Getting ESCHER Compliant

1
Getting ESCHER Compliant

• Don Hinton
• Institute for Software Integrated Systems
• EECS Department
• Vanderbilt University

6 May 2004
2
Problem

• How to help university research teams become
  ESCHER compliant, where ESCHER compliancy etails
  the following charateristics
• Transparency
• Accountability
• Repeatability
• Stability
• Compatibility

3
Desirable Characteristics

• Transparency
• Open source, source code freely available
• Accountability
• Bug tracking, New groups, Documentation
• Repeatability
• Versioned releases, SCM, e.g., CVS/VSS
• Stability
• Autobuilds, regression tests, metrics
• Compatibility
• Common data formats, e.g., XML, HSIF

4
Case Study

• ISIS-ESCHER Tools and
  Frameworks (Best Practices)
• ACETAO
• DESERT
• GME
• GReAT
• OTIF

5
Case Study (continued)

• ACETAO
• Transparency
• Open source
• Anonymous CVS, Viewcvs

6
Case Study (continued)

• ACETAO
• Accountability
• Bugzilla
• News group/mailing list

7
Case Study (continued)

• ACETAO
• Repeatability
• Source Code Management (CVS)
• Versioned Releases

8
Case Study (continued)

• ACETAO
• Stability
• Continuous distributed autobuilds
• Regression testing (via autobuilds)
• Metrics (performance, code size, compilation
  times)

9
Case Study (continued)

• ACETAO
• Compatibility
• Standards compliant, e.g., CORBA.
• Common data formats, e.g., XML
• CVS (viewcvs)
• Autobuilds (distributed,
• Metrics

10

• Scm (cvs/vss, viewcvs, open source)
• Autobuilds (testing, metrics, distributed)
• Bugtracking
• Web portal (visibility, availability)
• Release mechanism (numbered, tagged)
• Documentation (tutorials, etc...)
• Community (news groups, lists)