Thursday, August 27, 2015

[idea: product] NLP (natural language processing) application for life-restoring contents (e.g. the Bible)

idea: as soon as people type Bible related query, we get quick answers with some options (e.g. changing source version)

background: a plethora of people searching Bible texts everyday and cutting steps to find Bible text, one of the largely searched info (source needed), can increase access frequency and take out barriers to find the truth to be shaped by the truth so that we will live a life God originally intended us to live

UI: e.g. for Bible, answer text on top and changeable book and source info with version at the bottom 

  • saying or typing [john 3:16], [john 3 16] or [john three sixteen] on any platform (e.g. iOS, Android, Auto, Wear, Android TV, future virtual reality) will return the searched Scripture (with or without TTS)
    • TTS: e.g. [Romans 10 17], [Romans chapter 10 verse 17], [tell me Romans 10 17], [read Romans 10 17] triggers "Romans chapter 10 verse 17 Consequently, faith comes from hearing the message, and the message is heard through the word about Christ." 
  • changing book or source info will return relevant text
  • i18n and l10n: e.g. [요한복음 3장 16절] for Korean
  • default source version: displaying verse in previously selected version
  • query suffix/prefix : e.g. for [john 3 16 niv], display the version in users' typed version
  • finding book, chapter, verse: e.g. [where in the Bible is faith comes from hearing] or [faith comes from hearing] returns "Romans 10 17" 
tool: (link)

c.f. add gamification

      Tuesday, August 25, 2015

      [lecture: my note] 소프트웨어 테스팅 클래스 (translation: Software Testing Class) at Fast Campus

      Software Testing Class (link)
      Fast Campus Main Building 4th Fl.
      10:00 ~ 17:00 on August 25, 2015
      최준현 (Lindberg Choi)
      N/A (paid lecture: 150,000 KRW)
      QA basics, testing secret

      1. Lesson
        1. origin of software bug
          1. origin: vacuum tube
          2. term definition based on IEEE610.12-90
            1. Mistake: A human action that produces an incorrect result.
            2. Error: A difference...between a computed result and the correct result
            3. Fault/Defect/Bug: An incorrect step, process, or data definition in a computer program
            4. Failure: The [incorrect] result of a fault
          3. reasons for defect:
            1. engineers not complying with spec requirements
            2. product manager making mistakes in writing spec requirements
            3. engineers developing on their own due to absence of spec requirements
            4. engineers’ contrived interpretation in architecturing
            5. ...
          4. key principle:
            1. defect + regulatory violation => disabled service
        2. what is test process
          1. process: method to effectively increase task productivity
          2. software testing process: spec requirement analysis -> plan test plan -> create test case -> conduct test -> report result
        3. relationship between planning and testing
          1. deliverables of “planner”
            1. SRS (software requirement specification), IA (information architecture), business process, wireframe, SB (story board)
          2. quality management method for “planner” and product manager
            1. testing internal users
            2. reviewing proposal and development doc
            3. writing structured doc
            4. creating test case in the perspective of “planner”
          3. black-box testing method
            1. definition: Black-box testing is a method of software testing that examines the functionality of an application without peering into its internal structures or workings. <source: Wikipedia>
            2. example
              1. equivalence partitioning
              2. BVA (boundary value analysis)
              3. state transition testing
              4. decision table testing
              5. use case testing
              6. pairwise testing
        4. bug finding type per testing step
          1. understanding test cycle
            1. standard cycle: BAT -> main test -> ad-hoc test -> confirmation test -> regression test
            2. multiple test cycle: 1st (BAT -> main test -> ad-hoc test) -> 2nd (confirmation test -> main test -> ad-hoc test) -> 3rd (confirmation test -> main test -> ad-hoc test -> confirmation test -> regression test)
          2. bug finding type per testing step
            1. BAT (build/binary acceptance test): e.g. label differs per organizations: e.g. BAT at Samsung Electronics, dev to QA; dev review; hurdle at Nexon
            2. main test: conduct TC, finding function related bug
            3. ad-hoc test: exception test, negative action test (c.f. random: range is pre-determined; ad-hoc: testing under given time; monkey test)
            4. confirmation test: confirming bug fix, finding new bug due to unblocked functions
            5. regression test: testing previous feature due to bug fix; typically, new fix brings more bugs on previous feature
        5. bug finding principle
          1. principle of defect clustering
            1. principles of testing <source: link>
              1. Testing shows presence of defects
              2. Exhaustive testing is impossible
              3. Early testing
              4. Defect clustering
              5. Pesticide paradox
              6. Testing is context depending
              7. Absence-of-error fallacy
            2. most bug frequent area: upgraded/updated feature (e.g. At Google, engineers check file fix frequency in unit testing)
          2. magic of test data
            1. example: multimedia data by size and file types, member information, actual database data (e.g. 1234567890; 12345678901234567890, 일이삼사오육칠팔구십; ~!@#$%^&^&*((), null, 123456789 0, foreign language text, negative test case), physical data (e.g. OTP, credit card)
          3. targeting weak point
            1. reason: e.g. lack of connection feature to other modules; proposal issue; lack of state update test; batch function issue; incorrect data issue from admin and back office? (c.f. translation issue; number inconsistency issue)
          4. exception management
          5. targeting per product quality level
            1. many bugs at early development stage: focus on test case
            2. multiple test cycles: focus on unblocked feature
            3. stable quality: focus on UI/UX, compare similar feature from competitor, ask for testings to external user
          6. mindset of developer and tester
        6. tester secret

      1. Personal Takeaway
        1. types of bug: reliability, security, compatibility, usability, functionality, performance, language
        2. read
        3. UML (unified modeling language): ?
        4. app testing trends: e.g. testing with oscilloscope to check currents not to drain battery
        5. magic of test data
        6. secret: e.g. testing backdoor (e.g. CGV, Olleh, Auction sites for ticketing), leave the session
        7. establishing QA process at startups by StyleShare (link)

      Monday, August 10, 2015

      [lecture] 젊은 창업가가 들려주는 실리콘밸리 스타트업 경험기 (my translation: Silicon Valley Startup Story)

      Silicon Valley Startup Story
      D.CAMP 6th fl.
      August 10, 2015 (Mon) 19:30~21:30
      서승환 (Shaun Seo)
      Girls in Tech, Jobstick (link)
      general advice on how the speaker built his start-up

      lesson and takeaway: 
      1. what’s more important is how developers execute than a million dollar idea
      2. in selecting ideas to develop, try lean startup method
      3. combine marketing and product development together not separating the two