modelling t he cont ribut ion st ruct ure underlying
play

Modelling t he Cont ribut ion St ruct ure Underlying Requirement s - PDF document

1 Modelling t he Cont ribut ion St ruct ure Underlying Requirement s Orlena Got el & Ant hony Finkelst ein NFR memo RSD n.8 v.3 v. 66 REFSQ 94 2 To Avoid I nit ial Quest ions (1)... Requirement s Traceabilit y (RT) ref ers t


  1. 1 Modelling t he Cont ribut ion St ruct ure Underlying Requirement s Orlena Got el & Ant hony Finkelst ein NFR memo RSD n.8 v.3 v. 66 REFSQ ‘94 2 To Avoid I nit ial Quest ions (1)... “ Requirement s Traceabilit y (RT) ref ers t o t he abilit y t o describe & f ollow t he lif e of a requirement in bot h a f orwards & backwards direct ion” (i.e., From it s origins, t hrough it s development & specif icat ion, t o it s subsequent deployment & use, & t hrough all periods of on-going ref inement & it erat ion in any of t hese phases) REFSQ ‘94 Page 1

  2. 3 To Avoid I nit ial Quest ions (2)... “Post -RS Traceabilit y is concerned wit h t hose aspect s of a requirement ’s lif e t hat result f rom it s inclusion in t he RS” (i.e., Requirement deployment ) “Pre-RS Traceabilit y is concerned wit h t hose aspect s of a requirement ’s lif e prior t o it s inclusion in t he RS” (i.e., Requirement product ion) REFSQ ‘94 4 To Avoid I nit ial Quest ions (3)... “Sof t ware Qualit y (SQ) is t he: - Tot alit y of f eat ures & charact erist ics of a sw product t hat bears on it s abilit y t o sat isf y given needs - Degree t o which sw possesses a desired combinat ion of at t ribut es - Degree t o which a cust omer or user perceives t hat sw meet s his/ her composit e expect at ions - Composit e charact erist ics of sw t hat det ermine t he degree t o which t he sw in use will meet t he expect at ions of t he cust omer” REFSQ ‘94 Page 2

  3. 5 An Overview... Relat ion bet ween RT & SQ: - How post -RS t raceabilit y can improve SQ - How pre-RS t raceabilit y can improve SQ Requirement s Cont ribut ion St ruct ure: - What it is - Pot ent ial f or qualit y improvement Workshop quest ions (sort of ) answered REFSQ ‘94 6 Relat ion Bet ween RT & SQ... Reqs Qualit y-orient ed development : - Specif y reqs - Use t o drive, cont rol & evaluat e process Process t race Qualit y assurance/ conf ormance checks: - Meet s user needs & adheres t o qualit y at t ribut es - Support ed by met hods/ t echniques/ paradigms/ pract ices - Approaches depend on some f orm of RT - RT is primary qualit y-enabling t echnique SQ inf luenced by t echniques/ t ools used f or RT REFSQ ‘94 Page 3

  4. 7 Post -RS Traceabilit y & SQ (+)... reqs art if act s produced (relat ed t o reqs deployment process) RSD s0 s1 sn Qualit y reqs Can check qualit y Ef f ect s of changing can permeat e conf ormance qualit y reqs can be development at each phase analysed/ propagat ed REFSQ ‘94 8 Post -RS Traceabilit y & SQ (-)... Why does sw st ill f ail t o meet ant icipat ed levels of SQ? Def n SQ support ed - “ meet ing t he RSD” - no guarant ee Problems: User sat isf act ion Def n SQ/ met rics subj ect ive/ collect ive change/ const ruct ed / f ew met rics downst ream Reasons: Deals wit h rest rict ed Poor f oundat ion t o phases of a reqs lif e achieve/ assess SQ REFSQ ‘94 Page 4

  5. 9 Pre-RS Traceabilit y & SQ (+)... reqs art if act s produced (relat ed t o reqs product ion process) RSD changes Qualit y-orient ed Foundat ion Change f rom source & re-propagat e approaches built t o achieve & t hrough ALL phases in f rom onset assess SQ REFSQ ‘94 10 Pre-RS Traceabilit y & SQ (-)... 2 aspect s of QA act ivit ies Document at ion-relat ed Organisat ion-relat ed I nf o about RE process I dent if y t hose in a document ed/ st ruct ured/ posit ion t o assess maint ained/ navigable qualit y/ address def ect s Done out End-product Done by Loss of of cont ext cent red ot her part ies commit ment Subj ect of much work Basic working pract ice I mpact on SQ not account ed f or REFSQ ‘94 Page 5

  6. 11 Traceabilit y of Requirement s Cont ribut ors & Cont ribut ions... Model social inf rast ruct ure of RE process - as qualit y: Socially def ined Socially evaluat ed Socially accept ed I I can ident if y can ident if y sources of sources of reqs-relat ed reqs-relat ed work inf o We can locat e We can personnel ident if y sources of reqs I can t race personnel REFSQ ‘94 12 Basic Concept s Our Work (1)... Model Cont ribut ion St ruct ure underlying reqs art if act s Augment wit h agent det ails: - Schema of analyt ical capacit ies in which cont ribut e - Def ined t hrough int eract ive mark-up Development of CS: - I nf er f iner-grained social capacit ies & commit ment s - I mpact of t axonomy of int ra/ int er-art if act relat ions Requirement s: - Ongoing def init ion & re-def init ion - Evolvable & emergent - I nf erence & int errogat ion REFSQ ‘94 Page 6

  7. 13 General Approach Our Work (2)... (A) Def ine cont ribut ion f ormat of art if act Principal Cont ribut ed_by Art if act Cont ribut ed_t o Aut hor Document or P: agent whose posit ion or belief is est ablished by t he inf ormat ion t he art if act expresses A: agent responsible f or choosing & organising t he cont ent & st ruct ure of t he art if act D: agent responsible f or physical manif est at ion of t he art if act (i.e., capt ures or records it ) REFSQ ‘94 14 General Approach Our Work (3)... Cont ribut ion f ormat of art if act & int ernal component s Collect ion of PAD agent s: P A Artifact Art if act D P A D P A D P A D REFSQ ‘94 Page 7

  8. 14 P A Art if act D P A D P A D P A D REFSQ ‘94 15 General Approach Our Work (4)... (B) I nf er cont ribut ion roles of agent s wrt art if act Deviser Ghost Aut hor Sponsor P A Represent at ive Nominal Aut hor Aut hor Relayer D True Aut hor REFSQ ‘94 Page 8

  9. 16 General Approach Our Work (5)... Cont ribut ion roles of agent s 1 Examples: } P = Olly Olly is: A = Olly True Aut hor 2 D = Olly Olly is: } P = Olly Deviser 3 A = Olly Dave is: Olly is: D = Dave addy } P = Olly Relayer Sponsor A = P addy Paddy is: D = P Represent at ive REFSQ ‘94 17 General Approach Our Work (6)... (C) I nf er agent commit ment t o art if act Who t o involve/ inf orm in queries/ changes: � � True P hysical appearance Ant icipat ed/ realised ef f ect St ruct ural f orm Semant ic cont ent � � Aut hor: � P hysical appearance Ant icipat ed/ realised ef f ect � Deviser: St ruct ural f orm Semant ic cont ent � � � � P hysical appearance Ant icipat ed/ realised ef f ect Relayer: St ruct ural f orm Semant ic cont ent � � REFSQ ‘94 Page 9

  10. 18 Det ails More Our Work (7)... Qualif iers (or modes) f or cont ribut ion f ormat I nf er more about t ypes/ degrees of commit ment : - Principal st at us (signat ory-relat ed) approved/ pending/ not approved - Aut horial st at us (relat ed t o int er/ int ra art if act links) primary - creat or n-ary - adopt er/ adapt er/ ref erencer doc - Document or st at us (mood-relat ed) doc assumpt ive/ emphat ic xyz xy quot at ive/ report ive indef init e/ quest ioning REFSQ ‘94 19 Direct ions Current Our Work (8)... Markup t o overlay cont ribut ion f ormat , et c. Support ing CS evolut ion & maint enance: - Augment art if act -based t raceabilit y t o deal wit h associat ed CS at each st ep - Rules t o recomput e & updat e CS Meet ing t raceabilit y reqs: - Use of layers f or select ive t races - Priorit y st ruct s f or change handling Tool t o evaluat e & ref ine approach REFSQ ‘94 Page 10

  11. 20 Summarising Qualit y I mplicat ions... Qualit y depends on post -RS & pre-RS Traceabilit y Furt her leverage if t ie people int o t he t raceabilit y equat ion Modelling CS provides t raceabilit y of cont ribut ors & t heir cont ribut ions: - Fine/ rapid locat ion of personnel t o perf orm QA act ivit ies, address qualit y def ect s, et c. - Cont inuous def n, redef n, assessment & re-assessment of SQ t hroughout a proj ect ’s lif e REFSQ ‘94 q1 Quest ions (1)... Q: What is t he “right ” dat a t o be made t raceable? Traceabilit y & I n what way Of what (access & present at ional (inf ormat ional requirement s) requirement s) Depends on Who want s it Why t hey want it (user charact erist ics) (t ask charact erist ics) When t hey want it (proj ect & cont ext ual charact erist ics) REFSQ ‘94 Page 11

  12. a1 Answers (1)... A: No such t hing as “right ” dat a - only required dat a More pert inent quest ions we should ask: t oo much!!! (1) How can we ident if y required dat a? - More empirical st udies wit h proj ect -specif ic select ion (11) How can we deal wit h absence of required dat a? - Provide pot ent ial f or inf ormal communicat ion REFSQ ‘94 q2 Quest ions (2)... Q: How should t race dat a be recorded? - Manually - Aut omat ically Q: Who/ what should record t race dat a? - Those involved/ - Technology ot hers/ f orms Q: When should t race dat a be recorded? - During - Lat er process reviews REFSQ ‘94 Page 12

Recommend


More recommend