• Stephen J Mellor & Anthony N Clark & Takao Futagami
• Model-Driven Development
• IEEE Software Magazine V20n5(Sep/Oct 2003)pp14-18
• =INTRODUCTION OMG UML MOF MDA PIM PSM MDD
• sidebar p 16. Glossary
  
  1. MOF::="Meta-Object Facility", OMG's standards for managing model data.
  2. MDA::="Model Driven Architecture", OMG standard...
  3. PIM::="Platform-Independent Model".
  4. PSM::="Platform-specific Model", results from weaving a PIM with a platform.
  
  

Fowler03c

• Martin Fowler
• Who needs an Architect?
• IEEE Software Magazine V20n5(Sep/Oct 2003)pp11-13
• =ESSAY AGILE ARCHITECTURE
• So what is architecture any way? Things you want to get right early in a project because they are hard to change later.

DevlinEtal03

• Keith Devlin (ed)
• Why Universities require computer science students to take math (special section)
• Commun ACM V46n9(Sep 2003)pp37-55
• =ESSAYS MATH
• Devlin: "Being smart is about doing, not knowing"... need to open up (train) abstract ways of thinking.
• Bruce et al give examples of how math helps develop good software.
• Henderson argues for math in software engineering because need precise abstract thinking.
• Almstrum reports a survey of 500 computer professionals on why they were attracted to computing: 1:benefit, 2:Science and math, 3:experimenting, 4:being in the Vanguard

ShanksTansleyWeber03

• Graeme Shanks & Elizabeth Tansley & Ron Weber
• Using Ontology to Validate Conceptual Models
• Commun ACM V46n10(Oct 2003)pp85-89
• =ADVERT ONTOLOGY MODEL REALITY ERD
• Abuses the term "Grammar" to mean "modelling language". Gives a bad SSADM model but doesn't show corrected one.
• Notes risks of seeing everything as tables of data. Some relationships are best expressed as rules/code rather than tables.
• Quotes Bunges's theory of ontology that distinguishes: things, properties of things, states of things, laws, events in things, and couplings.
• Critiques ERD as not including rules.
• Important to:
  
  1. Distinguish rules from relationships
  2. Model composites and aggregates as entities not relationships.
  3. Don't model relationships using attributes.
  4. Don't use optional attributes in entities.
  5. Distinguish objects/instances from classses.
  6. Distinguish things from their properties.
  
  

Henderson-Sellers03

• Brian Henderson-Sellers
• Method engineering for OO Systems development
• Commun ACM V46n10(Oct 2003)pp73-78
• =ADVERT OPF METAMODEL METHODOLOGY OPEN
• First, choose your process...
• OPF::="OPEN Process Framework".
• OPEN::="Object-Oriented Process, environment and Notation".
• A framework for selecting process components to fit a development situation.
• define five major classes of components: work product, producer, work unit, language, stage,...
• For part of the framework supplies a set of component instances: examples: 30 activity instances, 160 task instances, ...
• Task/activity matrix. Technique/task matrix.

SoderborgCrawleyDori03

• Nathan R Soderborg & edward F Crawley & Dov Dori
• System Function and Architecture: OPM-Based Definitions and Operational Templates
• Commun ACM V46n10(Oct 2003)pp67-72
• =ADVERT SYSTEM SPECIFICATION DESIGN PQRST GRAPHIC OPD LANGUAGE OPLTOOL OPCAT
• OPM::="Object-Process Methodology", a way of separating and what from how using OPD and OPL.
• OPD::="Object-Process Diagram".
• OPL::="Object-Process Language".
• Functions (whats) are often to control attributes of objects connected by processes. Architecture (Hows) also defines a processes and objects.
• See [ http://www.ObjectProcess.org/ ]

MeredithBjorg03

• L G Meredith & Steve Bjorg (Microsoft)
• Contracts and Types
• Commun ACM V46n10(Oct 2003)pp41-47
• =ARTICLE WEB/NET SPECIFICATION PROCESS ALGEBRA XML WSDL BEPL4WS WSCI
• Web services allow components to send each other specifications. Current standards define how to share interface and protocols. However the protocol languages are to complex.
• Process algebras with types provide a formal basis that is powerful and simple enough to be used.
• Reduction rules for processes and process types.