Exercise lecture : Exercise 2 and 3

1
Exercise lecture Exercise 2 and 3

• Rune / Yun

2
Overview

• Intro to exercise 3
• Aspects from exercise 2

3
Exercise 3

• Available, extended deadline with one week (see
  updated exercise schedule on web)
• Task 1 Evaluate the quality of another groups
  model from exercise 2 (instance model in task 2)
• Task 2 Goal-modelling Work further on your
  result from exercise 2, task 1 (type level model)
• Task 3 Strategic dependancy model, link EEML and
  i in same model
• Task 4 As task 3 in assignment 2
• Possible to work on task 1 and task 2/3/4 in
  paralell

4
Task 1

• Will get the METIS model made by another group
  (or from us)
• Will not get information about which group
• Evaluate physical, empirical, syntactic,
  semantic, and pragmatic quality (not the others)
• Important to have in mind that there is a large
  number of ways to arrange such models !
• Although it is possible to add to the model based
  on information from for instance the web sites
  that it was linked to, this is not mandatory.

5
Task 2 Goal modelling

• Extend your own type-level model
• If you learn new things relative to the new
  information in this case, you can (but do not
  have to) update the class and tasks models you
  have already made.
• Focus is on goal-modelling (including relating
  goals to tasks)
• Good starting point A new container Goals
• Add goals, goal relationships and links from
  goals to tasks
• There might also be goals listed in the task
  description from exercise 2

6
Task 3 Strategic Dependancy Model

• Using i as supported in METIS
• In the same model as you worked on before
• Limited form of meta-modelling (see next slide)
• Recommend to save the model after task 2
  seperately!

7
Task 3 Augment meta-model

• Open your model (currently EEML-model)
• Click the meta-model tab
• Select i (if it says not loaded, load it
  clicking the -sign)
• Drag i metamodel onto the background of your
  model
• On question if you want to augment the
  meta-model, answer yes

8
Task 3 SD-modelling

• Good starting point Add a new container with
  name e.g. actor dependancies
• Use the i concepts within this
• You should link some of the concepts in your i
  model to the same concept in the EEML model
  (using for instance a general relationship)
• Note that i and EEML appearantly share concepts
  (task, goal), on the other hand, these have
  different rules and properties, probably best to
  do the modelling by dragging concepts from the
  menu (and not using the New Object)

9
Short on exercise 2 (relative to exercise 3 task
1)

• Will not show a possible solution of this today
• Will have an overview of a possible solution for
  exercise 2 and 3 after exercise 3 is submitted
• Some overall aspects
• Folder structure
• Use of sub-modelling

10
Exercise 3, task 4 lanuage quality

• Not here expected to develop detailed evaluation
  according to the quality framework/language
  quality
• Same applies to task 4 in exercise 3
• Both positive and negative aspects relevant
• Will perform more detailed evaluation of language
  quality in later exercises

11
Exercise 3, task 4

• Not expected that you have been using all the
  potentially relevant functionality in exercise 2.
• Validity more important than completeness
• Both positive and negative aspects
• Must try to differentiat aspects of the tool and
  aspects of the concrete modelling language and
  its implementation in the tool
• Some highlights (next slide)

12
METIS physical quality

• Store models as files, but also repository
  solution exist (not in our setup), including
  versioning, access controll etc.
• Also the reporsitory is unfortunately currently
  working on a file level.
• Propriatary file format, but possible to make
  models available on web (Web report), or to be
  navigated using METIS browser.
• Difficult to include live models in e.g. document
  (word, take screendump/snagit etc)
• Sub-modelling to make it possible for a team to
  work on seperate parts for later integration
• Meta-modelling to make it possible to extend and
  specialise the modelling language, difficult to
  manage
• Possible to import/export models to other tools
  (e.g. from database or to excel)

13
METIS empirical quality

• Good alignement and equal spacing facilities
• Possible to specify layout strategies, and have
  these enforced
• Possible to change anchor points of relations
• Can differentiate relationships etc through
  different fonts/colors etc
• Available swimlane-view
• No spell checking of texts

14
METIS syntactic quality

• Syntax error prevention Can restrict allowed
  relationships between different modelling
  concepts, types of concepts being included as
  part of other concepts
• Can be sidestepped using type change mechanism
• Syntax error checking through validate model

15
METIS semantic quality

• Limited support of advanced consistency checking
  in the tool itself
• Some support of quick modelling
• Stamping, new object etc
• Importing from excel and databases with automatic
  layout
• Some support of quick changes
• Swap direction

16
METIS pragmatic quality

• Grouping of part of the language in folders etc
  makes it possible to work with large models, some
  with large depth in decomposition structure
• Filtering Both predefined (Find) and throught
  the definition of own criteria
• Possible to make own views based on an existing
  model, changes to a modelling concept propagated
  to all views
• Can view the same concept in different ways using
  different viewstyles
• Difficult to manage view functionalty. View
  criterias not persistently stored
• Any model execution/simulation must be done using
  other tools

17
METIS social quality

• Model annotation to make it possible to gather
  input from a lot of different stakeholders linked
  to the model
• Can model exceptions (in language constructs)
• No explicit support for argumentation processes

18
METIS organizational quality

• Little support for expressing and making sure the
  goals of modelling is fulfilled (can be included
  as part of the language)
• Limited methodological support
• Relatively expensive tool
• Relatively steep learning curve
• Training services available
• Tool available for the foreseeable future, stable
  customer base