An Experience of Technologies Diffusion

1
An Experience of Technologies Diffusion

• Prof. Giuseppe Visaggio
• Department of Computer Science
• University of Bari
• visaggio_at_di.uniba.it

2
SETTING

• The research and experience packaging were
  carried out in Software Engineering Research
  Laboratory (SER-Lab)
• Domains dealt with
• renewal processes for geriatric software
• monitoring of software decay

3
Transfer Strategy ...

• Commitment. New technologies and methods are
  found efficacious for a risky situation.They are
  introduced, either in a pilot or a challenging
  project, and validated by experimentation.

4
... Transfer Strategy

• Packaging and dissemination.The innovative
  technologies are packaged, together with
  empirically-gained experience.The best practices
  are disseminated to other customers in the same
  or other organizations. New experience can then
  be included in the package to enhance its reuse
  potential in future projects.

5
Projects ...

• Renewal of a working Software System in a Bank in
  Southern Italy (APPBANK)
• Partial migration of a software system belonging
  to a Crédit Agricole (CHARTRE)

6
... Projects

• Analysis of a working package in a manufacturing
  company in Southern Italy (FILANTO)
• Feasibility study of renewal of a working
  software system in a French insurance company
  (RCV)

7
APPBANK experience packaging ...

• Metric plan for maintenance process evaluation
  (full)
• Reverse engineering process (partial).
• Restoration process (partial).

8
APPBANK experience packaging ...

• Templates for application architectures with a
  good level of information hiding (full)
• Reengineering process by increments (partial),
  experimented on part of the software system (weak
  evidence).
• Aging symptoms metrics and assessment tools
  (full).

9
APPBANK experience packaging

• Decision model for defining the renewal process
  adequate to age symptoms (partial)
• Value model for technical-economic evaluation of
  the system (full)
• Heuristic model for cost estimation (partial)

10
CHARTRE Reused Packages

• Value Model for technical-economic evaluation of
  the system, metrics used and baselines were
  adapted. This model will be used for Quality
  Monitoring.
• Templates for application architectures with a
  good level of information hiding.

11
FILANTO Reused Packages

• Aging symptoms, the metrics were adapted.
• Decision Models for defining the renewal process
  adequate to age symptoms.
• Reengineering process was suggested. This
  suggestion was not accepted.

12
RCV Reused Packages

• Aging symptoms, the metrics were adapted
• Decision Model for defining the renewal process
  adequate to age symptoms.
• Reengineering process was suggested. The
  management is considering its decision.

13
CONCLUSIONS ...

• The experience package needs to be adapted,
  especially when it has an impact on already
  active processes. At any reuse, the metric plan
  had to be adapted.

14
... CONCLUSIONS ...

• The customer management prefers to use packaged
  experience when the evidence is carefully earned.
  (RCV and FILANTO decide against the reengineering
  process although it has been shown to be the most
  efficacious).

15
... CONCLUSIONS

• Even when innovative technologies have the
  evidence of their efficacy, they are adopted when
  the problem has an economic value that is grater
  then the risk of technology adoption. (Only the
  value model used for quality monitoring was used
  after the end of the project in CHARTRE).