Datavetenskap Rogardt Heldal Introduction and pre-study - 1 - Lecture 1 Introduction to the course Model Driven Software Development (MDSD) Rogardt Heldal.

Datavetenskap

Rogardt Heldal Introduction and pre-study - 1 -

Lecture 1

Introduction to the course Model Driven Software Development (MDSD)

Rogardt Heldal

Datavetenskap


From Idea to system (small)

Idea Binary

Platform specific (PSM)Platform Independent (PIM)

C att

Datavetenskap


What if the system is a bit larger?

Datavetenskap


From Idea to system (small)

Idea BinaryC attReqs.

t

Datavetenskap


From Idea to system

Idea Reqs. BinaryCatt t

Datavetenskap


What if the textual reqs. are hard to understand

Datavetenskap


Analysis models

Idea Reqs. C BinaryUML+ At t at

Datavetenskap


Models for Communication• To make models readable is important: permit

communication. For example, to be able to discuss a new software system with the customer.

• Important that one don’t make the models too detailed.

• The contents of the models in the requirement/analysis phase are so far from code that there is need for humans refining the contents to obtain more precise models or code.

Datavetenskap


Analysis

Use Cases Diagram

Activity Diagram

Domain Model

Purpose: understand and describe the system

Descriptive Models

Datavetenskap


UML for Design

Idea Reqs. C BinaryUML+ A UMLDt t t

t

at

Datavetenskap


From Code Skeleton to Code

Code skeleton

Reqs. UML+ A

Code

Class Diagram

UMLD

Datavetenskap


Not using UML for Design

Idea Reqs. C BinaryUML+ A UMLDt t t

t

at

Datavetenskap


Productivity

Assembly

Fortran, C Productivity gain (factor of roughly 4)

C++, C# A small productivity gain (factor of roughly 1)

Models?

UML

Datavetenskap


Executable models• In this course we will also consider executable

models• As the name indicate, executable models are

models which can be executed • This require precise modeling• A subset of UML is used • Need to insert action language statements into the

models• Action language more abstract than higher

languages such as C++ and Java• Executable models can be translated to languages

such as C++, Java etc.

Datavetenskap


Executable Models

Idea Reqs. C Binary


UML+ A xtUMLt t at at

Datavetenskap


Executable Model Hierarchy

Prescriptive models

Datavetenskap


CxtUML at

Transformation

Java

C#

Platform1

Platform2

Platform2

mapTo

Datavetenskap


Domain-Specific Models• Work for a particular domain.• Can increase productivity a lot. • Often, relative easy to translate a domain-

specific model to code, because one knows what to produce.

• But: Very restricted language.

Datavetenskap


Domain Specific language

Idea Reqs. C Binary


UML+ A DSLt t at at

Datavetenskap


Domain Specific language

Idea Reqs. C Binary


DSL+ DSLt t at at

DSLDSL

Datavetenskap


What about xtUML?

Datavetenskap


Map to Hardware

Idea Reqs. C Binary


UML+ A xtUMLt t at at

Hardware

m

Datavetenskap


Mapping Component to Nodes

Some areas more than100 small computer and1000s of components.

Datavetenskap


MethodA method consists of:

• Modeling language (for example UML)

• Process (for example XP, RUP)

Datavetenskap


Course process

Analysis

Design

Requirement

Test

Water modelLearn topic by topic, so that force a bit of waterfall process,but try to be as agile aspossible within the project.

If you find mistakes, fix itwherever you find it!

Datavetenskap


Who am I (working)?• Teaching about object oriented system development for

13-14 years. For the last 10-11 years been the main instructor.

• Particular interested in modeling, the main focus of this course.

• Collaborate a lot with the industry, Ericsson, Volvo Car, Volvo Truck …

• Active in the modeling research community and have several papers in this area.

• Enjoy collaborating with industry via PhD students and master students. In particular when there are some research issues involved.

Datavetenskap


Course assistants

• Håkan Burden– Research area within Model Driven Engineering and

Natural Language Processing.

• Grischa Liebel– Research within Model-based requirements

engineering, took course in 2010.

Datavetenskap


Course structure• Two lectures a week

– Tuesday 13.15 – 15.00 – Thursday 13.15 – 15.00

• 4 voluntary exams– Will be done in combination with lectures

• Group work– Mandatory meetings every week with a supervisor

• support the development of the project

• Oral exam on the project work and course material– Time and place will be given later

Datavetenskap


Reading material • There will be no course book, but you are

responsible to find reading material on the topics covered in this course.

• Two books strongly related to this course:– Applying UML and Patterns (Craig Larman) (An

Introduction to Object-Oriented Analysis and Design)– Executable UML (Stephen J. Mellor and March J. Balcer)

(A Foundation to Model-Driven Architecture)

• We will add other recommended readings on the course home page during the course.

Datavetenskap


Course Home Page• The course home page can be found at:

– http://www.cse.chalmers.se/edu/course/TDA593/Year2013/start.html

• In this lecture we will highlight important issues considering the course, but all of you should also read the course home page carefully. There will be more details.

• The course home page will be updated during the course.• Changes done to the course home page will be indicated

as news in the beginning of the course home page, with the exception of lecture notes, project assignments which will be added without notification.

Datavetenskap


Work packages• Project work

– Weekly assignments– Contribution and time report– Final report – Oral examination

• 4 Voluntary written Examination• Oral examination on the project and course

material– Will be done two at a time– This will be in the examination week (week 8)

Datavetenskap


Project work• Done in groups of 10 students. Groups will be

formed today. • New assignment every week, see course home

page.• Mandatory meetings with supervisor every week,

but you should have more meetings yourself to discuss and solve the week’s assignment.

• Examination of project and course material:– Final report together with an oral examination– Contribution and time report

• For more details see the course home page.

Datavetenskap


Name SSNr Project

1 2 3 4 5 6 7

Voluntary exams1 2 3 4

pass

sum

Grade

Contribution

Grade

Gra

de

Group Nr:_______

Time:__________________________ Place: ______________________Supervisor: _____________________________

Contact person: _____________ email:____________telephone:___________Reserve: ___________ ___ email: ______________ telephone:____________

Executable

Datavetenskap


Examinations• To pass the course with the lowest mark

(3 or G) a student has to have: participated in writing a report on the group work – attended at least 6 of 7 group meetings – contributed a time report with stating the personal

contribution to the project – passed an oral examination on the project and course

work

• In border cases the voluntary examinations can help, if the student has obtained at least 24 points in total on 3 written examinations.

Datavetenskap


Examinations• To pass the course with a higher mark (4,5

or VG) a student has to have: – fulfilled all the commitments above and – obtained at least 36 points on 3 written examinations

for grade 4 and at least 48 points on 3 written examinations for the grade 5 (VG).

– Some extra points can be given for the project: a good solution, extra functionality, etc.

Datavetenskap


Contribution among members• You need to rank the contribution among the

group members• Indicate what mark the other group members

deserve

Datavetenskap


Groups • Groups:

– Tuesday 08.00 – 09.45 (2 groups)– Tuesday 10.00 - 11.45 (4 groups)– Tuesday 15.15 - 17.00 (2 groups)– Thursday 08.00 – 09.45 (2 groups)– Thursday 10.00 – 11.45 (2 groups)– Thursday 15.15 – 17.00 (2 groups)– Friday 08.00 – 09.45 (2 groups)

Datavetenskap


What should you do first week?

• A text description of your system. • Functional requirements • Non-functional requirements that will be

important.• A discussion or motivation of your

requirements.• Include documents that makes your report

stronger, like mind-maps, interviews and activity diagrams

Datavetenskap


Why hotel domain?

Datavetenskap


Modelling

• The first week there are focus on textual descriptions and textual requirements.

• But, we want to start to use models even at this stage, activity models.

• We will use simplified (subset) activity diagrams.

Datavetenskap


Example: Taxi company

Per

Kari

Lise

Datavetenskap


Mind Maps

Taxi

order

customer

regularTime

Start address

end address

Bus-taxi

Datavetenskap


Activity diagrams

• Part of UML• Can be useful for business modelling• Can be used to:

• Describe sequences of activities • Both sequential and parallel

Datavetenskap


Work flow/processes

Analysis

Design

Requirement

Test

Implementing

Start point

End point

Water model

Activity

Call actionnode

Datavetenskap


Action nodes

Some action

Signal name

Accept Event

Time expression

Call action node

Send signal

Accept event action node

Accept time event action node

Datavetenskap


Example

ProcessOrder

Request Payment

Paymentconfirmed

ProcessOrder

Cancelorder

Contactcustomer

Endof month

Pay outsalary

Datavetenskap


Control nodes

Decision node or Merge node

Fork node, join node

Initial node

Activity final

Flow final

Datavetenskap


Work flow/processes

Analysis

Design

Requirement

Test

Implementing

Start point

End point

Water model

Datavetenskap


Decision node• The output edge whose guard condition is true

is traversed.

[Condition 1]

[Condition 2]

Datavetenskap


Insert coins inthe machine

Test if machinehas received the right

amount of money.

Indicate that beveragecan be selected.

Select beverage.

Dispense beverage.

Indicate that beverageis not available.

[beverage not available]

[beverage available]

Example: Conditions and Iteration

Coffee machine

Datavetenskap


Decision input

[true]

[false]

<<decisionInput>>age > 25

Datavetenskap


Pre and Post-condition

Write grantapplication

Address grantapplication

Post grantapplication

<<localPrecondition>>address is know

<<localPostcondition>>Grant application is addressed

Send grant applicationPrecondition: known research areaPostcondition: grant application sent to address

Datavetenskap


Activity A. Activity B.

Concurrent fork

Concurrent join

Parallel activities

Datavetenskap


”Swimlanes”

Order productHandle order

Obtain materials

Package order

BillReceive product

Finish order

Pay

Customer Order Processing Store

Product process

Datavetenskap


Object nodes

InvoiceSendinvoice

Make Payment

Sendinvoice

Make Payment

Using pins:

Invoice

One can send objects:

Datavetenskap


Object in State

Create taxiorder

Order[new]

Assign taxi

Order[Assign]

Datavetenskap


Activity parameters

Order Order[Delivered]

Acceptpayment

Manufactureproduct

Deliverproduct

Order[Paid]

marketing Manufacturing Delivery

Handle order

department

Datavetenskap


Activity Diagram

Give card ValidateCard

Enter amount

Wrong pin less than 3 times

Wrong pin more than 3 times

Not enoughmoney on accountTake money

and card

Withdraw Money

Cancel

Enoughmoney onaccount

Correct pinCardDetails

Datavetenskap


Multicast and multireceiveRequest for Proposals process

Technical Group Member

Identify need

Request for Proposal RFP Create Proposal

Proposal[Candidate]

AssessProposals

Proposal[Accepted]

<<multicast>>

<<multireceive>>

Datavetenskap


Semantics• Action nodes execute when

– There is a token simultaneously on each of the input edges AND

– The input tokens satisfy all of the local action node precondition

Action node Action node

Action node

Action nodeDoes not execute

Action nodeexecutes

Datavetenskap


Summary• Introduced the course• Shown why models are important• Introduced the project • Shown activity diagrams

• By searching for Activity diagram on Google you will obtain many matches: introductions, tutorials etc.

• For more in depth information you can look at the UML specification. There is link from the course home page.

• Spend some time to understand activity diagrams before the first project meeting.

Datavetenskap Rogardt Heldal Introduction and pre-study - 1 - Lecture 1 Introduction to the course Model Driven Software Development (MDSD) Rogardt Heldal.

Documents