1. 1 SAIV/CAIV/SCQAIV LiGuo Huang USC University of Southern California Center for Software Engineering CSE USC

2. 2 Outline Why SAIV/CAIV/SCQAIV? The SAIV Process Model Fulltext Title Database –A SAIV Case Study The CAIV/SCQAIV Process Model University of Southern California Center for Software Engineering CSE USC

3. 3 Is This A Problem? The clients propose cost, schedule, or quality constraints — Want the system on a ridiculously short schedule And they aren’t even willing to commit on just exactly what they want University of Southern California Center for Software Engineering CSE USC

4. 4 Is This A Problem? Yes, when you accept the schedule target before understanding the requirements – Retrofit the required capabilities in an ad-hoc way – Reduce the scope or dropping features – Success is more a matter of luck and heroic efforts –Usually, this leads to a Death March No, if you use the SAIV process model University of Southern California Center for Software Engineering CSE USC

5. 5 The SAIV Process Model “If schedule is your independent variable, then just modulate your functionality to meet schedule.” 1. Shared vision and expectations management 2. Feature prioritization 3. Schedule range estimation 4. Core capabilities determination 5. Architecture Flexibility Determination 6. Incremental development 7. Change and progress monitoring and control University of Southern California Center for Software Engineering CSE USC

6. 6 Core Capability Determination Core capability not just top-priority features – Coherent and workable end-to-end operational capability –Facilitate evolution to full operational capability –Architected for ease of adding, dropping marginal features University of Southern California Center for Software Engineering CSE USC

7. 7 Fulltext Title Database – A SAIV Case Study University of Southern California Center for Software Engineering CSE USC Fixed 24-week schedule 12 prioritized features

8. 8 University of Southern California Center for Software Engineering CSE USC Prioritized Product Features PriorityFeatures Very highF1Search and locate full-text journal titles by title keywords Very highF2Search and locate full-text journal titles by title keywords and date, title keywords and volume or number, title keywords with any combination of the other three attributes Very highF3Provide hyperlinks to vendors’ databases in the searching results Very highF4Update Fulltext Title Database using current vendors’ title lists HighF5Automatically FTP downloaded title lists from administrator’s local machine to remote server HighF6System administrator authentication HighF7Administrator password maintenance MediumF8Add new vendor’s title list profile MediumF9Delete existing vendor’s title list profile MediumF10Modify existing vendor’s title list profile MediumF11View existing vendor’s title list profile LowF12Allow more searching options starting with searching by ISSN, etc.

9. 9 Core Capability Determination Four capabilities of the Fulltext Title Database system — Provide a full-text journal title search capability(High) — Update the Fulltext Title Database (High) — Administrator password maintenance (High) — Accommodate vendors’ title lists in various formats (Medium) The first three form a coherent core capability! University of Southern California Center for Software Engineering CSE USC

10. 10 COCOMO II Analysis With 90% confidence, 12 features — 40 weeks 7 features — 24 weeks University of Southern California Center for Software Engineering CSE USC

11. 11 Architecture Choices University of Southern California Center for Software Engineering CSE USC F1, F2, F3F4, F5 F6, F7 F8, F9, F10, F11 F12 Intelligent Router … Rigid Hyper-flexible Flexible

12. 12 Architecture Flexibility Determination University of Southern California Center for Software Engineering CSE USC

13. 13 Architecture Flexibility Determination (cont.) University of Southern California Center for Software Engineering CSE USC E – total change effort N – total number of features (N =12) m – modularity factor (m = 2.74)  – a fraction of anticipated changes (  = 5/12) c 1 – average effort to develop a feature c 2 – average development effort of per inter-module interaction c 3 – average development effort of per intra-module feature interaction c 4 – average effort to accommodate a new feature with each existing feature in a module

14. 14 Incremental Development, and Coping with Rapid Change Initial Operational Capability Iteration I (70~80% of the schedule) — Provide a full-text journal title search capability — Update the Fulltext Title Database — Administrator password maintenance Initial Operational Capability Iteration II — Accommodate vendors’ title lists in various formats — Negotiated feature demotion  Add F13: Partial keyword searching  Drop F12: Allow more searching options starting with searching by ISSN, etc. University of Southern California Center for Software Engineering CSE USC

15. 15 Revised Architecture University of Southern California Center for Software Engineering CSE USC Intelligent Router … Rigid Hyper-flexible Flexible F1, F2, F3F4, F5 F6, F7 F8, F9, F10, F11 F13 F12 ? F13 F13

16. 16 Conclusions: SAIV Critical Success Factors Working with stakeholders in advance to achieve a shared product vision and realistic expectations; Getting clients to develop and maintain prioritized requirements; Scoping the core capability to fit within the high- payoff segment of the application’s production function for the given schedule; Architecting the system for ease of adding and dropping features in an appropriate level of flexibility; Disciplined progress monitoring and corrective action to counter schedule threats University of Southern California Center for Software Engineering CSE USC

17. 17 CAIV and SCQAIV SAIV process model also works for Cost as Independent Variable SCQAIV model is a straightforward extension of CAIV and SAIV –And “Cost, Schedule, Quality: Pick Any Two?” University of Southern California Center for Software Engineering CSE USC

18. 18 University of Southern California Center for Software Engineering CSE USC Cost, Schedule, Quality: Pick any Two? Consider C, S, Q as Independent Variable – Feature Set as Dependent Variable C Q C QS “Cost, Schedule, Quality: Pick All Three”