opim316 lectures 24 project review

OPIM 316/661Systems Analysis, Design, and ImplementationClass Web Pages: http://opim.wharton.upenn.edu/~opim316/spring04/Class WebCafe: http://webcafe.wharton.upenn.edu/eRoom/opim/661-sp04-1 : 

OPIM 316/661 Systems Analysis, Design, and Implementation Class Web Pages: http://opim.wharton.upenn.edu/~opim316/spring04/ Class WebCafe: http://webcafe.wharton.upenn.edu/eRoom/opim/661-sp04-1 Lecture 24 Project Review Alan Abrahams Room 557, Jon M Huntsman Hall, 3730 Walnut Street, Phila., PA 19104 Telephone: 215.573.5046 (W) Email: asa28[[at]]wharton.upenn.edu 1

Outline: 

Outline 2 Assessing goals achieved and missed Why software projects fail Learning and improving: Software Capability Maturity Model Discussion of problems you personally encountered Maintenance Phase: Project Review

Where are we in our Systems Architecture Framework in this lecture?: 

Where are we in our Systems Architecture Framework in this lecture? 3 Bus. Case Analysis Design Build Assess Deploy Maintain STORE WHAT DO WHAT WHO WHERE WHEN WHY HOW MUCH WHAT IF Project Review

Goals Achieved and Missed: 

Goals Achieved and Missed 4 Business goals achieved and missed (costs-down, profit or market share up?) Project goals achieved and missed What were we good at and how can be repeat it ? (Standards) What were we bad at and how can we avoid doing it badly next time ? Functionality implemented and omitted Assess performance of individual staff (goals achieved and missed) and: reward retrain, or retrench Actuals vs estimates: Times Costs (money and other resources) Determine why estimates were wrong, and improve for next time. e.g. did we consistently underestimate for a certain type of technology, programmer, or task ? How do we make better estimates for next time ?

Why Software Projects Fail: 

Why Software Projects Fail 5 Incorrect or fuzzy requirements definition Bad estimates: Cost overruns, schedule delays Skills deficit: untrained project managers, analysts, programmers, … Bad implementation quality: design or programming errors Collapse under complexity: as new components are added links between components multiple exponentially. Difficulties in deployment: e.g. Software Distribution: rolling out dynamic link libraries (DLLs) can be hard; hence the move to zero footprint browser-based solutions (nothing to install) Timing: Hershey’s chocolate deployed their SAP system in a big-bang approach just before Halloween and reportedly lost millions when the system went down, as this is one of their peak times of year.

Why Software Projects Fail: 

Why Software Projects Fail 6 Unclean data: garbage in, garbage out Implementing before design stabilizes: systems are like cement – soft during design are being, but hard and brittle once set in code ! Staff redundancy, presence of consultants, new business processes causes internal organizational unrest. Lack of management or user support and involvement in system design and in use (operation) Poor management of user expectations Company Politics

Why Software Projects Fail: 

Why Software Projects Fail 7 External risks materialized: Market risk: Customer adoption too low or too high (‘victim of own success’) Physical Environmental risk: fire, flood, earthquake Legal and Political Risk: e.g. change in tax laws Technology risk: e.g. software becomes obsolete or unsupported, or promised patches delayed Vendor risk: e.g. vendor goes insolvent or is bought, products phased out Terrorism: Sep 11th destruction of systems, and overload on CNN.com causing collapse Closed upgrade paths: If you bolt-on a custom-developed module to a standard COTS software product it’s not easy to upgrade to the next release of the COTS product, because your bolt-on needs to be updated separately. Possible solution: SAP co-developed a custom footwear module with Nike (saving development cost); Nike allowed SAP to include functionality in the next standard release, subject to a ‘wait period’ which gave Nike a head start on the competition.

Learning and Improving:The Software Capability Maturity Model (CMM): 

Learning and Improving: The Software Capability Maturity Model (CMM) 8 From Carnegie Mellon’s Software Engineering Institute (CM-SEI) Source: Adapted from http://www.sei.cmu.edu/cmmi/presentations/euro-sepg-tutorial/sld052.htm Structure & Productivity Novelty & Risk

Improving Schedule & Budget Estimates:The Software Capability Maturity Model (CMM): 

Improving Schedule & Budget Estimates: The Software Capability Maturity Model (CMM) 9 Source: http://www.sei.cmu.edu/cmmi/presentations/euro-sepg-tutorial/sld022.htm

Improving Software Productivity and Quality:The Software Capability Maturity Model (CMM): 

Improving Software Productivity and Quality: The Software Capability Maturity Model (CMM) 10 Source: http://www.sei.cmu.edu/cmmi/presentations/euro-sepg-tutorial/sld025.htm

Discussion: 

Discussion 11 What problems did you encounter which affected the success of your projects, (and how should project grading be adjusted to account for these mitigating factors) ?