• Steven J Vaughan-Nichols
• Building Better Software with Better Tools
• IEEE Computer Magazine V36n9(Sep 2003)pp12-14
• =EXPERIENCE TESTING DEBUGGING TOOLS QUALITY IBM Eclipse Hyades Microsoft Prefix Prefast SLAM Sun Jackpot
• IBM Hyades tool uses UML testing profile.
• Microsoft PREfix & PREface slow but thorough source code analysis. SLAM use simulation for model checking.
• Sun Jackpot JSCM code analysis: metrics and dataflow.

Doernhoefer03b

• Mark Doernhoefer
• Surfing the Net for Software Engineering Notes: Database Design
• ACM SIGSOFT Software Engineering Notes V28n4(Jul 2003)pp11-20 [ http://www.acm.org/sigsoft/SEN/ ] ACM DL subscribers: [ ??? ]
• =SURVEY DATA DESIGN KNOWLEDGE
• 17 good WWW sites

RyderSoffa03

• Barbara G Ryder & Mary Lou Soffa
• Influences on the design of exception handling...
• ACM SIGSOFT Software Engineering Notes V28n4(Jul 2003)pp29-35 [ http://www.acm.org/sigsoft/SEN/ ]
• =HISTORY EXCEPTIONS RESEARCH PRACTICE LANGUAGES LISP PL/I Ada C++ Java
• simple_maturation_model::= basic_research; concept_formation; introduced_into_a_programming_language; #(solution_refinement; (extend_language | develop_ne_language)).

WoodA03

• Allen P Wood
• Software reliability from a customer view
• IEEE Software Magazine V20n4(Jul/Aug 2003)pp37-42
• =EXPERIENCE QUALITY USER RELIABILITY HP/Compaq
• RVU::="Release Version Update".
• Customer installation is skewed hump peaking 8 to 12 months after release, and spreading over at least 21 months.
• Monthly failure rate very variable.
• Three month moving average shows a peak at the start and about 1 year later. then a decline.
• Larger changes in functionality leads to larger failure rates.
• Nearly 50% of failures are a new defect, 25% a rediscovered and fix is known, 30% is rediscovered and the fix is known.

Hohman03

• Luke Hohman
• The Difference between Marketecture and Tarchitecture
• IEEE Software Magazine V20n4(Jul/Aug 2003)pp51-53
• =EXPERIENCE MARKET REQUIREMENTS <> TECHNICAL ARCHITECTURE
• THere is more to software than writing code.
• Example of neat technical evolution that would have lost $5M dollars in revenue.
• Example of marketing modules that were technically boolean flags.
• example of an API that was technically convenient but never sold to application programmers!
• Marketecture::="software as imagined by marketing and sold o customer".
• tarchitecture::="software structure as imagined by developers".
• Compare with Davis03 phenotype and genotype.