© 2007MetaCase 1

Domain-Specific Modelling:The Killer App for Method Engineering?

12th September 2007

9:30 — 10:30

Steven Kelly, MetaCase

© 2007MetaCase 2

1

10

100

1,000

10,000

100,000

1,000,000

10,000,000

100,000,000

1,000,000,000

10,000,000,000

1950 1960 1970 1980 1990 2000

programmers

computers

Moore

Exponential growth is not smooth

� Moore: 18 months

� Computer: 33

� Programmer: 88

� 1980s PC revolution

� 2000 mobile revolution

� Computers/user:

– 1970: 1:10

– 2010: 10:1

� Customise with software

� Software crisis worsening!

© 2007MetaCase 3

How has productivity improved?

� "The entire history of software engineering is that of the rise in levels of abstraction"

Grady Booch

� New programming languages have not increased productivity

� UML and visualization of code have not increased productivity

� Abstraction of development can be raised above current level...

� ... and still generate full production code (and ignore it!)

*Software Productivity Research & Capers Jones, 2002

© 2007MetaCase 4

Contingency theory and software development

� One size does not fit all: Diversity due to:

– type of systems built

– technology used to implement them

– organizations, cultures, tools etc.

� Laboratory studies show that developers understand and use methods differently

– Extend, give new meanings, create own interpretations etc. for modeling constructs (e.g. in studies by Wijers, Verhoef)

� Contingency theory advocates flexible languages:no single language gives best result in all situations

– IFIP WG conferences (Olle et al. 1982, -83, -86, -88)

� Empirical studies show that companies prefer own methods

– 2/3 use internal, home-grown methods, Russo et al., Fitzgerald

� Create your own modeling language and generators?

– Domain-Specific Modelling

© 2007MetaCase 5

What is Domain-Specific Modelling?

Introducing Domain-Specific Modelling (DSM)

Where to apply – and why

Real life examples from various domains

© 2007MetaCase 6

DomainIdea

FinishedProduct

Solve problem in domain term

s

AssemblerMap to code, implement

UML ModelMap to UML

Generate,Add bodies

Domain Framework

Model inDSM

language

Generate codeNo need

to map!

CodeMap to code, implement

Modelling functionality vs. modelling code

© 2007MetaCase 7

Why is the vision possible (now)?

� Need to fit only one company’s requirements!

� Modelling is Domain-Specific (unlike 90’s UML)– Works for one company / product group,

application domain / framework / product family

– Language has concepts people are already familiar with

– Models used to solve the problem, not to visualize code

� Generator is Domain-Specific (unlike 80’s CASE)– Generate just the code needed from models

• Efficient full code

• No manual coding afterwards

• No reason for round-tripping

– Generator links to existing primitives/components/platform services etc.

– Can generate 3GL, Assembler, object-oriented, XML, etc.

© 2007MetaCase 8

Real-life cases, various domains

� Insurance products & eCommerce Java

� VoiceMenu for microcontroller Assembler

� Business Process Modelling for XPDL XML

� Call Processing Language XML

� Web application Java, XML

� Robotics C

� Infotainment systems

� Enterprise apps in Smartphones Python

� Symbian native applications C++

� IMS Service Creation

© 2007MetaCase 9

Insurance products & eCommerce ���� Java

© 2007MetaCase 10

VoiceMenu for microcontroller ���� Assembler

© 2007MetaCase 11

Business Process Modelling for XPDL ���� XML

© 2007MetaCase 12

Call Processing Language ���� XML

© 2007MetaCase 13

Web application ���� Java, XML

© 2007MetaCase 14

Robotics ���� C

© 2007MetaCase 15

Infotainment systems

© 2007MetaCase 16

Enterprise apps in Smartphones ���� Python

© 2007MetaCase 17

.RSS

.HRH

.PKG

.MMP

.INF

....

.LOC

.CPP

.H

Symbian native applications ���� C++

© 2007MetaCase 18

IMS Service Creation

© 2007MetaCase 19

Where to apply?

� Repetitive development tasks

– Large portion of the work similar to earlier products

– Several products made in parallel

– Useful to narrow down the design space

� Domain expertise needed

– Non-programmers can participate

� These normally include:

– Product Family

– Platform-based development

– Configuration

– Business rule definitions

– Embedded devices

© 2007MetaCase 20

Tools for DSM:Building a modeling language —i.e. real Method Engineering!

© 2007MetaCase 21

DomainIdea

FinishedProduct

Model inDSM

language

Easy!Normal(many)

Done a few times before!

Codegenerator

DSMlanguage

Frameworkcode

How to implement DSM

Domain Framework

Generate code

Expert(few)

© 2007MetaCase 22

Tools for DSM creation and use

� 6 ways to get the tools we need for DSM

1. Write own modelling tool from scratch

2. Write own modelling tool based on frameworks

3. Metamodel, generate modelling tool skeleton, add code

4. Metamodel, generate full modelling tool over a framework

5. Metamodel, output configuration for generic modelling tool

6. Integrated modelling and metamodelling environment

� Good tools minimize resource use (few man-weeks)

– create modelling tools automatically

– guide in DSM creation

– allow you to test DSM throughout domain design process

� Good tools allow DSML to evolve

© 2007MetaCase 23

A Brief History of Tools

� Tools for textual languages (late 70’s ->)

– SEM (Teichroew and Yamato)

– Others include Plexsys, Metaplex, Quickspec, PSL/PSA

� Tools for graphical languages (mid 80’s)

– Swedish Ramatic: set theory for metamodeling

– British Eclipse (not THAT Eclipse!): directed graphs

� Tools for graphical metamodeling (late 80’s)

– Finnish Metamodeling Editor MetaEdit: OPRR

� + dozens of others over the years: MetaView, Kogge, Virtual Software Factory, Paradigm+ SDK, Customizer in Excelerator, IPSYS ToolBuilder, ConceptBase, Dome, GME etc.

– Most of the tools focus on initial language specification and editor construction

© 2007MetaCase 24

Currently many 'version 1.0' tools— again!

� Single user

� Single modelling language at a time

� Simple metamodels

– Focus on objects, basic properties, binary relationships

� Simple notation

– Single graphical element + label

� Generator is text-to-text

– Or hand-written code to read models

� Resulting modelling tool primitive

– Missing majority of functions users expect in such a tool

� Upgrading modelling language invalidates models

� Upgrading tool framework invalidates tool

© 2007MetaCase 25

MetaEdit+ Eclipse EMF+GEF

1 hour 10kLOC=1yr=2000hComparison of EMF+GEF and MetaEdit+ for DSM, MDSD Workshop, OOPSLA ’04

© 2007MetaCase 26

MetaEdit+ Eclipse EMF+GEF/* Abridged version of LED.java, showing just parts related to property value. *//* Whole file is 3 times as long. In total, the 4 LED classes are 663 lines. */package org.eclipse.gef.examples.logicdesigner.model;import org.eclipse.gef.examples.logicdesigner.LogicMessages;import org.eclipse.ui.views.properties.IPropertyDescriptor;import org.eclipse.ui.views.properties.PropertyDescriptor;import org.eclipse.ui.views.properties.TextPropertyDescriptor;public class LED

extends LogicSubpart{public static String P_VALUE = "value"; protected static IPropertyDescriptor[] newDescriptors = null;static{

PropertyDescriptor pValueProp = new TextPropertyDescriptor(P_VALUE, LogicMessages.PropertyDescriptor_LED_Value);

pValueProp.setValidator(LogicNumberCellEditorValidator.instance());if(descriptors!=null){

newDescriptors = new IPropertyDescriptor[descriptors.length+1];

for(int i=0;i<descriptors.length;i++)newDescriptors[i] = descriptors[i];

newDescriptors[descriptors.length] = pValueProp;} else

newDescriptors = new IPropertyDescriptor[]{pValueProp};}public Object getPropertyValue(Object propName) {

if (P_VALUE.equals(propName))return new Integer(getValue()).toString();

if( ID_SIZE.equals(propName)){return new

String("("+getSize().width+","+getSize().height+")");}return super.getPropertyValue(propName);

}public void resetPropertyValue(Object id){

if (P_VALUE.equals(id))setValue(0);

super.resetPropertyValue(id);}public void setPropertyValue(Object id, Object value){

if (P_VALUE.equals(id))setValue(Integer.parseInt((String)value));

elsesuper.setPropertyValue(id,value);

}}

/* Abridged version of LEDFigure.java, showing just unselected symbol display. */

/* Whole file is 4 times as long, plus 65 lines for dragged symbol display. */

protected void paintFigure(Graphics g) {

Rectangle r = getBounds().getCopy();

g.translate(r.getLocation());

g.setBackgroundColor(LogicColorConstants.logicGreen);

g.setForegroundColor(LogicColorConstants.connectorGreen);

g.fillRectangle(0, 2, r.width, r.height - 4);

int right = r.width - 1;

g.drawLine(0, Y1, right, Y1);

g.drawLine(0, Y1, 0, Y2);

g.setForegroundColor(LogicColorConstants.connectorGreen);

g.drawLine(0, Y2, right, Y2);

g.drawLine(right, Y1, right, Y2);

// Draw the gaps for the connectors

g.setForegroundColor(ColorConstants.listBackground);

for (int i = 0; i < 4; i++) {

g.drawLine(GAP_CENTERS_X[i] - 2, Y1, GAP_CENTERS_X[i] + 3, Y1);

g.drawLine(GAP_CENTERS_X[i] - 2, Y2, GAP_CENTERS_X[i] + 3, Y2);

}

// Draw the connectors

g.setForegroundColor(LogicColorConstants.connectorGreen);

g.setBackgroundColor(LogicColorConstants.connectorGreen);

for (int i = 0; i < 4; i++) {

connector.translate(GAP_CENTERS_X[i], 0);

g.fillPolygon(connector);

g.drawPolygon(connector);

connector.translate(-GAP_CENTERS_X[i], 0);

bottomConnector.translate(GAP_CENTERS_X[i], r.height - 1);

g.fillPolygon(bottomConnector);

g.drawPolygon(bottomConnector);

bottomConnector.translate(-GAP_CENTERS_X[i], -r.height + 1);

}

// Draw the display

g.setBackgroundColor(LogicColorConstants.logicHighlight);

g.fillRectangle(displayHighlight);

g.setBackgroundColor(DISPLAY_SHADOW);

g.fillRectangle(displayShadow);

g.setBackgroundColor(ColorConstants.black);

g.fillRectangle(displayRectangle);

// Draw the value

g.setFont(DISPLAY_FONT);

g.setForegroundColor(DISPLAY_TEXT);

g.drawText(value, valuePoint);

}

© 2007MetaCase 27

What is needed in a mature tool?

� Meta-metamodel concepts– Explicit concepts for Graph, Role, Port

– N-ary relationships

� Constraints– Easy & powerful for relationships, property values

– Language for arbitrary constraints

� Model integration– Objects as property values

– Reuse of objects across models

– Links to subgraphs

� Notation and modelling tool functionality– Models automatically update when metamodel changes

� Purpose-built generator language and tools

© 2007MetaCase 28

Research Issues and Challenges

� Reuse

– Whole models and individual model elements

– Sequential and parallel

– Tool support for modeller vs. rules by metamodeller

� Versioning

– Metamodel level, upgrading the language

– Model level, with graphical diff

� Scaling

– Multiple users, concurrency vs. split & merge

– Multiple projects, sequential and parallel

� These all conflict with each other!

– At least if we try to apply familiar code-based practices

© 2007MetaCase 29

Roadmap hints for ME research

1. Understanding the state of the art

– mature tools

– previous research

2. Empirical research on the use of DSM in the field

3. Condensing 1 & 2 into advice for method engineers

4. Hypothesizing from 1 & 2 a set of new requirements for the next generation of tools

� Prerequisites for success:

– Don’t skip steps!

– Industry maturing past ”everything is UML”

– Academia going beyond ”ME = mix&match existing bits”

© 2007MetaCase 30

Try MetaEdit+ for free, buy your own for 150€,get 20 for 50€ each!

MetaCase

Ylistönmäentie 31FI-40500 Jyväskylä, Finland

Phone +358 14 4451 400, Fax +358 14 4451 405www.metacase.com

Thank you!

© 2007MetaCase 31

Further reading� Brinkkemper, S., Lyytinen, K., Welke, R., Method Engineering – Princi-

ples of method construction and tool support, Chapman & Hall, 1996.

� Czarnecki, K., Eisenecker, U., Generative Programming, Methods, Tools, and Applications, Addison-Wesley, 2000.

� Gray, J., Rossi, M., Tolvanen, J-P, (eds.) Special issue of Journal of Visual Languages and Computing on Domain-Specific Modelling with Visual Languages, Vol 15 (3-4), 2004.

� Kelly, S. & Tolvanen, J.-P., Domain-Specific Modeling, Wiley, 2008

� Kieburtz, R. et al., A Software Engineering Experiment in Software Component Generation, Proceedings of 18th International Conference on Software Engineering, Berlin, IEEE Computer Society Press, 1996.

� Pohjonen, R., Kelly, S., Domain-Specific Modelling, Dr. Dobb's, 8, 2002.

� Tolvanen, J.-P., Pohjonen, R., Automated Production of Family Members: Lessons Learned. Proceedings of International workshop of Product Line Engineering, Technical Report at Fraunhofer IESE (eds. K. Schmid, B. Geppert) 2002.

� Weiss, D., Lai, C. T. R., Software Product-line Engineering, Addison Wesley Longman, 1999.