open, when to Close source - OS -- Peer Review, CS -- rent from bits
- OS is only scalable method for high quality peer review -- The Silver Bullet
- Network effects desired (e.g., email, TCP/IP)
- Customer gains -- there is no Lock In
- Secrecy is enemy of quality
- OS SW is future-proof
- OS is high payoff when:
- Quality and reliability is paramount
- Peer review is only means of ensuring it
- Software is a biz-critical capital good
- SW establishes common comm and computing infrastructure
- Key methods and algorithms are common knowledge
- CS makes sense in vertical markets, narrow, non-critical segments or apps with specific, proprietary methods (biz secrets)
Knowing when to let go - DOOM
- OS as stretegic weapon: Lower cost, future proofed
- Resetting competition -- DEC vs. X-windows
- Growing the pond -- Red Hat and RPM installer
- Preventing a chokehold -- Mozilla www.mozilla.org
- OS and Strategic business risk -- CS is risky
- Biz ecology of OS: developers, distributors, users; loose coupling (no one is indispensible); no tar pit, no deadlines
Coping with success - Licenses ensure that code is shared
- no monopoly develops
- Distributors compete on service
- Even Red Hat
Open R & D -- the reinvention of patronage - Developers get paid for what they would do for free
- Why fund research avbl to others for free: dominance helps
- Buying goodwill (as in academia)
Open Source movement moving along - Task markets for funding OS development:
- Net surveys:
Suitable areas for OS - Apps (least) -- MS Powerpoint
- Middleware -- database management systems (e.g., Oracle -- CS, and MySQL -- OS)
- Infrastructure (most) -- Internet, Web, Operating systems, development tools
In the future, a program competing with IS will either die or become a part of the infrastructure |
