Download Introducing Reuse in Companies: A Survey of European Experiences

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Introducing Reuse in Companies: A
Survey of European Experiences
Maurizio Morisio, Michel Ezran,
and Colin Tully
Overview
• Presents initial results of the Surprise project
• The Surprise project aimed at identifying key
factors in adopting or running a company wide
reuse program
• Key factors derived from empirical evidence of
best reuse practices
• More than 20 European performed 94-97 were
analysed
Introduction
• Systematic reuse is generally considered a key
technology to improve the software process
• OOT is seen as an essential enabler by many,
others argue that OOT alone does not guarantee
reuse
• Factual evidence is contradictory
• The basic idea of the Surprise project is to derive
key factor for reuse success or failure form
industrial reuse projects
Research approach
• Identification of projects
– specifically ESSI Process Improvement Experiments (PIEs)
– most of the projects discarded were only introducing OOT and
claiming reuse
• Development of the questionnaire
– several iterations and testings
• Reading of reports and interview
• Validation of the interview report
– interview report validated by the interviewee
• Consistency check
– attribute values synthesized were agreed upon by at least two team
members
• Data analysis
– extracting findings and trends
The questionnaire
• General information
– about the company
• Non reuse information
– software process, etc.
• Reuse information, non technical
– motivation for reuse at the company level, rationales, expectations,
goals
• Reuse information, technical
– categorisation of the reuse approach, type of assets used, reuse
roles and processes added, repository, metrics, cost models
• Qualitative and quantitative PIE results
Data from the analysis
• Table 1 (state variable)
• Table 2 (control variables, dependent variables)
• Success is defined by:
– continuation of the reuse project after the PIE is over
– on the soundness of the approach
– if assets actually are reused
• Not possible to define return of investment as a
measure of success (only one company performed
an economic analysis)
Analysis (state variables)
• Software staff
– successful projects belong to all categories
– un successful projects belong to all categories
– size should impact on the choices of control variables
• Software production
– most companies in the product-family
– both success and failure suggest that a product family oriented
production is not a success factor
• Software process maturity
– majority of companies with a good maturity level
– process maturity appears to be useful but not sufficient
Analysis (state variables)
• Application domain, software type
– are not a factor of success or failure
– two cases of embedded software was related to failure
• Size of baseline, language, analysis and design
– diversity of value suggest no factor of success of failure
– success and failure happens for both object oriented and
procedural
– object orientation equals reuse could be a failure factor
Analysis (control variables)
• Commitment of top management
– no top commitment projects failed
– both failure and success for projects with top
commitment
– commitment from the leader is a key success factor (but
not alone)
Analysis (control variables)
• Reuse technical approach
– Tight approach
• frameworks, product baseline, composition language
• all projects with tight approach were successful
• the real key factors could be: long term commitment and
investment, skilled architects and domain experts
– Loose approach
•
•
•
•
no common architecture defined
both success and failure
failure due to wrong focus (e.g. aut. repository)
may be caused by less commitment
Analysis (control variables)
• Rewarding, training, awareness
– most companies did not use reward policy, not a key factor
– training and awareness are key factors
• Repository
– some companies are successful without dedicated repository,
focusing on repository but overlooking other factors caused failure
– not a key factor
• Number of reusable assets
– not a key factor for success
– hypothesis: limited number can be a condition for success
Conclusion
• Reuse can happen in a variety of ways, in a variety
of companies
• Key factors appearing with more frequency in
success stories
– commitment of management
– training and awareness actions
– tightly coupled approach appears to work best