エンタープライズOSS 2006
エンタープライズOSS2006「マイケル・ティーマンが語る!オープンソース3つの視点/Q&A」というのに参加した。
http://www.seshop.com/event/eoss/
会社のブログに若干の蛇足を書いたのでそちらも参考にしてほしい。
以下はログ。
Software Industry History 1946: The First Computer 1961-1965: IBM OS 360 1975: Fred Brooks Today - $500B HW/SW - $500B people and support Money No Match for Security Design Flawas $100M MS $200M $400M $800M Software Industry Not Alone Education - treats teachers like factory workers/students raw materials - bureaucracy - policy not science Agriculture - old: harmony, sustainable - now: one kind of food Manufacturing - Deming: The Courage to See The Crisis - low quality --> bad industry model - loyalty and trust - improve people withe training - transformation is the everybody's job As We May Think 1947: Vannervar Bush manifesto for World Wide Web 1984: France:MiniTel 1990: World Wide Web User Innovation Toolkits (von Hippel, 2002) - 85% of all breakthorough innovation is user-driven - Proprietary innovation strategies stifle 5/6 of all innovation - some innovations (like the web) impossible with proprietary models old ad/dev, design, build, test ---------------------- ---- supplier customer new ad/dev, design, build, test ------- ------------------ supplier customer democrotize innovation Unleashing Innovation - sourceforge.net 1.2M users, 100K projects - 2002 and 2005 FLOSS surveys of OSS developers - 490K OSS developers spend 11+ hour/week on OSS - MORE THAN 5 Million HOURS PER WEEK - it's fun - it improves their skills - it's good for society - Microsoft employs 61,000 and spend $6B on R&D - 80 hours/week * 61,000 < 5M hours/week - $6B/Year @ $25/hour not enough to afford 5M hours/week - $2B fixing security bugs Software Quality -- What Would Deming Do? - Typical software defect density is 20-30 defects per 1000 lines of code - 2004 -- Linux Kernel is 5.7 MLOC - should 100,000 defects - only found 985 found - 2005 -- 100% of serious bugs fixed in first 6 months - complexity increased 4.7% - density decreased 2.2% - end of mythical man month - 2006 -- 32 other OSS projects measured - worst (PHP) was still 50x better than typical (or 98% fewer bugs per LOC) Open Source is Changing The World - AIDS/ secrecy and competition Improving Security by Making it Free - Deming: Security is everybody's job New and Changed System Call Q&A topics - OSS Merit and Risk - OSS Project Management - Open Standards & Open Source - About MT and MT in the future OSS Risk is ... - transparent development or broadcasting the vulnerability -- code review is the best practice -- not ideal situation -- there is no hide -- thouthands reviewers OSS -- Can they develop/create the software - money > off shore - compete added value > best; compeate cost > worst - innovation > creation value more smart people out side OSS model optimise boundary - control --> power lose/get RH strategy $1B cach in Bank RH can not change the world RH meets the market/revolutional mind capitalism Open Standards - Open Standard Requirement - The Criteria - No Secrets - Availability - Patents - No Agreements - No OSR-incompatible Q&A session - RH does not accept patch etc from Ikeda san's question RH partner / open source development best way to find it propriety model service improvements < open source model Open Source GIS OSGES.org R-project.org 2,500 pages GRASS project brender.org 3D About MT blog
マイケル・ティーマン氏のプレゼンの後に岡田さんが質問するという形式だったのだが、フロアからも2つ質問する機会があったので、ずーずーしくも質問した。