Fifty Enterprise Architect Tricks

Fifty Enterprise Architect Tricks Probably the most popular book about Enterprise Architect in the world Peter Doomen This book is for sale at http://leanpub.com/entarch This version was published on 2014-03-01

This is a Leanpub book. Leanpub empowers authors and publishers with the Lean Publishing process. Lean Publishing is the act of publishing an in-progress ebook using lightweight tools and many iterations to get reader feedback, pivot until you have the right book and build traction once you do. ©2012 - 2014 Peter Doomen

Tweet This Book! Please help Peter Doomen by spreading the word about this book on Twitter! The suggested hashtag for this book is #entarch. Find out what other people are saying about the book by clicking on this link to search for this hashtag on Twitter: https://twitter.com/search?q=#entarch

Contents Introduction . . . . Preface . . . . . . Copyright notice Thanks to . . . . About the author Online . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

1 1 2 2 2 3

Chapter 1: Elementary, my dear Watson . . . . . Trick 1: Adding an element of the same type . . Trick 2: Adding related elements to the diagram Trick 3: Put target element in diagram . . . . . . Trick 4: Bookmarking elements . . . . . . . . . Trick 5: Element styles . . . . . . . . . . . . . . Trick 6: Finding elements in the tree . . . . . . . Trick 7: Nesting elements . . . . . . . . . . . . . Trick 8: Locking elements . . . . . . . . . . . . . Trick 9: Adding elements directly to a package . Trick 10: Mimic multilanguage support . . . . .

4 6 7 11 13 15 19 21 23 25 27

Chapter 2: Connecting things . . . . . . . . . . .

29

CONTENTS

Trick 11: Exact placement of connectors . . . . . Trick 12: Repeat connector . . . . . . . . . . . . Trick 13: Pressing shift while dragging to add a bend . . . . . . . . . . . . . . . . . . . . . Trick 14: A bezier curve . . . . . . . . . . . . . . Trick 15: Prompt on connector deletes . . . . . . Trick 16: Adding notes to a connector . . . . . . Trick 17: Docking elements . . . . . . . . . . . . Trick 18: Parent-child relations, drag and drop style Trick 19: Converting linked element to local copy Trick 20: Connecting notes to internal documentation . . . . . . . . . . . . . . . . . . . .

31 32 33 35 37 40 42 44 47 51

Chapter 3: Diagrams . . . . . . . . . . . . Trick 21: Copy a diagram to the clipboard Trick 22: Same size . . . . . . . . . . . . Trick 23: Save as shortcut . . . . . . . . . Trick 24: Change diagram type . . . . . . Trick 25: Package on a diagram . . . . . Trick 26: Quick zooming . . . . . . . . . Trick 27: Shallow or deep copy? . . . . . Trick 28: A legend . . . . . . . . . . . . . Trick 29: Status colors on diagrams . . . Trick 30: Manual layouting . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

54 56 59 62 63 66 68 69 72 76 80

Chapter 4: Toolbox, types and tree Trick 31: Toolbox visibility . . . . Trick 32: Create your own toolbox Trick 33: General types . . . . . . Trick 34: Auto name counters . .

. . . . .

. . . . .

. . . . .

. . . . .

84 86 88 92 95

. . . . .

. . . . .

. . . . .

. . . . .

CONTENTS

Trick 35: Stereotypes . . . . . . . . . . . . . . . 98 Trick 36: Tagged value types . . . . . . . . . . . 101 Trick 37: Inheriting tagged values . . . . . . . . 103 Trick 38: Apply a tagged value to multiple elements 106 Trick 39: Sorting tree . . . . . . . . . . . . . . . 107 Trick 40: Finding orphans . . . . . . . . . . . . . 110 Chapter 5: A bit more advanced . . . . . . . . Trick 41: Slideshow . . . . . . . . . . . . . . Trick 42: Searches . . . . . . . . . . . . . . . Trick 43: Relationship matrix . . . . . . . . . Trick 44: Hierarchy window and traceability Trick 45: List view . . . . . . . . . . . . . . . Trick 46: Filtering . . . . . . . . . . . . . . . Trick 47: Printing large diagrams . . . . . . . Trick 48: The Team Review . . . . . . . . . . Trick 49: Database access . . . . . . . . . . . Trick 50: Visualize workflow . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

112 114 116 118 122 124 126 129 132 136 138

Bonus Chapter 1: various extra tips . . . . . . . . Trick 1: a simple todo list . . . . . . . . . . . . . Trick 2: quickly copying an element with its properties . . . . . . . . . . . . . . . . . . Trick 3: Hide stereotypes for information items on information items conveyed by a connector . . . . . . . . . . . . . . . . . . . . Trick 4: Show Namespaces . . . . . . . . . . . . Trick 5: A bunch of relations . . . . . . . . . . . Trick 6: Reporting on some specific elements . . Trick 7: Inserting images the easy way . . . . . .

141 141 142

143 145 148 151 153

CONTENTS

Trick 8: Updating the status of all requirements in a package . . . . . . . . . . . . . . . . Trick 9: How to make a diagram the model default, and how to undo it . . . . . . . . . . Trick 10: coupling roles to swimlanes . . . . . .

155 158 159

Bonus Chapter 2: Still more various extra tips . . 166 Trick 1: semi-batch creation of elements . . . . . 166 Trick 2: sorting a list on tagged values . . . . . . 167 Trick 3: using lists for easy export to spreadsheets 172 Trick 4: really advanced search . . . . . . . . . . 174 Trick 5: convert a scenario to a diagram - automatically! . . . . . . . . . . . . . . . . . . 177 Trick 6: how to make structured text from diagrams 180 Trick 7: Importing visio drawings . . . . . . . . 184 Trick 8: Links in notes . . . . . . . . . . . . . . 193 Trick 9: Operations as activities . . . . . . . . . 197 Trick 10: Working with working sets . . . . . . . 200 Bonus Chapter 3: Building your own toolbox . . 205 Trick 1: the order of the elements in the toolbox 206 Trick 2: setting a default element size . . . . . . 208 Trick 3: an icon for elements in the toolbox . . . 210 Trick 4: working with metatypes . . . . . . . . . 213 Trick 5: creating autohiding connectors . . . . . 215 Trick 6: how to create an arrow with shapescripting 219 Trick 7: the shapescripting canvas . . . . . . . . 224 Trick 8: some shapescripting limitations . . . . . 230 Trick 9: hiding connector labels . . . . . . . . . 233

CONTENTS

Trick 10: what to do when an MDG Technology is lost . . . . . . . . . . . . . . . . . . . .

237

Overview of shortcuts . . . . . . . . . . . . . . .

240

Introduction Preface “Enterprise Architect is an intuitive, flexible and powerful UML analysis and design tool for building robust and maintainable software. From requirements gathering, through analysis, modeling, implementation and testing to deployment and maintenance, Enterprise Architect is a fast, feature-rich, multi-user UML modeling tool, driving the long-term success of your software project”, states the Enterprise Architect Manual. Hmm. As an everyday Enterprise Architect user, I know that Enterprise Architect is very feature rich, but it can be hard to get the most out of the tool initially because there is so much functionality to learn. I once prepared a document with four Enterprise Architect tricks and presented these to a user forum. I was surprised that no participant, though every single one of them an expert user, knew all four. So I concluded there was a real need for a book like this. It is suited for those who have at least some working knowledge of Sparx Systems Enterprise Architect but don’t want to struggle their way through the 2000+ pages electronic manual that comes with the current version, 9.1. With these fifty different Enterprise Architect tricks, I barely scratched the surface of all Enterprise Architect possibilities. On the other hand,

Introduction

2

I carefully selected these tricks because they will be useful for most daily users. And be assured that I have enough material for another book like this, with more tricks and more advanced stuff. Happy reading!

Copyright notice © 2011 Peter Doomen. Sparx and the names of Sparx’s products referenced in this book are trademarks or registered trademarks of Sparx Systems Pty Ltd.

Thanks to Thanks to Estelle Bell and Scott Hebbard from Sparx Systems for their help. Thanks to Jonathan Feinberg for his famous Wordle application that generated the word clouds in this book. Thanks to Helmut Ortmann for providing me with additional tricks, and to Philippe Backx, Thomas Kilian and Ramsey Millar for their support in marketing this book. Thanks to my family and colleagues for their support and advice. Special thanks to my colleague Stephan for proofreading the manuscript and to colleague Herman for tip 9 in bonus chapter 2.

About the author Peter is a full-time enterprise architect at SD Worx, a leading HR company, and vice president of the Belgian Chapter of the Association of Enterprise Architects.

Introduction

Online Website of the author¹ Section about Enterprise Architect² ¹http://www.AbOrigineMundi.com ²http://www.AbOrigineMundi.com/index.php/tag/enterprise-architect/

3

Chapter 1: Elementary, my dear Watson

Chapter 1: Elementary, my dear Watson

5

Chapter 1: Elementary, my dear Watson

6

In which we find ten tricks to work with elements, the basic stuff of any Enterprise Architect model.

Trick 1: Adding an element of the same type When you are building a model, more often than not you need exactly the same type of element more than once. Instead of returning to the toolbox every time, just press “ctrl” and click on the diagram. For example: you need five “objects” to appear on the diagram.

Screenshot 1-1

First, create a new object: click “object” in the toolbox and click on the diagram. A rectangle appears and a dialog window pops up. Name the object “app1” and press “OK” so that the dialog window closes. Now keep the ctrlkey pressed and click on the diagram again. A second rectangle appears and the dialog window opens again. You just created a second “object” without having to visit the

Chapter 1: Elementary, my dear Watson

7

toolbox again. As simple as this sounds, it struck me that even experienced Enterprise Architect users did not know this little trick. A related trick is Trick 12: Repeat connector.

Trick 2: Adding related elements to the diagram This trick goes right into the heart of what a modeling tool is about: modeling instead of just drawing. The difference? When modeling, you understand there is a difference between the model and the diagrams, while when you are just drawing the model and the diagram are the same. Let’s say you are building a model for an organization that releases the wine magazine “Vinimble” six times a year. You have modeled the subscription process, a part of the application landscape, and some roles that work with the software to support the business process. The model quickly becomes complicated. The solution? Split up the diagram in separate diagrams, each showing a part of the model. The function “add related elements” is very useful for that. For example, let’s focus on the “Staff” role. Drag the “Staff” role from the model tree to a new diagram. Right click on it and choose “Advanced | Add | Related elements”.

Chapter 1: Elementary, my dear Watson

8

Screenshot 2-1

A dialog window “Insert related elements” appears.

Screenshot 2-2

For the moment, just press OK. The result is a diagram with all elements directly linked to the “Staff” role.

Chapter 1: Elementary, my dear Watson

9

Screenshot 2-3

The options on the “Insert related elements” dialog window will be useful when there are lots of elements linked to the source element on the diagram.

Insert Elements to X Levels Choose 2 as a value and Enterprise Architect will also show elements that are not directly linked to the source element, but to the elements that are linked to it.

For Link Type This option limits the action to elements that are linked with a specific link type. For example, choosing “aggregation” in our model would show only “VinimbleGroup”

Chapter 1: Elementary, my dear Watson

10

since the link type between VinimbleGroup and Staff is of that type. The links to the process activities “Check subscription”, “Register new subscriber” and “send previous issues for free” are typed as “Association” hence these process activities would not show up.

With Link Direction The default for this option is two-way, but you can limit this to source->target or target->source.

Limit to Element Type The same as link type, only that you limit the action depending on the type of the connected elements.

Layout Diagram when Complete You probably will want to turn this option off unless you trust the autolayout function of Enterprise Architect really well. I don’t, so I turn it off unless I have a diagram that has not been layouted by hand so far.

Limit to Namespace Another way of dividing complex projects so that the pieces become manageable. This will only work if you have namespaces defined. To define a package as a namespace root, right-click on the package in the model tree, choose

Chapter 1: Elementary, my dear Watson

11

“Code Engineering” and press “Set as Namespace Root”. The package icon changes so that you can recognize it as a namespace root. Useful if the model is very complicated and you use the “Add | Related Elements” function a lot and don’t have the time to delete elements on a diagram that you don’t need.

Screenshot 2-4

Trick 3: Put target element in diagram When consulting a complex model, it pays to have the relationships pane in front of you. It shows an overview of all elements connected to the currently selected element on

Chapter 1: Elementary, my dear Watson

12

the diagram. Open it by choosing “View | Relationships”. It will dock to the bottom of the Enterprise Architect window, which is just fine.

Screenshot 3-1

It not only shows the element, but also the connector type and other information you often need. Select an element on the diagram. Go to the Relationships

Chapter 1: Elementary, my dear Watson

13

pane and select any element that is not shown on the current diagram and right click on it.

Screenshot 3-2

Then choose “Place Target Element in Diagram”, go back to the diagram (the cursor will change) and click. The connected element is shown now. Together with Trick 2: Adding related elements to the diagram on page 14, this allows you to quickly build a new diagram based on an existing model.

Screenshot 3-3

Trick 4: Bookmarking elements Building models involves being able to focus on aspects of a problem. Often you want to focus on one aspect and suddenly you get information on another aspect. A bookmark is what you need, as a reminder that you still have to work on an element. Putting a bookmark on an element is easy: select the element and press shift+space bar. The same key combination removes the bookmark. Later, in Trick

Chapter 1: Elementary, my dear Watson

14

42: Searches on page 97, we discover how we can come up with a list of all bookmarked elements in the model. I discovered this trick by accident. A colleague asked how he could get rid of the “red triangle” that suddenly showed up on a diagram. I said I had the same problem, and realized that it was probably some odd key combination. Looking up the table of Enterprise Architect shortcuts revealed the culprit. But now, I use this function fairly often either as a reminder to myself or to point a colleague to “some work to do”.

Chapter 1: Elementary, my dear Watson

15

Screenshot 4-1

Trick 5: Element styles The success of a diagram is strongly related to its readability. Especially if your diagram has to convey a message to non-experts. That’s when colors and fonts come into play: use them to support meaning and the diagram instantly becomes more readable.

Chapter 1: Elementary, my dear Watson

16

Screenshot 5-1

Element styles can save you a considerable amount of time when layouting diagrams. Color an element, save this as a style and use it to color other elements in a consistent way.

Chapter 1: Elementary, my dear Watson

17

Screenshot 5-2

In the diagram above, I’ve used two styles: a gray background with a bold font means “Organization”, while a light blue background with a normal font means “Business Process Activity”. The difference with the previous diagram is obvious. Suppose we want to distinguish visually between automatic and manual process activities: manual processes are light blue, automated ones are light green. Select the activity “Check subscription”, which is supported by software, and color it green the usual way. Now click on the arrow in the pane at the bottom of the window, next to the style list. Select “Save as New Style” and give it a name of 12 characters or less (like “BusProcAut”). The style list is saved on a project base, which means that (1) you can use different conventions for different projects and (2) when

Chapter 1: Elementary, my dear Watson

18

you work on a project as a team and you use the same underlying Enterprise Architect database, you can share conventions so that the model is consistent with the visual style guide you probably have to use.

Screenshot 5-3

Then we are going to use our new style. Select the activity “Register new subscriber” and click on the styles list. Select “BusProcAut” and the activity turns light green.

Chapter 1: Elementary, my dear Watson

19

Screenshot 5-4

If we would have selected a different font or font color or a border color and style, it would change too. It even works if you select multiple elements with ctrl-click or by dragging around them. Neat huh?

Trick 6: Finding elements in the tree As we noted in Trick 2: Adding related elements to the diagram, there’s a difference between the model, as represented in the model tree, and the diagrams. An element can appear on one diagram, on many diagrams and even on no diagram. An element can even be twice in the model: nothing prevents you from having two actors

Chapter 1: Elementary, my dear Watson

20

“Subscriber”, for example. If this is the case, then how do you know which diagram element corresponds to which model element? Simple: select the diagram element and press “alt+g”. The focus will switch to the model tree and the corresponding element is selected. You will find yourself using this shortcut often to delete elements from the model, but you can save yourself time by selecting such an unwanted element from the diagram and pressing “ctrl+delete”. Only do this if you really want to delete an item from the model, not only from the current diagram! A related trick is Trick 40: Finding orphans on page 92. For the reverse action of alt+g, knowing which model element corresponds to which diagram element, there is also a shortcut. Select the element in the tree and press “ctrl+u”. If the element appears on exactly one diagram, that diagram is shown and the element is selected. If the element appears on more than one diagram, a dialog window with the list of these diagrams is shown. Double click the diagram (or select it and press “open”) to open the diagram you were looking for.

Chapter 1: Elementary, my dear Watson

21

Screenshot 6-1

If the element does not appear on any diagram, the same dialog window opens but with an empty list. I consider this a bug in Enterprise Architect: Enterprise Architect should tell you that the element is an orphan.

Trick 7: Nesting elements Often you need to visualize that an element is composed of other elements. For example, a component is composed of other components. Sometimes you only need the master component, sometimes you want to know the details and thus have the smaller components on a diagram. For such cases, you use element nesting. Suppose we have a cus-

Chapter 1: Elementary, my dear Watson

22

tom built application that uses a SubscriberManagement component. This component is composed of several other components: one that takes care of the subscriber lists, one that archives old subscriptions etc.

Screenshot 7-1

We would like to represent this visually by drawing the main component around the other components. Just put the components on a diagram and drag the child components into the parent component. The border of the parent changes to a dotted line, indicating that it is ready to accept children. In the model tree, or the project browser in Enterprise Architect speak, the component structure is reflected by a hierarchical relationship. When you drag the child component outside its parent’s border, it will become detached in the model tree too.

Chapter 1: Elementary, my dear Watson

23

Screenshot 7-2

Trick 8: Locking elements Once you will have done real hard work modeling a piece of reality or of your dreams, you will want to protect it against unwanted changes. For example changes by your project team colleagues who always argue with you about… well, you know what I mean. That’s the right time to start locking your work. Select the element to protect, right click on it and press “lock element”. Until you get the lock off, you and others won’t be able to change its properties (name, notes etc). The OK button on the properties window is grayed out.

Chapter 1: Elementary, my dear Watson

24

Screenshot 8-1

However, you can still move the element around in diagrams and even delete it from diagrams and from the model. If you want to protect the diagram, you guessed it, you have to lock the diagram by right clicking on empty diagram space and selecting “Lock Diagram”.

Chapter 1: Elementary, my dear Watson

25

Trick 9: Adding elements directly to a package You don’t have to open a diagram to add elements to a package. You can also right-click on the package in the tree, then select “Add | Add Element”.

Screenshot 9-1

A dialog window will open, asking you to provide Enterprise Architect with the information about the new element.

Chapter 1: Elementary, my dear Watson

26

Screenshot 9-2

Why would you do that? You don’t use every element every day. Therefore, sometimes it can be hard to remember where in the toolboxes an element is hidden. In this dialog, when you click the “Type” drop down box, a list with lots of elements shows up, making it easier to find the right one.

Chapter 1: Elementary, my dear Watson

27

Saves time and frustration!

Trick 10: Mimic multilanguage support Enterprise Architect does not support multiple (natural) languages. So if your organization happens to have let’s say a Spanish subsidiary, and you need to translate diagrams in Spanish, you’re in for trouble in the form of manually keeping two models in sync. Or do you? A little trick to mimic multilanguage support is by adhering to the following convention: put the English name of an element as the “Name” and the foreign name as the “Alias”. Or vice versa, it does not matter as long as you are consistent.

Screenshot 10-1

Then, when a translated diagram is needed, open the diagram properties window by right clicking on an empty

Chapter 1: Elementary, my dear Watson

28

part of the diagram and check the option “Use Alias if Available”.

Screenshot 10-2

The diagram will instantly use the Spanish translations, as far as you have used them in the alias field of the different elements. Neat!

Chapter 2: Connecting things

Chapter 2: Connecting things

30

Chapter 2: Connecting things

31

In which we discover ten tricks to connect elements with each other.

Trick 11: Exact placement of connectors For drawing connections between elements, Enterprise Architect’s default behavior is to draw the lines from the middle of one element to the middle of the other element. This can be a nuisance when you work with large elements that have lots of links, because you typically want the connector lines to be well separated. Open the menu “Tools” and choose “Options”. In the dialog window that opens, choose “Links” (left side). Then change the radio button for the option “New Connector End-Points:” to “Exact placement”. From now on, Enterprise Architect will draw the connection where you put it. Quite a time saver!

Chapter 2: Connecting things

32

Screenshot 11-1

Trick 12: Repeat connector Just like you can repeat the last element you drew (see Trick 1: Adding an element of the same type), you can repeat a connector. The shortcut is even simpler: just press F3. Then point to the source element and the cursor will indicate that you can start drawing the connector. You really need to memorize that one, because you’ll find yourself drawing the same connectors over and over again.

Chapter 2: Connecting things

33

Screenshot 12-1

Trick 13: Pressing shift while dragging to add a bend When a diagram becomes really crowded, adding a new connector without crossing existing ones is not easy. But if you press and release “shift” while dragging the connector, Enterprise Architect will add a bend point to the connector. You can even add multiple bend points, just press and release shift every time you want to have another one.

Chapter 2: Connecting things

34

Screenshot 13-1

You can bend a connector by pressing ctrl and clicking on the connector. However, when the connector style is set to “direct”, this is not possible. Right click, choose “connector properties” and set the style to custom to be able to bend it properly.

Chapter 2: Connecting things

35

Screenshot 13-2

Trick 14: A bezier curve A colleague of mine once told me, that IT people like their connectors straight and business people like to have them rounded. I don’t know if that is true or not, but in some diagrams, it would be nice if you could work with curved lines to connect elements. Here’s how to do it: open the connector properties window by double clicking on it, select the tagged values tab and add a new tagged value named “_bezier” with value 1.

Chapter 2: Connecting things

36

Screenshot 14-1

The result will be a line with an extra handle that changes the curvature of the bezier curve. This trick is little known and hard to find. Be proud you know this and use it to impress your colleagues with your in-depth knowledge of Enterprise Architect.

Chapter 2: Connecting things

37

Screenshot 14-2

Extra tip: if you keep forgetting how to do this, consult Trick 37: Inheriting tagged values.

Trick 15: Prompt on connector deletes When you’ve just installed Enterprise Architect, the tool will warn you when you delete a connector from the model.

Chapter 2: Connecting things

38

Screenshot 15-1

I suggest you turn this feature off. After a while, you will know exactly what happens when you press delete so you save time again. Then which default option should you choose? Hiding the relationship or deleting it altogether? It depends:

Chapter 2: Connecting things

39

Screenshot 15-2

Complicated diagrams If you work with complicated diagrams where you often need to hide relations because you are not interested in them (but you want to keep them in de model), it’s best to have “Hide the connector” as a default. This is probably when you already have a model and your main work is to exploit it. That way, you don’t have to use the cumbersome “right mouse click - visibility - hide connector” again. Deleting relations is then a matter of pressing ctrl-delete with a quick confirmation.

Chapter 2: Connecting things

40

Delete connectors often used On the other hand, when you find yourself needing to delete connectors often, obviously it’s better to check “Delete the connector from the model” to save time. This is probably when you are still building a model and your main work is to add new things to it. In any case, you can return to the default Enterprise Architect setting with the warning by checking Tools - Options - Links - Prompt on connector deletes. Or you can switch between the default options “delete” and “hide” whenever you need to, by visiting “Tools | Options | Links | Prompt on Connector Deletes”.

Trick 16: Adding notes to a connector Attaching notes to elements is easy and there are several ways to do it. But can a connector have notes attached to it? Probably, otherwise you would not be reading this trick.

Chapter 2: Connecting things

41

Screenshot 16-1

Here’s how to do it. Right click on the connector, choose “attach note or constraint” then check all relations that the connector should be attached to and confirm with OK. When you need to edit the links, right click on the note and choose “Advanced | Set Attached Links”. The Link relations dialog appears again and you can check and uncheck relations.

Chapter 2: Connecting things

42

Screenshot 16-2

Trick 17: Docking elements Strictly speaking, docking elements is not really connecting them. The “connection” between docked elements is not included in the model, it does not show up in any relations window. But sometimes it’s useful to have some relation between elements that is very loose. More lose than even the weakest real connector. That’s when you use docking. Docking can be compared to the grouping function of Microsoft Office, only that docked elements do not move together. They just stick to each other’s sides. Here’s how to dock elements. Select the elements to be docked, right mouse click on any one of them, and choose Dockable. Now deselect the elements, select one again and drag it to the side of the other. It snaps to the side as soon as it is close enough.

Chapter 2: Connecting things

43

Screenshot 17-1

Docking is useful for several reasons:

Docking saves space Suppose you have a list of hundred or more applications. With docking, you can neatly arrange them in the smallest space possible.

Chapter 2: Connecting things

44

Docking can be used to mimic tables or columns This is why elements like the activity partition are dockable by default.

Screenshot 17-2

Trick 18: Parent-child relations, drag and drop style The fastest way to organize elements, is by having support for composite objects turned on. You can then drag and drop elements to each other and Enterprise Architect will automatically arrange them hierarchically in the model tree, as we saw in Trick 7: Nesting elements on page 24. First check in “Tools | Options” if “Support for Composite Objects” is checked:

Chapter 2: Connecting things

45

Screenshot 18-1

Then drag an element to another element that would accept it. You can see that it works when the border of the parent element changes from normal to dotted. UML savvy readers will have objected to my using an aggregation relation between the actor and the object representing the actor’s organization in Trick 2: Adding related elements to the diagram on page 14. But there’s a good reason for that and it has to do with nesting: when building diagrams, I try to show hierarchical relations by nesting elements. If you use an aggregation relationship, Enterprise Architect will automatically hide the relation. The readability of the diagram improves and there’s less work to do, that’s what I call “pragmatism”! Here’s the normal view, showing the

Chapter 2: Connecting things

46

relation explicitly:

Screenshot 18-2

And here’s the composite view, showing the relation implicitly:

Chapter 2: Connecting things

47

Screenshot 18-3

This is much more readable, especially when you have lots of elements, and on top of that, when the elements are nested, the child elements move with their parent, saving time when making diagrams. It is important to know that these are just alternate views of the same model. In both cases, the relation window will show the aggregation relation between “Layouter” and “SurePrint”.

Trick 19: Converting linked element to local copy Any element on a diagram that belongs to another package than the package of the diagram itself, is considered a “linked element” by Enterprise Architect. For example, the element “Editor in chief” on the diagram below is “linked”, while the element “Approve for press” is not.

Chapter 2: Connecting things

48

Screenshot 19-1

You can recognize linked elements by their “namespace indicator”: the label that starts with “(from”. By the way, you can remove these namespace indicators by right-clicking on the diagram, then choosing “Properties | Diagram” and removing the check before “Show Namespaces”. Suppose you now want to have a duplicate of the role “Editor in chief”. Right click on the element, choose “Advanced | Convert Linked Element to Local Copy”. Enterprise Architect will react with a rather cryptic message:

Chapter 2: Connecting things

49

Screenshot 19-2

It actually just means that the new copy will inherit all local connections from its parent element. In this example, the “Editor in chief” was copied and also both links to the business process were automatically copied, but not the aggregation link to the VinimbleGroup element, since it is not a local connection.

Screenshot 19-3

From a modeling point of view, you should only create

Chapter 2: Connecting things

50

copies of elements when you discover that something which you thought was one thing is actually two separate things. It is not a good modeling practice to duplicate elements that refer to one and the same thing: it limits the maintainability of the model and you also lose some information that is contained in the links.

Screenshot 19-4

Chapter 2: Connecting things

51

Trick 20: Connecting notes to internal documentation You probably know that certain elements, like classes, can have attributes. For example, we are modeling the subscription procedure. Our class “Person” has attributes like “FirstName” and “FamilyName”. We will even keep extra info like the reason why we have a separate first name from a family name in our model. Now it would be nice when we could also show this documentation on a diagram. That’s when we use notes connected to the internal documentation of an element, like in the diagram below.

Screenshot 20-1

The advantage of using connected notes is of course that they will synchronize. For example, if we would change the attribute “Name” to “FirstName”, the note would also show “FirstName”, like below.

Chapter 2: Connecting things

52

Screenshot 20-2

Here’s how to make connected notes: On the diagram, find the element and select the feature you want to connect a note to. It will light up. Then right click and select “Create Linked Note”.

Chapter 2: Connecting things

53

Screenshot 20-3

It will show not only the name of the feature, but also the notes attached to it, exactly what we had in mind.

55

Chapter 3: Diagrams

Chapter 3: Diagrams

Wordle 3

Chapter 3: Diagrams

56

In which we work with diagrams, the main reason we use a modeling tool after all.

Trick 21: Copy a diagram to the clipboard This one is a beauty: simple, but versatile. The shortcut ctrl+b copies the current diagram to the clipboard. That way, you can easily use diagrams in other applications such as a drawing tool, a text processor or presentation software. Here’s a simple diagram copied to Microsoft Visio:

57

Chapter 3: Diagrams

Screenshot 21-1

If you don’t like the frame around the diagram, you can turn this off via “Tools | Options”. Remove the checkmark before “On Clipboard Images”.

58

Chapter 3: Diagrams

Screenshot 21-2

Also the format is determined in “Tools | Options”. You can choose between “Metafile”, a vector format, and “Bitmap”, which is, well, a bitmap. Usually you will need the metafile format, because it allows you to resize the diagram without quality loss. Also the file size of a metafile is a lot smaller than that of a bitmap. In rare cases, for example when the target application cannot work with metafiles, change the setting to “Bitmap” (but don’t forget to reset it afterwards).

59

Chapter 3: Diagrams

Screenshot 21-3

Trick 22: Same size Experienced users will know this one for sure, but even they will underestimate the visual impact of giving diagram elements the same size.

60

Chapter 3: Diagrams

Screenshot 22-1

For example, this diagram explains the business process activities of the editor in chief. The first version has unequally sized elements. It looks as if the “Register new subscriber” is a very important activity to the editor in chief (which is not the case, he is simply informed of new subscribers). Due to the different format (height larger than width) the “Write redactional column” seems to be something totally different from the other two activities, while in fact it is not. The solution is to make all three activities the same size.

61

Chapter 3: Diagrams

Screenshot 22-2

Here’s how to do it: Step 1: Select the elements you want to make the same size. You can do this with ctrl+click or by dragging around the elements to be selected. Step 2: Right click on the element that is the norm (all other elements, if possible, will get their height and width from this element). A striped band will appear around the element that is the norm. Step 3: Choose “Same Height and Width” from the context menu.

62

Chapter 3: Diagrams

Screenshot 22-3

Trick 23: Save as shortcut This trick can save you time when you are working on a project and you have different hats on. For example, you are business analyst but also technical architect. As a business analyst, you would probably need diagrams like the organizational model, the business processes and the requirements model. As a technical architect, you’ll open the component diagram, the technical architecture diagram and the like. Did you know you can have two shortcuts to the same Enterprise Architect model, each leading to a different set of diagrams open by default? Here’s how to do that. Step 1: In the File menu, choose “Save As”

63

Chapter 3: Diagrams

Step 2: As “Type of Save” choose “shortcut” (right side of the dialog box). Give the shortcut any name you remember and link to the diagrams it will open. In our example, the shortcut “business analysis.eap” will only open those diagrams that are relevant to the business analyst. Step 3: Select some diagrams and press “OK”. The link will be created. Test it by closing Enterprise Architect and clicking on the link.

Screenshot 23-1

Trick 24: Change diagram type One frustrating aspect of Enterprise Architect is that there are many toolboxes but it is difficult to make your own (see Trick 32: Create your own on page 75). On top of that, each time you activate a diagram, the toolbox settings

64

Chapter 3: Diagrams

change. Luckily, they don’t change in a random way. The toolboxes shown are determined by two factors: Visibility settings (see page 74 for Trick 31: Toolbox visibility) and the diagram type.

Screenshot 24-1

So you can influence what toolbox is shown by giving a diagram the right type when it is created. But you can also change the diagram type. Just press “Diagram | Change Type…” and you get the dialog window below. The tooltip warns you for restrictions: you cannot change any diagram type in any other type. Unfortunately, it is not clear what is allowed and what not. When you see the following warning:

65

Chapter 3: Diagrams

You will have to use another trick to have Enterprise Architect do what you want: Step 1: Make a new diagram of the type you need and open it. Step 2: Open the old diagram, select all objects with ctrl+a and press ctrl+c. Step 3: Activate the new diagram and press ctrl+v. The contents will be copied. If needed, you can delete the old diagram; remember that diagrams are nothing more than a view on a model so you don’t lose information apart from some diagram-owned elements, possibly.

Screenshot 24-2

Old diagram of type “Documentation”, to be changed to a sequence diagram:

66

Chapter 3: Diagrams

Screenshot 24-3

New sequence diagram:

Screenshot 24-4

Trick 25: Package on a diagram Enterprise Architect has a nice feature to make lists from elements that are present in the model tree. Just make a new diagram and drag one or more packages to the diagram. The result will be something like this:

67

Chapter 3: Diagrams

Screenshot 25-1

Note that the packages are kept up to date automatically: for example if you add an actor to the roles package, the diagram will reflect this change immediately. There are several uses for this feature. To name but two:

Combination with “Diagrams Only Export” In combination with the “Diagrams Only Export” function, you have an automatic table of contents that gives the reader a bit of context.

Combination with html export In combination with the html export, an overview diagram serves as a landing page for the reader. Just indicate this diagram as the default diagram (see below).

68

Chapter 3: Diagrams

Screenshot 25-2

Trick 26: Quick zooming If you happen to have a mouse with a mouse wheel, you can easily zoom in Enterprise Architect diagrams. Just keep the ctrl key pressed and use the mouse wheel to zoom in (roll from bottom to top) or out (roll from top to bottom). This diagram is at 50% zoom, useful if the diagram is complicated and you want to have an overview:

69

Chapter 3: Diagrams

Screenshot 26-1

Returning to the normal zoom factor (100%) is as easy: just ctrl+click with the mouse wheel and the diagram returns to standard zoom.

Trick 27: Shallow or deep copy? In some cases it might be necessary to copy an entire diagram. There are two types of diagram copies, which are somewhat mysteriously named “shallow” and “deep”. Right click on the diagram, then choose “copy diagram”. Go to the target package in the model tree, right click again

70

Chapter 3: Diagrams

and choose “paste diagram”. You get the following dialog window.

Screenshot 27-1

The diagram below has two elements, one in the same package and the other from a different package (Enterprise Architect sometimes calls this a “namespace”, confusingly).

71

Chapter 3: Diagrams

Screenshot 27-2

The first copy is a shallow copy, the second a deep. Each copy is made in a separate package. As you can see, the shallow copy has not copied the “owned” element (it links to the old element “WebBuilderPro” in the tree), but it has made a copy of the “non-owned” element “Vinimble.com”. If you duplicate the diagram in the same package, all elements will be duplicated irrespective of copy mode. As stated before, it is not good modeling practice to duplicate elements when they represent one and the same thing in the real world. Therefore, only use copies when you really want to have different elements:

One thing is actually two Either when you discover that one thing is actually two (you might for example learn that there are two applications called “subscribermanagement”). In that case, it is

72

Chapter 3: Diagrams

good modeling practice to name them differently, if only with a version number or something like that.

Playing with elements Or when you want to “play” with the elements on a diagram, for example when you want to develop different scenarios. In that case, either delete the copy after your study is done, either associate the corresponding elements with each other so you can keep the overview when the model evolves.

Screenshot 27-3

Trick 28: A legend Adding a legend to a diagram can improve its readability, but unfortunately, making legends in Enterprise Architect version 8 is largely a manual business. First, find the “Diagram Legend” icon in the toolbox “Common”. Click the icon, go to the diagram and click where you want the legend to appear.

73

Chapter 3: Diagrams

Screenshot 28-1

The dialog window “Legend” appears.

Screenshot 28-2

With the screen elements “Name”, “Fill Color”, “Line Color”

74

Chapter 3: Diagrams

and “Line Thickness” and with the controls “New”, “Save”, “Delete” and both hand icons, you can create and edit legend elements. There are two kinds of line elements: “Fills” and “Lines”. This might not be too obvious, but “Fills” stands for elements and “Lines” for connectors”. Elements will come first, lines second. You cannot change that. Each time you added a legend element, press “new” to add a new element. If you forget, the current element will be overwritten. Here’s what the Legend dialog box and the legend box on a diagram look like:

Screenshot 28-3

You can change the order of the diagram elements by using the up hand and down hand icons. The “Style Options” window serves for giving the legend box on the diagram its own colors. The different options here speak for their

75

Chapter 3: Diagrams

own:

Screenshot 28-4

Once you’ve made a legend, you can change it by double clicking it on the diagram. Now what’s more important: you can use the same legend on multiple diagrams. You can even choose to paste it as a copy or as a link. If you paste it as a copy, the resulting legends become unrelated. Change one, and the other one does not change with it. When you paste it as a link, however, the two legends will keep a relation. Change one and the other one changes too. This is handy when you work with multiple diagrams of the same type. So if you work with legends a lot, you can reuse them by making a collection of legends and copying/pasting them into the diagrams where you need them. It will help you to

76

Chapter 3: Diagrams

maintain consistency and therefore improve the readability and ultimately the usability of your diagrams.

Screenshot 28-5

Trick 29: Status colors on diagrams Certain elements, such as requirements, have a “status” and are able to show this status by means of a color on a diagram. At least when the setting “Show Status Colors on Diagrams” is turned on. Select “Tools | Options”, go to the “Objects” pane and check this option:

77

Chapter 3: Diagrams

Screenshot 29-1

From then on, each requirement on a requirements diagram will show its status, like this: By default, these are the statuses and their associated colors: Approved: Blue Implemented: Gray Mandatory: Orange Proposed: Yellow Validated: Green Now how can you define the statuses available and the colors that go with these? Click “Settings | General Types” and you get this dialog window:

78

Chapter 3: Diagrams

Screenshot 29-2

To me, the following statuses make sense:

Proposed A requirement is proposed, but not yet validated let alone implemented.

Validated The requirement is validated as being useful to the project, but business has not yet released it to IT so IT cannot yet implement it.

Chapter 3: Diagrams

79

Approved The requirement is handed over to IT so that they can work on it to implement it.

Implemented IT has implemented the requirement.

Mandatory: something else entirely The status “Mandatory” does not fit into this list: it is not a status, but a kind of priority assignment. I would suggest you remove this from the status list. Furthermore, you would probably need some way of saying that a requirement has not been validated or approved, so it becomes obsolete, and some way of saying that it has been postponed. With the General Types dialog window, it’s easy to add these types if needed. Until the colors you choose are common knowledge in the project team and to the relevant stakeholders, it also might be a good idea to show a color legend on each requirements diagram. See Trick 28: A legend on page 62 to learn how to work with legends and how to reuse them. Be sure to follow the lifecycle of the requirements as described above.

80

Chapter 3: Diagrams

Screenshot 29-3

Trick 30: Manual layouting You’ve probably discovered the autolayout task panel (in case you have not, it’s under “View | Layout tools”, after

81

Chapter 3: Diagrams

which a task pane appears to the left hand side of the workspace). Then you’ve also discovered that this function does not always work. And that you need to do some manual work to get the layout you want. This section gives you an overview of shortcuts. I suggest you learn them by heart, especially if you want your diagrams to be economic (in terms of space used) and readable (in terms of readers rolling their eyes and mumbling “I am too dumb for this”). Besides these two reasons, there’s another good reason to layout a diagram nicely. A sloppy diagram like the one below looks like you did not do your homework well. It encourages your readers to treat your work as an “alpha version”, and they won’t read it because they don’t have the time�hey’ll wait for the final version instead!

Screenshot 30-1

Basically, there are three actions you can take to give a diagram a nice layout:

82

Chapter 3: Diagrams

Sizing elements Adjusting the height and the width of elements such that they are equal in one or in two directions.

Aligning elements So that they line up nicely along an (invisible) edge.

Spacing elements Adjust the spaces between a series (3 or more) elements so that the spaces are equally wide. Usually, layouting a diagram is done in the order shown in the diagram below.

Screenshot 30-2

So you resize the elements, then loop through aligning and spacing them until it’s done. The final result will be like this. To resize elements, select them, then right click on the element that already has the right height and width, then

83

Chapter 3: Diagrams

choose “same height and width”. Since you’ll do this only once, you don’t really need a shortcut for that. Then, to align the elements, select them (in our example, first align the top elements, then the bottom elements, then the right elements, then the left elements) and press alt+ctrl+arrow. For example, to align the top edges of the top elements, press alt+ctrl+arrow up. Again, the element with the thick black-and-white striped border is the reference element. It is the element clicked on or selected last. To finish, adjust the position of the elements by selecting them and pressing shift+

Screenshot 30-3

Note that only the middle elements will move when you space items across. The left and the right item keep their places. The same holds for the top and the bottom elements when spacing vertically.

Chapter 4: Toolbox, types and tree

Chapter 4: Toolbox, types and tree

Wordle 4

85

Chapter 4: Toolbox, types and tree

86

Were we work with the magical three t’s: toolbox, types, and the tree.

Trick 31: Toolbox visibility Getting the right toolbox is as important for the Enterprise Architect user as it is for the carpenter. As we discussed in Trick 24: Change diagram type on page 55, it depends on the diagram type which toolboxes are visible and which ones are not. But there is an exception. You can force certain toolboxes to be visible in any case. Click “More tools” at the top of the toolbox pane to get the dialog window “Visible Toolbox Pages”.

Chapter 4: Toolbox, types and tree

87

Screenshot 31-1

Then simply check those toolboxes you need. At the bottom of the window, check “Show Relationship Pages” if you need specific relationships that go with element toolboxes you don’t need. As you will see when you close the “Visible Toolbox Pages” and you change diagram type, the diagram specific toolbox comes first, then the toolboxes you checked to be always visible. Good to know: instead of dragging elements from the toolbox to a diagram, you can also right click on the diagram and choose “New element or

Chapter 4: Toolbox, types and tree

88

connector”. The toolboxes shown are the same as those that are shown on the toolbox pane.

Screenshot 31-2

Trick 32: Create your own toolbox It is quite possible, though not straightforward, to develop your own Enterprise Architect toolbox. Follow these steps:

Create a package First, create a package and stereotype it as “Profile”. Add a diagram of type “Class” to it. The toolbox “Profile” will open automatically.

Chapter 4: Toolbox, types and tree

89

Screenshot 32-1

Add metaclasses Second step: for every element or connector in the toolbox, do the following 1. Add an element of type “MetaClass” and choose the right metaclass to start from. For example: the metaclass of “MyApplication” is “Object”, since we model an application as “Object”. 1. Add an element of type “Stereotype” and give it the name you want to appear in the toolbox. 2. Add an “Extends” relationship from the stereotype to the metaclass. 3. Optional but nice: change the default color of the stereotype to the default color you want these elements to get when using the toolbox. The final result looks like this:

Chapter 4: Toolbox, types and tree

90

Screenshot 32-2

Save the package Third step: right click on the package in the model tree and choose “Save package as an UML profile”. Go to the Resources pane (menu “View | Other Project Tools | Resources”) and right click on “UML Profiles”. Choose “Import UML profile”. Choose the xml file you just saved and click “Import”.

Chapter 4: Toolbox, types and tree

91

Screenshot 32-3

Import the package Fourth step: you can now open the toolbox by clicking “More Tools…” in the bar on top of the toolbox pane and start using it. There’s a lot more to be said about toolboxes: you can define your own shapes, add tagged values to your own stereotypes (see Trick 37: Inheriting tagged values) and more. But this basic introduction will probably be enough to start experimenting.

Chapter 4: Toolbox, types and tree

92

Screenshot 32-4

Trick 33: General types

Screenshot 33-1

We’ve met one general type in Trick 29: Status colors on diagrams, but there are five in total.

Status The one we already discussed.

Chapter 4: Toolbox, types and tree

93

Constraint A constraint is something that limits something else. For example, in requirements analysis, a constraint can limit the solutions you can come up with: you have to use a specific technology, the solution needs to be implemented on specific hardware,…

Constraint status Like requirements, constraints can have statuses. The list of statuses is by default the same as the requirements statuses mentioned above, but you can manage each list separately.

Requirement A requirement is a need that the solution you will build has to comply to. Enterprise Architect requirements can be internal (belonging to an element) or external (existing on their own, so you can connect them to more than one element). If you really want to work with requirements (which is a good idea if you want your customer to be happy with the project result), I suggest you change the default list of requirement types in Enterprise Architect with the Volere classification (http://www.volere.co.uk/); functional, data, look/feel, usability, performance, operational, maintainability, security, cultural, legal.

Chapter 4: Toolbox, types and tree

94

Scenario Working with scenarios is useful if you work with use cases as an analysis tool, but most elements in Enterprise Architect can have scenarios included (see the properties window of the element). The default list provided with Enterprise Architect includes simple, basic path and alternate. Usually, having a basic scenario with exceptions (alternate) is enough even for fairly complicated analysis types. Adapting these general types to your own needs can greatly increase the usefulness of Enterprise Architect as a modeling tool. Furthermore, since these general types are set per project, they help impose standards for a project. The weights indicated are useful if you calculate metrics. The dialog window of general types can be found in the menu Settings - General Types.

Chapter 4: Toolbox, types and tree

95

Screenshot 33-2

Trick 34: Auto name counters In the menu “Settings”, there is an option “Auto Name Counters” that is very useful when you work with elements

Chapter 4: Toolbox, types and tree

96

you need to refer to by name. The usual suspects are requirements, changes, and issues. Auto name counters give each element of that type a unique number, so you can make lists and refer to each item by its number without worries that you point to the wrong one.

Screenshot 34-1

First, indicate the type of element you want to setup an auto name counter for. Then give it a prefix (like “issue”) and a suffix (for example, to refer to the project in case you have multiple projects running at the same time). As a counter, type as many zeroes as needed to “pad” the counter, then a “1”. This indicates that you want the auto name counter to start with the number 1, appearing as “issue 0001 prX” on the diagram.

Chapter 4: Toolbox, types and tree

97

Screenshot 34-2

Of course, you can start with any value you want. If you input “0100” as a counter value, the first issue will be “0100”, the next one “0101” and so on. You don’t have to give a suffix and a prefix: neither of them or one of them is okay too. Put a checkmark before “Active”, confirm with “Save” and you will be able to use the auto name counter.

Chapter 4: Toolbox, types and tree

98

Screenshot 34-3

Trick 35: Stereotypes In Trick 5: Element styles, I showed you how to work with styles to give elements the appearance you want.

Chapter 4: Toolbox, types and tree

99

But there’s an even better way of accomplishing this. It is somewhat more complicated, but much more flexible. You cannot only change the colors of the text, the line and the background of an element, but you can even give an element exactly the appearance you want.

Screenshot 35-1

You could for example draw a nice picture for all actors that belong to the Vinimble staff. When you use a stereotype, it will automatically be the default appearance for all staff actors. Here’s how to do it: Step 1: Use any painting tool that supports metafiles (en-

Chapter 4: Toolbox, types and tree

100

hanced metafiles or *.emf are widely spread) to draw the icon you want to be associated with a stereotype. Save it as an emf file. Step 2: In Enterprise Architect, click “Settings | UML…” Step 3: In the dialog window that appears, type the name of the new stereotype, indicate a base class and choose “Metafile”. Click “assign” and point to the file you just made. Step 4: Don’t forget to press “save” when you are done.

Screenshot 35-2

When you now use an actor that is part of the Vinimble staff, give it the stereotype “ActorStaff” and it will appear as the image you made up. There’s more to be said about stereotypes. They are indeed one of the three mechanisms to extend UML in Enterprise Architect. But this little trick gives at least an idea of how powerful stereotypes are.

Chapter 4: Toolbox, types and tree

101

Trick 36: Tagged value types Enterprise Architect users will know that they can add tagged values to elements to keep information about specific properties of elements. For example, we could define a tagged value called “BusinessValue” and attach it to the requirements.

Screenshot 36-1

With tagged value types, we can go even further. When we define a custom type, we can enforce a standard and protect the user from entering wrong data. The example below shows how to limit the values of the tag “BusinessValue” to 1, 2 or 3.

Chapter 4: Toolbox, types and tree

102

Screenshot 36-2

As you can see, the spin control does not allow other values than the ones we like to have as values for the tag BusinessValue. Here’s how to do it: Step 1: Click Settings - UML and choose the tab Tagged Value Types. Step 2: Enter a new tagged value type. Name it BusinessValue and give a description. Step 3: As detail, add the following lines: Type=Spin; UpperBound=3; LowerBound=1; Step 4: Press save. Step 5: Open the properties of a requirement, choose the tab with tagged values, and add the tag “BusinessValue”. You’ll see the spin controls appearing on the screen.

Chapter 4: Toolbox, types and tree

103

Screenshot 36-3

Instead of working with a spin, you can also work with a drop down box: a discrete list of possible values, like ‘high’, ‘medium’, ‘low’. The details field reads: Type=Drop down; Values=high, medium, low;

Trick 37: Inheriting tagged values If you develop your own toolbox (see Trick 32: Create your own), you can add tagged values automatically to elements. Turn to the profile package again and choose the stereotype element of the item you want to add an inherited

Chapter 4: Toolbox, types and tree

104

tagged value to. Right click on the element and choose “Attributes”. For example, we will add the tagged value “BusinessValue” to all our requirements. Add an attribute named “BusinessValue” to the attributes of MyRequirement. The BusinessValue is of type “char” and the initial value is “medium”.

Screenshot 37-1

Delete the old profile in the resources pane and re-import

Chapter 4: Toolbox, types and tree

105

the profile (see Trick 32: Create your own). If you now add a requirement from your own toolbox, it will have the tagged value “BusinessValue” and it is set to “medium”. Nice, but it can be even nicer. Remember that we want to protect our users from typing in wrong data. So we’ll limit the options a business value can have. Add BusinessValueOptions (an element of type “enumeration”) to the profile diagram and add as its attributes the value list. In our example, we would add three attributes to the enumeration “BusinessValueOptions”: “Low”, “Medium” and “High”. Also indicate “BusinessValueOptions” as the “Type” for the BusinessValue attribute of the MyRequirement element.

Screenshot 37-2

Re-import the profile after having deleted the old one, and you can start using the BusinessValue tagged value with limited options.

Chapter 4: Toolbox, types and tree

106

Screenshot 37-3

Trick 38: Apply a tagged value to multiple elements You’ve gathered a bunch of requirements and want them to have the tag “BusinessValue”? Here’s how to do that: Step 1: Select the elements, either by dragging over them or by using ctrl+click if you want to select them one by one. Step 2: Press ctrl+shift+t. Step 3: The tagged value dialog window appears. Now just select the tagged value from the list (or type it in the combo box) and confirm. Step 4: You can now set the value of the tag per element.

Chapter 4: Toolbox, types and tree

107

Screenshot 38-1

If you work with tagged values a lot, you can open the Tagged Values pane and dock it to one side of the workspace. The shortcut is ctrl+shift+§ but you’ll probably use the menu View - tagged values.

Trick 39: Sorting tree There are at least two good reasons to sort the model tree in a convenient way:

Visual inspection and finding an element When a package has lots of elements, it’s much easier to find your way in it when it is sorted somehow.

Output By default, the order of the output is the order of the tree. When the order of the elements follows roughly the order of the diagrams, output is easier to understand.

Chapter 4: Toolbox, types and tree

108

Screenshot 39-1

The tree above is not sorted in any meaningful way, except that the elements are grouped by type. When you generate a rtf document from this package, readers will have difficulties deciphering the logical order of the subscription process. To get a more meaningful element order, first be sure that the option “Tools - Options - General - Allow Free Sorting” is checked.

Chapter 4: Toolbox, types and tree

109

Screenshot 39-2

Then use the green arrows on top of the model tree to put items in the right order.

Screenshot 39-3

The result would be this:

Chapter 4: Toolbox, types and tree

110

Screenshot 39-4

Trick 40: Finding orphans While modeling, the analyst sometimes adds elements to the model tree to find out later that they are not needed. Often, the element gets deleted from the diagram but not from the model. The result is what we call an orphan: elements that appear in the tree, but not on any diagram. With one of the default searches, Enterprise Architect helps us to find orphans like these. First, press ctrl+alt+a to open the model search window in the main workspace. Then, select as search type “Find orphans” and press “run”.

Screenshot 40-1

Chapter 4: Toolbox, types and tree

111

From this window, you can manipulate the object and even delete it if it is really not needed anymore. Select the element, click “delete” (the red cross icon) and it’s gone. There are other searches available and even a sql-based search builder. See Trick 42: Searches on page 97 for more about searching.

Chapter 5: A bit more advanced

Chapter 5: A bit more advanced

113

Chapter 5: A bit more advanced

114

Where we tackle some advanced topics, and also meet a few strange uses of Enterprise Architect.

Trick 41: Slideshow If you need to present your work, you can export your work (for one nice way of doing that, see Trick 21: Copy a diagram to the clipboard). But it is equally possible to stay in Enterprise Architect and give a custom presentation right from your workspace. Here’s how to proceed:

Screenshot 41-1

Step 1: Open the model view pane (menu “View | Model Views”). Step 2: Either select an existing model view or add a new one, for example “Presentations” in the model view tree by pressing the “New Model View” icon.

Chapter 5: A bit more advanced

115

Screenshot 41-2

Step 3: Add a Views folder to this model view by pressing the “New Views folder”. Step 4: Add a slideshow folder to this views folder with the “New Slideshow Folder” icon.

Screenshot 41-3

Chapter 5: A bit more advanced

116

Step 5: Now drag diagrams from the project browser to the slideshow you just created. You can add as many as you like. Sort them with the green arrows. You can choose whether to have an automatic presentation or a manual one by double clicking its icon in the model view tree. For automatic presentations, you can choose the number of seconds that pass before the next slide is shown.

Screenshot 41-4

To run a slideshow, just right click it in the model view tree and choose “Run Slideshow” or “Run Slideshow Fullscreen”. Enterprise Architect then opens the diagrams in sequence and closes them when the next one is shown. When you disable the automatic presentation, each time you click the diagram a small toolbar with controls appears next to the cursor. Press the play button to advance to the next slide (or press the spacebar).

Trick 42: Searches We already encountered one search option in Trick 40: Finding orphans on page 92. Here, we look for some more nice search features. After all, when your model has become an accurate representation of reality, you will consult it to get answers to your questions. Searching for specific

Chapter 5: A bit more advanced

117

information is then one option to really use the model. Now there’s a lot to say about searches. We will discuss only the most useful aspects of it here.

Screenshot 42-1

Step 1: First, open the model search window by pressing ctrl+alt+a. Step 2: Then, choose any of the predefined searches. For example, if you work with bookmarks as explained in Trick 4: Bookmarking elements on page 19, we can look for all elements that bear a bookmark. Very useful if you need a list of all model elements that still need work.

Screenshot 42-2

Step 3: Press “Run”. In the model search pane, a list of corresponding elements appears. Step 4: To group elements by any property, drag the column

Chapter 5: A bit more advanced

118

heading to the grouping row on the top of the results. Step 5: To sort elements, just click on the column. When you’re happy with the result, there are a few output options. For example, you can copy and paste the list to any other application, like a spreadsheet application, by selecting the elements, right click and “Copy Selected to Clipboard”, then paste it in the spreadsheet application. This is how the list looks when pasted in the spreadsheet application:

Screenshot 42-3

Alternatively, you can choose “Documentation” and send the results to an rtf file. Note that in the search result list, you can still edit the elements and their properties. This gives you a quick way of accessing only those elements you need, without having to scroll through the project browser.

Trick 43: Relationship matrix Modeling is all about defining the right relationships between elements. Sometimes, we are lucky and the relationships are hierarchical. Like for example the organizational tree view: each unit is part of a department, every

Chapter 5: A bit more advanced

119

department is part of the company. In most cases, however, relationships are many-to-many: every role acts in more than one process, and every process has more than one role taking part in it. A component is used by more than one application, but an application can use no components at all, a single component or several different components. When we want to work with complex relationships like these, we often want to represent them as a table. Enterprise architect calls this the relationship matrix. From version 8 onwards, this view has interesting functionalities. Let’s explore a few of them.

Screenshot 43-1

Step 1: First, select a package with the source elements in the model tree. Right click it and choose “Documentation | Open in Relationship Matrix | As Source…”.

Chapter 5: A bit more advanced

120

Screenshot 43-2

Step 2: Then, select some appropriate values for the fields on top of the relationship matrix. The source has already been filled in and we don’t bother with the profile yet. The result should be a matrix like this one. A first nice feature is that you can create and delete relationships from this view, at least when we did specify a direction (source -> target or the other way round). Right click any cell to edit the relationship (if there is one) or create it (if there isn’t). Then we can set some options. Click options and choose “Matrix Options” to make this dialog window appear. The options speak for themselves.

Chapter 5: A bit more advanced

121

Screenshot 43-3

When we achieve the result we had in mind, we can save the matrix either as a drawing (metafile or png) or as a CSV so we can import the matrix in our spreadsheet application.

Screenshot 43-4

Finally, we can save the options we have set to a profile. This is just a way to quickly set the same options again

Chapter 5: A bit more advanced

122

in the future, for the same or for other packages. It has nothing to do with UML profiles. Press “Profiles” to access this feature. Saved profiles will be available in the main relationship matrix window (top right).

Trick 44: Hierarchy window and traceability The traceability window (in Enterprise Architect below version 8 called “Hierarchy View”) is a nice feature for when you have a complicated model that you want to consult. Select any element on a diagram and press “View |Traceability” to get the traceability pane docked to the side of the workspace.

Chapter 5: A bit more advanced

123

Screenshot 44-1

The relationship buttons at the top allow you to only show those relationships that are currently of interest. Furthermore, you can: Step 1: right click any element in this tree to get a context menu where you can switch to the related element so you can navigate through the model, each time focusing on a different element Step 2: view the properties of the element

Chapter 5: A bit more advanced

124

Step 3: locate it in all diagrams (“Find in all Diagrams…”).

Screenshot 44-2

Trick 45: List view Diagrams, while flexible to work with, can become so complicated that it is difficult to keep the overview. For example, in a diagram with different element types and lots of connections, it can be hard to locate an element, or to count how many items of one type are present. The solution is to switch to list view by right clicking on the diagram: The result is a list like this:

Chapter 5: A bit more advanced

125

Screenshot 45-1

Here you can group by columns, change the placement of columns, sort columns (by clicking the black arrow that appears when you hover over a column header), edit items, make rtf reports of the list and more. Changing back to diagram mode is a matter of right clicking the list and choosing “Switch to Diagram Mode”. Warning: in previous versions of Enterprise Architect, you could copy the list to the clipboard so as to paste it in other applications. If you try the same in Enterprise Architect version 7.5 or higher, you will get undesired effects: Step 1: Pressing ctrl-c empties the description of the first item in the list and “undo” is not available. Step 2: Choosing “Edit | Copy” copies the diagram to the clipboard as a diagram, not as an element list.

Chapter 5: A bit more advanced

126

Screenshot 45-2

Trick 46: Filtering Filtering is a feature new to Enterprise Architect 8, it might even be the most useful new feature of that software version. Before version 8, we had no way but the manual and tedious one to show the difference between “as is” and “to be”, or between the deliverables of different project phases, or of different scenarios. So filtering is what we needed and finally got. Here’s how to work with filters. First, we’ll define some filters. Click “View | Diagram Filters”. The filter pane appears to the side of the workspace. Press the “New Filter” icon, choose a name for the filter (for example, “Phase 1”). Then choose the filter conditions.

Chapter 5: A bit more advanced

127

Screenshot 46-1

Here, I only require that an element’s phase description contains the character ‘1’. When I use the filter on a diagram, this is the result:

Chapter 5: A bit more advanced

128

Screenshot 46-2

So we see that in phase 1 of the project, we only have the SubscriberList component available. The archive component and the link to the website will come in phases 2 and 3. We can combine filters to show the full result:

Chapter 5: A bit more advanced

129

Screenshot 46-3

Instead of showing the deliverables for next phases as faded or gray, we can also hide them or select them. By cleverly combining filters and properties of elements and connectors, we can show exactly what we want with little effort. A truly useful new feature!

Trick 47: Printing large diagrams This trick has little to do with Enterprise Architect itself, but if you don’t have the money to buy a big printer, it can help you to print large diagrams so you can communicate effectively. Here’s an example of an impact diagram that won’t fit on a standard page from your printer, especially after the analysis is done:

Chapter 5: A bit more advanced

130

Screenshot 47-1

Now how to print this? We’ll need a standard printer, scissors and glue for that! First, print the diagram on multiple pages. Then, cut off the edges this way: Step 1: Top left page: don’t cut off the edges Step 2: Top pages starting from the second one: cut off the left edge

Chapter 5: A bit more advanced

131

Step 3: Second and following rows: for the leftmost page, only cut off the top edge. For all other pages, cut off the top and the left edge. (edges to be cut off are indicated in black). Step 4: Now glue the pages together, each time using the blank edge to put the glue on. (glue is indicated in green). Step 5: Tip: it helps if you have a diagram frame around the diagram. That way you can at least position the outer pages correctly. See Trick 21: Copy a diagram to the clipboard.

Printing large diagrams

Chapter 5: A bit more advanced

132

Trick 48: The Team Review Formerly called the “Discussion Forum”, the “Team Review” pane facilitates teamwork on a model. The integration of a discussion forum in the Enterprise Architect interface has two advantages:

Ease of use No need to leave your daily work environment and switch to another tool.

Links You can put links to elements from the model tree in the discussion forum, so it’s easy to point to elements that need work.

Chapter 5: A bit more advanced

133

Screenshot 48-1

To start, open the Team Review pane by clicking “View | Team Review” (or with the shortcut ctrl+alt+u). In the empty team review pane, make a new discussion and a new post. Double click the post and a blank sheet of paper will open in the main workspace. Here you can type the body of the post. To pinpoint the element or the diagram you are talking about, select a word or a group of words, right click the selection and press “Create | Link to Existing Element”.

Chapter 5: A bit more advanced

134

Screenshot 48-2

From the dialog window that appears, choose the right element and a link will be created in the document. When you or a teammate clicks it, he will be brought to the right element in the tree, making it easy to discuss the model. There are lots of other features connected with the team review function. To name a few useful ones:

Rich text Each discussion post is a “rich text” fragment, so you can copy/paste objects from other applications. Sounds obvious, but it makes posts lively and rich.

Chapter 5: A bit more advanced

135

Glossary definitions You can create glossary definitions from a text in a discussion post. Often you’ll find yourself talking about a word without having given it a clear definition. Right click the word, choose “Create | Glossary Definition” and give in the definition in the dialog window that pops up. The glossary is available in the menu “Project | Documentation | Glossary”.

Standard workflow There is a standard workflow available for the posts in the team review pane. Right click a post, choose “Review Status” and indicate the status. Depending on the status chosen, the post will show a different icon: a yellow triangle for posts awaiting approval, a green checkmark for approved posts, and a red cross for rejected posts.

Chapter 5: A bit more advanced

136

Screenshot 48-3

Trick 49: Database access Some operations would be very difficult, time consuming or nearly impossible to carry out in Enterprise Architect. Especially difficult to work with are connectors, since they don’t show up in the repository. That makes batch operations, like for example, finding all connectors of type “controlflow” and stereotyping them as “Businessflow”, tedious and error-prone. If you work on a local database, the trick is to access the database directly. As an example, let’s change all connectors of type “UseCase” to “Dependency”: Step 1: Close Enterprise Architect. Step 2: Make a backup copy of the file containing the model.

Chapter 5: A bit more advanced

137

Step 3: Rename extension of the original file from *.eap to *.mdb. Step 4: Open this file in Microsoft Access. Step 5: Locate the table t_connector and open it. Step 6: Filter the table so as to show only the UseCase connectors. Step 7: Change these to “Dependency”. Step 8: Close the file and re-rename it to *.eap. Step 9: Open it in Enterprise Architect and verify that the model is still okay.

Screenshot 49-1

Warning: Not all changes will lead to the desired results. Always have a backup copy of your model, always verify that the model is not corrupt and that there are no undesired side effects. Use this trick only if you understand how the Enterprise Architect datamodel works. If you don’t feel safe, don’t do it.

Chapter 5: A bit more advanced

138

Trick 50: Visualize workflow Agile methods have become increasingly popular in IT and rightfully so. When used well, methods like scrum and kanban can really help you achieve succes. One common guideline among agile methods is to visualize workflow. This basically means: keeping an overview of what needs to be done, what is being done and what has been done. Surprisingly, Enterprise Architect is a reasonably good tool to keep an overview like this. For modeling projects, a big advantage is that you don’t need to switch to another tool: you can keep your work list in your main working environment. The task list presented here might not be as flexible as a wall with post it notes, but it has the advantage that it works even for distributed teams. Here’s how to do it: Step 1: Make a new root level folder called “TO DO”. Add a diagram of type “UML Behavioral | Activity”. Step 2: Add three elements of type “Partition”. Name them “TO DO”, “In progress” and “Done” respectively. Change their orientation to vertical (by clicking the fourth little icon on the top right of the partition element on the diagram, or by right clicking and choosing “Advanced | Vertical Partition”). Step 3: Stitch the partitions to each other (they are dockable by default, see Trick 17: Docking elements on page 39). Step 4: Now for each task, add an element of a type you otherwise don’t use (like “Action”). This way, you can

Chapter 5: A bit more advanced

139

use searches (see Trick 42: Searches97 on page 97) without getting your model mixed up with your task list. The text of the item should reflect the task. Step 5: Put each task in the right partition: TO DO if it is untouched yet, In progress if someone is working on it, and Done if it’s finished. To change the status of a task, simply drag it to the next partition. The diagram will look like this:

Screenshot 50-1

And the model tree looks like this:

Screenshot 50-2

As you can see, the actions are “owned” by their partitions, so it is easy to keep an overview or to extract reports.

Chapter 5: A bit more advanced

140

Some more tips Tip 1: If a task is related to a diagram or a specific element, include the path to this diagram or element in the notes (right click “Copy Reference | Copy Node Path to Clipboard). Tip 2: If you have a specific workflow each task goes through, split the “In progress” column in more columns, each indicating a workflow stage. For example, you might have “Analysis”, “Documentation” and “Review”. Tip 3: For each “work in progress column”, add a maximal number of tasks that can be here at the same time. In our example, we have only one such column and it has a work in progress limit of 4. This simple addition implements the kanban guideline “Limit work in progress”. When you stick to it, it helps you to keep focus and become aware of bottlenecks in your workflow. Tip 4: If a task is high priority, give it a distinct color (in our example: orange). When a task moves from “In progress” to “Done”, there is room for a new task. Which one to choose from the TO DO column? Easy: the one that is high priority, in this case “complete application landscape”.

Bonus Chapter 1: various extra tips Trick 1: a simple todo list When you are developing a model, you’ll want to put some things off your head. Suppose you are currently designing the system and you suddenly remember that you have still to flesh out the code for a component you use on the diagram.

Screenshot 1-1

I’ve just attached a note to the component, but I could have put this in the notes of the component description as well.

Bonus Chapter 1: various extra tips

142

Now you can easily search your entire model by pressing ctrl-f and using “todo:” (without quotes) as a search term. Enterprise Architect will come up with a list of items where todo: appears in the body or in the description.

Screenshot 1-2

Of course this works with all search terms, not just “todo:”. You could even use @tags if you like (like, “@john”, to indicate John has to work on this). Unfortunataly, hashtags won’t work (“#urgent”) so I stick with the “:” marker.

Trick 2: quickly copying an element with its properties To quicly copy an element with all its properties, select the element, press ctrl+c and shift+ctrl+v. You will get a dialog window like this:

Bonus Chapter 1: various extra tips

143

Screenshot 2-1

Now you can change the name if you like. You probably will, since having two elements with the same name does not make much sense from a modeling perspective. It can only create confusion. The element and its properties, but not its links, will be copied. You can repeat the shift+ctrl+v to get additional copies.

Trick 3: Hide stereotypes for information items on information items conveyed by a connector Some connections, such as “information flow”, allow to indicate the information that is conveyed by the flow in a more structured way than by just naming the flow. That

Bonus Chapter 1: various extra tips

144

way, you can reuse the information model you built by having the classes or objects appear as flow labels. Now if such a class or object has a stereotype, it will show on the connector, like here:

Screenshot 3-1

While sometimes this is a useful piece of information, at other times it clutters the diagram. Luckily, there is a way to turn these stereotype labels off. Open Tools | Options and uncheck “Show Stereotypes in the box “Project Browser”. Restart EA. The stereotype label has gone!

Bonus Chapter 1: various extra tips

145

Screenshot 3-2

As a side effect, the stereotypes disappear from the tree as well. Here is what you get:

Screenshot 3-3

Trick 4: Show Namespaces By default, Enterprise Architect diagrams show for every element outside the diagram package the name of the

Bonus Chapter 1: various extra tips

146

package it belongs to. Especially for diagrams with lots of foreign elements, this adds to the clutter and readability suffers.

Screenshot 4-1

It is good to know you can turn this off in the diagram properties window. For some reason, Enterprise Architect calls the package name a “namespace” even if you did not indicate it as such. Open diagram properties by rightclicking on an empty part of the diagram, and clicking on “Properties”, choose the “Diagram” tab and uncheck the option “Show Namespace”.

Bonus Chapter 1: various extra tips

Screenshot 4-2

The result is a much more readable diagram:

147

Bonus Chapter 1: various extra tips

148

Screenshot 4-3

Trick 5: A bunch of relations In my work as a business analist for Vinimble, I discovered that there are three main groups of stakeholders, each with their own specific requirements. On the other hand, I also discovered that there are some requirements that are shared by all stakeholders. Let’s model that with Enterprise Architect.

Bonus Chapter 1: various extra tips

149

Screenshot 5-1

How to connect the general requirements on top to all three stakeholders? Drawing the relations one by one is cumbersome and there’s a risk you miss one. The solution is to turn to the relationship matrix. Rightclick on the package with the general requirements in the model tree and choose “Documentation | Open in Relationship Matrix | As Source”:

Bonus Chapter 1: various extra tips

150

Screenshot 5-2

Then select all cells by clicking the topmost left cell, pressing shift and clicking the bottom right cell. Release the shift button and right click any of the selected cells. Then choose “Create Relationship | UML::Dependency”:

Screenshot 5-3

Turn to the diagram again, and there’s the result we expected:

Bonus Chapter 1: various extra tips

151

Screenshot 5-4

Or, as my colleague Erwin put it, “the result is an ugly diagram”. Indeed, but I am sure you can beautify it! And, being an architect, you are well aware that it is the model that counts, not the diagram…

Trick 6: Reporting on some specific elements Sometimes you need reports. For example when meeting with non-EA users about a specific piece of your model. Suppose you have this big package full of well-documented elements. How to pick only those that are relevant for the review meeting? Of course, you could just generate the whole document and then strip out what you don’t need.

Bonus Chapter 1: various extra tips

152

But that is rather time-consuming and error-prone. There is a better way: • Right click the package in the tree • Choose “package browser”. The middle pane will now show the list of elements in the current package instead of a diagram. • Select the relevant elements with ctrl+click • Right-click any selected element and choose “Documentation” | “RTF Report on Selected Items…”.

Screenshot 6-1

Now just continue generating the RTF report. The result will be a nicely formatted RTF document with exactly the elements you need!

Bonus Chapter 1: various extra tips

153

Screenshot 6-2

Trick 7: Inserting images the easy way In my work as a business analyst, I’ve created diagrams, models, drawings and other artifacts outside Enterprise Architect. For example, I’ve made this Mind Map with seven types of waste typically found in processes. Now I would like to add this picture to a diagram, for example to take it to a stakeholder meeting as one single piece of paper. How to do that? This is really easy. Grab the picture in Explorer and drag it to the diagram:

Bonus Chapter 1: various extra tips

154

Screenshot 7-1

Then choose “Insert”:

Screenshot 7-2

The result is exactly what we want: we have a diagram with the addition image. It is even resizable. It shows up as a “boundary”. You can change its properties like its name by right-clicking it and choosing “Properties”, like any other EA element.

Bonus Chapter 1: various extra tips

155

Screenshot 7-3

Trick 8: Updating the status of all requirements in a package After the meeting with the business, all of the general requirements I gathered so far have been approved by the business stakeholders. How can I update their status without the time consuming action sequence “open requirement - change status - close requirement”? Simple: right click the package, select “package control” then “update package status…”

Bonus Chapter 1: various extra tips

156

Screenshot 8-1

The dialog that opens now lets you update the statuses of all the elements, but also the version and phase:

Bonus Chapter 1: various extra tips

157

Screenshot 8-2

In case you have a diagram with requirements open, and the requirements show a color to reflect their status (see trick 29), you need to close and reopen the diagram to see the effect:

Screenshot 8-3

Bonus Chapter 1: various extra tips

158

Trick 9: How to make a diagram the model default, and how to undo it Do you find yourself opening the same diagram each time you start up Enterprise Architect? Or do you want to focus the team on the to do list when they open the model in EA? Here’s how you can indicate the default model to open when starting up the tool. Open the diagram. In the “Diagram” menu, click “advanced” and “make model default”.

Screenshot 9-1

Now every time anyone opens the diagram, this diagram will be shown. This is especially useful for overviews or to do lists. Remark that there is no way to undo this. Or is there? Just create some empty diagram, indicate it as the model

Bonus Chapter 1: various extra tips

159

default, and delete it.

Trick 10: coupling roles to swimlanes In business modelling, there are two competing ways to model a process:

Implicit Put activities in swimlanes, representing a role or an actor. This is the “implicit” way of working. A clear advantage of this approach is that for simple processes, it is a very simple, understandable picture. People won’t need uml training to grasp the meaning: the swimlane represents the responsibility area of a role. Screenshot 1 gives an example of a simple process description using swimlanes.

Bonus Chapter 1: various extra tips

160

Screenshot 10-1

There are also disadvantages: • Shared responsibilities: It is difficult to represent shared responsibilities. Suppose you want to indicate that both the reviewing editor and the editorin-chief are responsible for the final review of an article, you’ll have to clutter the diagram either by duplicating the activity, either by adding a swimlane for the combination “reviewing editor” and “editorin-chief”. • Traceability: a modeling tool ultimately leads to a model that is more than the sum of the diagrams.

Bonus Chapter 1: various extra tips

161

For example, once you start designing the application landscape, you could benefit from the business process model to connect the applications with the business activities. Later, you could easily find which roles are most likely users of what applications by exploiting the model. • Space: often, the swimlane notation takes a lot of space, most of which is empty. Especially when there are actors who don’t do a lot in the process. There are a few tricks you can use so your diagrams don’t grow too large (like having a swimlane “others” where you combine roles that appear only once), but even then the problem remains.

Explicit In the explicit approach, you connect the activities directly to elements representing the roles. This gives compact diagrams with full traceability, but they can be somewhat harder to read, especially for those who are used to swimlane style process diagrams. Screenshot 2 gives an example of a simple process description using the explicit notation.

Bonus Chapter 1: various extra tips

162

Screenshot 10-2

Combining the explicit and the implicit way of working Is there a way to have the advantages of both modelling approaches, without too much duplication or other overhead

Bonus Chapter 1: various extra tips

163

work? Yes there is! First, describe the process using the swimlane notation. Add the roles to the diagram and connect them to the activities. Then delete the roles and open the swimlane editor (right click on an empty space of the diagram, choose “Swimlanes and Matrix…”). Now edit the swimlanes one by one by clicking the swimlane and choosing “Modify”. Next to the “Classifier” label and text box, there is a button with “…” (screenshot 3). Click it, and search for the role in the tree that appears. Click OK and the role will be connected to the swimlane.

Bonus Chapter 1: various extra tips

164

Screenshot 10-3

It will even be indicated in the diagram (see screenshot 4: instead of just “editor” the swimlane header label says “editor:reviewing editor”).

Bonus Chapter 1: various extra tips

Screenshot 10-4

165

Bonus Chapter 2: Still more various extra tips Trick 1: semi-batch creation of elements Sometimes, you need to add a batch of elements that have a few properties in common, but not all. Like when you want to create ten or more actors stereotyped as “BusinessRole”. Here’s how to do that efficiently: • switch to list view: right-click on the diagram and choose “List view” • choose “New element” (ctrl+N or press the “new” icon, top left of the list) • Indicate the element type (“Actor”) and the stereotype (“Businessrole”) • uncheck “Open Properties Dialog on Creation” and “Close Dialog on OK” • Now give the name of the first element, press create, change the name to that of the second element and repeat.

Bonus Chapter 2: Still more various extra tips

167

Screenshot 1-1

Simple, but very effective! I even use this instead of importing via csv if the list to import is relatively short.

Trick 2: sorting a list on tagged values Suppose we keep track of our main technologies. Currently, we use Java, Python and HTML. Most components and applications use only one of these technologies, some can use two or even all three. We also want to keep track of the version of the technology used: SE 6 and SE 7 for Java, 4 and 5 for HTML and so on.

Bonus Chapter 2: Still more various extra tips

168

A good way to do that is to use tagged values. The tags are “JavaVersion”, “PythonVersion” and “HTMLVersion”. Values are “SE 6” and “SE 7” for Java and so on. We now want to sort a list of all our components and applications based on these tagged values, so we can discuss with the architect which elements need an upgrade and which are fine for now. Here’s how to do that. Step 1: Define it in the UML types window (“Settings | UML Types | Tagged Value Types”)

Bonus Chapter 2: Still more various extra tips

169

Screenshot 2-1

Step 2: Use the tagged value by accessing the properties dialog of the element and clicking on “Tagged Values”.

Bonus Chapter 2: Still more various extra tips

170

Screenshot 2-2

Step 3: Sort the list view: open the diagram in list view, right click and choose “Add tag value column”

Bonus Chapter 2: Still more various extra tips

171

Screenshot 2-3

Extra tips • It is also possible to use a tagged value that is not defined in the UML types window. Select an element in the list view, right click it and indicate “Tagged Value types from the Selected Element”. Then choose the tagged value from the list. • You can even drag the tagged values to the column of the list window, so as to hide any elements for which a certain technology is not used. This allows you to focus quickly on those applications that use HTML, for example.

Bonus Chapter 2: Still more various extra tips

172

Screenshot 2-4

• If you leave the tagged values window opened (“View | Tagged values”), it is easier to manage the tagged values for the elements you are working with.

Trick 3: using lists for easy export to spreadsheets Exporting data from Enterprise Architect can be a time consuming process. Like when you want to export all elements from a diagram (not a node in the tree). And when you want specific data like tagged values to be exported as well. Here is a little trick that can save you some valuable time.

Bonus Chapter 2: Still more various extra tips

173

Open the diagram in list view, apply the previous tricks to get the right data in the list. Then select the first row, press shift and select the last row.

Screenshot 3-1

Now press ctrl-c, switch to the spreadsheet and press ctrl-v. The result will be a nicely formatted table with exactly the data you wanted.

Screenshot 3-2

Note: depending on the spreadsheet settings, you might not get a properly formatted table in the spreadsheet application. In that case, paste the data in a text processor like

Bonus Chapter 2: Still more various extra tips

174

Notepad. Save the file as a csv file and import it into the spreadsheet application.

Trick 4: really advanced search Underneath the graphical user interface of Enterprise Architect, there is a database. Wouldn’t it be fantastic if you could query this database from within the user interface? Well, you can. Here’s how: Step 1: press ctrl-f to open the search window. Then click the builder button (the one with the blue triangle).

Screenshot 4-1

Step 2: switch the tab from “search builder” to “sql” by clicking the sql tab

Bonus Chapter 2: Still more various extra tips

175

Screenshot 4-2

Step 3: start typing “select * from”. Bring up the list of tables by clicking ctrl+space and choose a table, for example “t_authors”

Bonus Chapter 2: Still more various extra tips

176

Screenshot 4-3

Step 4: click “Run SQL” or F5. The list of currently defined project authors appears.

Extra tips • Since this result list is a normal Enterprise Architect list, you can select some items, press ctrl+c, open up a spreadsheet and paste them in the spreadsheet with ctrl+v. • Relating what you see in the table list and what is there in the user interface is tricky. For example, tagged values related to elements are not stored in the t_taggedvalues table, but in the t_objectproperties table. Thomas Kilian has published a valuable leanpub book about querying Enterprise Architect. You can find it here: http://leanpub.com/InsideEA

Bonus Chapter 2: Still more various extra tips

177

Trick 5: convert a scenario to a diagram - automatically! How often are you pondering “should I model this as a diagram, or present it as a text”? The answer is, of course: it depends. It depends for a large part on the demands of the audience. Some people prefer to read text, other immediately grasp the meaning of a well-structured diagram. Thanks to a nice Enterprise Architect feature, you don’t even have to choose. You can please your IT colleague, who wants the details in the form of text, and your manager who never has the time to read (but inspects any diagram with great care). Here is how it works. Suppose we are modeling the functionality of the Vinimble website with use cases. Open the properties of a use case and choose “Rules - Scenarios”. Then click the tab “Structured specification”.

Bonus Chapter 2: Still more various extra tips

178

Screenshot 5-1

Now gradually add the scenario steps. When the scenario is complete, press the “Generate Diagram” button and the scenario will appear as a diagram:

Bonus Chapter 2: Still more various extra tips

179

Screenshot 5-2

Extra tips • Next to the “Generate Diagram” button, there is a dropdown button that allows you to choose the type of diagram. Default is a simple activity diagram, but other options include ruleflow, state and robustness diagrams. • For each scenario step, you can indicate whether it is a human being who carries out the step, or a

Bonus Chapter 2: Still more various extra tips

180

computer. Just right click the icon in front of the step and choose the appropriate option. • You can add alternate scenarios and exceptions as well. If you add steps to these alternate scenarios, they will show up as decision diamonds in the generated diagram.

Trick 6: how to make structured text from diagrams Once you have run the diagram export as explained in tip 5, you can choose if you want to work on the text or on the diagram. Enterprise Architect has this nice little feature that allows you to change the diagram you generated from a structured scenario, then reimport this diagram into the structured text. Here’s how to do that.

Bonus Chapter 2: Still more various extra tips

Screenshot 6-1

First, make a change to the generated diagram.

181

Bonus Chapter 2: Still more various extra tips

182

Screenshot 6-2

For example, let’s add a step. If you are done, save the diagram.

Bonus Chapter 2: Still more various extra tips

183

Screenshot 6-3

Then, open the structured specification editor and right click. Choose “Create structure from generated activity diagram”. The validation window will open, and Enterprise Architect asks to confirm that you want to overwrite the old structured specification. Click “yes” and the step has been added to the structured specification!

Bonus Chapter 2: Still more various extra tips

184

Screenshot 6-4

Alas, this won’t work if you make up your own diagram instead of generating it from a structured specification. So the trick is to generate a diagram as fast as possible, then continue working on the diagram, regularly going back to the structured specification to reimport it.

Trick 7: Importing visio drawings During my work as a consultant for Vinimble, the chief editor told me someone already studied the “Call for Articles” process. “I’ll send you the Visio diagrams,” he said. And he kept his promise. Would there be some clever way to import these diagrams in Enterprise Architect, saving me time from redrawing them? Yes there is! First, we’ll find and install the Visio

Bonus Chapter 2: Still more various extra tips

185

importer on the SparxSystems website. Then, I’ll show you an example of how to work with it. And last, I tell you how the importer works for three frequently used Visio stencils: the Basic Shapes, the Basic Flowchart Shapes, and the UML Activity Diagram. Importing from diagrams made with other Visio stencils behaves the same way as these three, so you can experiment with it yourself to find out what works and what not.

Finding and installing the Visio Importer Finding the Visio Importer on the SparxSsytems website is not trivial, even if you know it exists and how it is called. Here are the links: Link to the products page: http://www.sparxsystems.com/products/#link Direct link: http://www.sparxsystems.com/bin/VisioImporter.exe

An example of Visio imports First step: install the MDG Link for Microsoft Visio.

Bonus Chapter 2: Still more various extra tips

186

Screenshot 7-1

Second step: this is the Visio process model we will import, as seen in the Visio GUI:

Bonus Chapter 2: Still more various extra tips

187

Screenshot 7-2

Third step: In Enterprise Architect, select “Extensions |

Bonus Chapter 2: Still more various extra tips

188

Import Visio Documents…”.

Screenshot 7-3

Fourth step: choose a default model. If you want to deviate from the suggested default Diagram Type, Object Type and Connector Type, choose “Custom…”.

Screenshot 7-4

Fifth step: take a look at the report after the import is finished. This will tell you how Enterprise Architect interpreted the Visio model.

Bonus Chapter 2: Still more various extra tips

189

Screenshot 7-5

Sixth step: the final result. As you can see, most elements have been translated correctly, be it that only the appearance is correct (color, shape, position and connections). If I want the element type to be correct as well, I need to instruct the creator of the Visio diagram to work with an other Visio stencil (UML Activity) or I can tweak the results

Bonus Chapter 2: Still more various extra tips

190

after import.

Screenshot 7-6

Basic shapes Basic shapes from Visio are not recognized by Enterprise Architect. Importing shapes like squares, rectangles,… results in an element of a default type (for example, “activity” when the type of diagram is “activity diagram”) with a stereotype that is equal to the Visio shape name. Spaces

Bonus Chapter 2: Still more various extra tips

191

in the shape name result in underscores. The exact result depends on the stereotypes present in Enterprise Architect. By default: • Visio circles, squares, rectangles and triangles result in transparant Enterprise Architect elements with a standard border • Rounded rectangles: stereotyped “rounded_rectangle” results in a filled Enterprise Architect element with the stereotype and the element name shown.

Basic flowchart shapes • Process and decision seem to be recognized and get the correct Enterprise Architect element type (an “activity” and a “decision” respectively) • Other visio shapes (Document, Start/End, Subprocess, Database, Data, External Data) are not recognized and get a default element type with a stereotype that reflects the visio shape name, spaces becoming underscores.

UML Activity • Swimlanes become boundaries, resulting in a correct looking diagram • ActivityInitial, ActivityFinal, State and Decision from Visio are translated properly into Enterprise Architect elements.

Bonus Chapter 2: Still more various extra tips

192

• Action States from Visio become Activities in Enterprise Architect • Transition (Fork) and Transition (Join) get lost and the importer window shows an error message • Signal Send, Signal Receipt, Object in State and Constraints are converted to Activity in Enterprise Architect An example of a simple Activity Diagram with two swimlanes, an activityinitial, an activityfinal, two activities and a decision, imported from a Visio model based on the UML Activity stencil.

Bonus Chapter 2: Still more various extra tips

193

Screenshot 7-7

Trick 8: Links in notes Most Enterprise Architect users have noticed the hyperlink icon that appears in the toolbar on top of every notes field. Those who have tried it, surely have wondered what

Bonus Chapter 2: Still more various extra tips

194

it does. Nothing, most of the time. Unless you link to elements, or when you start exporting the model to rtf (Microsoft Word) or html (website).

Screenshot 8-1

Open an element, type some text in the notes field and select a part of this text. Then click the hyperlink button (the globe with a keychain on top of it). A dialog window appears that allows you to choose a link type and some parameters. Most of these links do not work in the Enterprise Architect GUI itself. Some links work when output is generated, either to rtf either to html. This is true only if the target is also included in the output, naturally.

Attribute Links to an attribute of an element. Only works in output and insofar the target attribute is present in the output.

Bonus Chapter 2: Still more various extra tips

195

Diagram Links to a diagram. Only works in output and insofar the target diagram is present in the output.

EA Command Links to an EA Command. Makes sense only in the EA GUI, but it does not work.

Element This is the only link that actually works in de Enterprise Architect GUI. When you hover over the element link, three buttons appear. • Open element properties: the leftmost button opens the target element’s properties window on top of the source element’s properties. • Find element in diagrams: the middle button opens a dialog with a list of all diagrams in which the target element appears. You can then double click a diagram to open it. • Look up element in tree: the rightmost button looks up the target element in the tree and brings the focus to this element.

Bonus Chapter 2: Still more various extra tips

196

File Links to any file. Only works in output and insofar the target file can be reached from the location of the output file.

Help Links to the EA Help. Makes sense only in the EA GUI, but it does not work.

Matrix Links to any EA Matrix Definition. Makes sense only in the EA GUI, but it does not work.

Operation Links to an operation of an element. Only works in output and insofar the target operation is present in the output.

Package Links to a package. Only works in output and insofar the target package is present in the output.

Search Links to any EA Search Command. Makes sense only in the EA GUI, but it does not work.

Bonus Chapter 2: Still more various extra tips

197

Team Review Links to a post in the Team Review. Makes sense only in the EA GUI, but it does not work.

Website Links to any website or webpage. When generating html output, it depends on the browser and its settings if the target website can be opened in the frame.

Trick 9: Operations as activities Analysts add operations to classes to document their behaviour. For example, I have the class “PublicityCalculator” that will store data about the publicity that appears in Vinimble. Some of its operations are intended to calculate the ROI and to estimate the total printing costs of the magazine.

Bonus Chapter 2: Still more various extra tips

198

Screenshot 9-1

Now I would like to indicate visually that some of these operations are linked. For example, I need the total publicity figures to calculate the ROI for the magazine. To achieve that, I just add an activity diagram and drag the operations from the tree to this diagram. Enterprise Architect will then add an Action to the diagram (and to the model tree). This way, I have a visual representation of the behaviour of this operation. For instance, I could use it to graphically represent the calculation procedure, or to link it to other elements in the three.

Bonus Chapter 2: Still more various extra tips

199

Screenshot 9-2

When inspecting the properties of the action, you will find an action pin tied to the action. This action pin is called “target” and it’s behaviour links to the operation. Starting from a diagram, I can also find out what operations are linked to which actions. Just select the action, right click it and choose “Find | Locate Operation in Project Browser”.

Bonus Chapter 2: Still more various extra tips

200

Screenshot 9-3

Extra tip Roughly the same holds for attributes that belong to a class. Only they are represented as an object instead of an action.

Trick 10: Working with working sets Suppose you are working on two different aspects of the same project. Of you fulfill two different roles within the team, so you need two “views” on your model. In that case, you can save time with “working sets”. A working set is a collection of diagrams, matrix profiles, searches and items from the team review that can be opened together. Here is how to make a working set:

Bonus Chapter 2: Still more various extra tips

201

• open some diagrams you want to include in the working set. (tip: if you find yourself repeatedly opening the same diagrams, this is a good indication you would benefit from a working set) • click “View | Personal Information” then choose the last tab called “Working Sets”. • click the “Create Working Set” icon (the only one that is active, when no working sets already exist)

Screenshot 10-1

Now you can give the working set a name (like “Business analysis”) and indicate which diagrams should be part of it.

Bonus Chapter 2: Still more various extra tips

202

Screenshot 10-2

Add other The “Add Other…” button shows four options:

Screenshot 10-3

• Diagram: add any diagram from the current model

Bonus Chapter 2: Still more various extra tips

203

• Matrix profile: add a matrix profile (relationship matrix) • Search (see screenshot 10-3): adds a search. You can indicate the type of search (simple, extended,…) and the search parameter. • Team Review: any topic in the team review window.

Using and changing existing “working sets”

Screenshot 10-4

Once you have created a working set, you can start using it. Right click the working set and choose one of these actions:

Bonus Chapter 2: Still more various extra tips

204

• Open working set: opens the diagrams and other views that are part of this working set • Create working set and Create working set from: to create new working sets, possibly based on an existing working set or a set of open diagrams. • Edit: edits the current working set (to add or remove diagrams or views) • Copy as New: to copy the current working set, so you can make a new one based on this • Add Active View: if a diagram or view is currently open that does not already belong to this working set, this menu option becomes available to add the diagram or view to the current working set. • Apply when Model opens: this toggle decides if the current working set will be opened when the model is opened. If the toggle is on, the button becomes green. • Store Main Tab View History: also a toggle. When active, captures which diagrams and views are open when you close Enterprise Architect. Upon reopening, creates a read-only working set that contains these diagrams and views. • Locate in project browser (only available if the active item is a diagram): looks up the diagram in the project browser and puts focus on it. • Delete Working Set Item (only available if the active item is a diagram or a view): deletes the item from the working set.

Bonus Chapter 3: Building your own toolbox For the tricks in this chapter, we will build on trick 32 in chapter 4. A quick review of the steps necessary to build an elementary toolbox: • create a package stereotyped “profile” and add a class diagram to it • add metaclasses, one per element or connector, add a stereotype and change its name to the name you want to appear in the toolbox, and draw an “Extends” relationship from the stereotype to the masterclass • save the package as a UML profile and import it via the Resources pane (UML Profiles - right click and choose “Import profile”) • open the toolbox via “More tools” Before you start reading the rest of this chapter, make sure you know these steps by heart, since you will find yourself repeating them over and over. Especially the export-import steps.

Bonus Chapter 3: Building your own toolbox

206

Trick 1: the order of the elements in the toolbox Wouldn’t it be nice if you could determine what order the elements in the toolbox would have, so that you can put the most frequently used ones on top? Or such that you can put them in some “logical” order, for example ordered by level of abstraction? There are two ways to determine the order of the elements in the toolbox. The first one is the better one, since you only have to do this once. The second is useful if you want to play with different order options without having to importexport the whole package over and over again.

Method one: stereotype Z-order Every element or connector stems from a “stereotype” element in the generating diagram. The Z-order of the stereotypes determines in which order they are exported and therefore, the order of the elements in the resulting toolbox. Adapting the Z-order is easy: take the stereotypes one by one, starting with the element you want on top, right click the element and choose “Z-order | send to bottom”.

Bonus Chapter 3: Building your own toolbox

207

Screenshot 1-1

Method two: the xml file The generated xml file is nothing more than an text file with xml markup tags. Open it with a text editor, then manually cut and paste the element sections in the order you want them to appear in the toolbox. Make sure you don’t inadvertently scramble the xml code, because this will most likely render the file invalid and you’ll have to regenerate the file.

Bonus Chapter 3: Building your own toolbox

208

Screenshot 1-2

Trick 2: setting a default element size Probably you will want to deviate from the default sizes elements get when they are created. Suppose, for example, that we have an element that “hosts” other elements. Or you want to visually distinguish elements based on their shape. It is good to know that you can include default element sizes in the profile definition.

Screenshot 2-1

Bonus Chapter 3: Building your own toolbox

209

Setting the default size is easy if you follow these steps: • Right-click the stereotype (“MyApplication” in this case) and choose “Attributes…” • For the default width of the element, add an attribute “_sizeX” of type int and give it the default value (20, for example). Don’t forget to save the attribute and to press “New” when going to the next step. • For the height of the element, add an attribute “_sizeY” of type int and give it the default value (100, for example).

Screenshot 2-2

Bonus Chapter 3: Building your own toolbox

210

Regenerate the toolbox, delete it from the resources and import it again. Any MyApplication element you now drag from the toolbox to a diagram will have a narrow and high appearance.

Screenshot 2-3

Trick 3: an icon for elements in the toolbox Once your toolbox starts to grow, you will feel the need to have separate icons for the elements and connectors in your toolbox.

Bonus Chapter 3: Building your own toolbox

211

Screenshot 3-1

Here is how you can include them in your toolbox. • add the following attribute Icon: int = icon_name.bmp to the element • also add an attribute called “_metatype: string = metatype name” to the element (do not worry about this metatype for the moment, we will talk about metatypes in a following trick) • put a suitable file in the location of the xml file that contains the toolbox definition. This icon should be a 16x16 bmp file with transparant background (RGB 192,192,192). If you make this file by

Bonus Chapter 3: Building your own toolbox

212

resizing a larger picture, pay attention to the color of the border. It really should be transparant or a border will show. For example, this is a large file with a light grey background that will turn transparant:

Screenshot 3-2

But if you resize the picture, some border pixels take a slightly different color. Correct this in your graphic editor.

Bonus Chapter 3: Building your own toolbox

213

Screenshot 3-3

Delete the old profile, import the new and your icons should all be set up!

Trick 4: working with metatypes In the previous trick, I promised to tell you more about metatypes. Time to do exactly that! Metatypes have following advantages:

Bonus Chapter 3: Building your own toolbox

214

• Default name: new elements derived from the toolbox element will get a new name based on the metatype (for example, MyApplication1) instead of the base type (for example, Object1). This will make it easier for toolbox users to provide a useful name for the element. • Relationship matrix: metatypes can be used in the relationship matrix feature (also see Trick 43). That way, it is far easier to select the right element type if you have multiple types derived from the same base type. Also, the matrix output will be easier to interpret. • Profile image: unless you provide an element type with a metatype, EA will not always apply the visual properties you define for the stereotype. • Icon: when EA creates an project browser icon for a new element, this will be based on the icon you made up for the element instead of the standard base type icon (also see the previous trick).

Screenshot 4-1

Bonus Chapter 3: Building your own toolbox

215

Metatypes are not difficult to define. In fact, it is so easy and so valuable that you should give every element you define in your own toolboxes their own metatype, unless you have very good reasons not to do so. Here is how to define a metatype: • Add an attribute of type “string” named “_metatype” and with initial value the name of the metatype (for example, MyApplication) • Just make sure the name of the metatype does not conflict with an already existing name (such as Object, Process and other names that are already present in the standard EA toolboxes). If you still do so, some of the metatype features will not work. • Regenerate and reimport the toolbox as usual. The metatype should be all in place now!

Trick 5: creating autohiding connectors Remember the “implicit” relationships of trick 18? The idea was that diagrams are simpler to read when relationships are hidden, or at least shown implicitly, by nesting elements inside each other. It’s perfectly possible to have relationships like these in your own toolbox. Just derive the relationship from any of these two base class connectors:

Bonus Chapter 3: Building your own toolbox

216

• composition: “composing” means relating elements such that the composition is a thing on its own, but the composing elements cannot exist without their composition. Like the departments of an organization, for example. A department only exists when it is part of an organization: it cannot exist on its own. UML compositions are shown as black diamonds attached to the parent element. • aggregation: “aggregating” is a more general and “looser” way of relating elements. The aggregated elements make sense on their own. Like the people who form a department: even without the department, they still exist. Aggregated elements are shown as unfilled diamonds attached to the parent.

Bonus Chapter 3: Building your own toolbox

217

Screenshot 5-1

Add a base class with the core connector “composition” checked and a stereotype “MyComposition” that extends it.

Screenshot 5-2

This is the result if you regenerate the toolbox. The diagram shows the relationships explicitely.

Bonus Chapter 3: Building your own toolbox

218

Screenshot 5-3

But when you enlarge the parent object, the composition and the aggregation disappear.

Bonus Chapter 3: Building your own toolbox

219

Screenshot 5-4

Trick 6: how to create an arrow with shapescripting Probably you want your own elements and connectors to deviate visually from the standard EA elements and connectors. One way of achieving that is explained in trick 35, about stereotypes. There’s another way of doing that, however. A more powerful way since it allows you to draw almost anything you could want. That’s when shapescripting comes into play. Some of the next tips will cover more aspects of shapescripting, but this tip teaches you the basics.

Bonus Chapter 3: Building your own toolbox

220

Adding a shapescript to a connector First, add a metaclass Informationflow to the toolbox diagram. Also add two stereotypes: one for an open arrow and one for a closed arrow.

Screenshot 6-1

Now add an attribute “_image” to each stereotype. Click the three dotted “initial value” button. The shapescripting window appears.

Bonus Chapter 3: Building your own toolbox

221

Screenshot 6-2

Click “refresh” to preview the shape, click “next shape” to go to the next subshape that belongs to this shapescript. If you have made an error in the script, clicking “refresh” will give you an error message with the line number of the error. For example, if you type “shaep” instead of “shape”, EA will give this error message: 1 2

Error parsing ShapeScript: There was an error pa\ rsing on line 1: Unexpected symbol: shaep

Code for an open arrow

Bonus Chapter 3: Building your own toolbox

1 2 3 4 5 6 7 8 9

shape main { // draw a dashed line noshadow=true; setlinestyle("SOLID"); hidelabel("MiddleBottomLabel"); moveto(0,0); lineto(100,0); }

10 11 12 13 14 15 16 17 18 19 20

shape target { // draw an arrowhead at the target end rotatable = true; noshadow=true; moveto(0,0); lineto(6,4); moveto(0,0); lineto(6,-4); }

Code for a closed arrow

222

Bonus Chapter 3: Building your own toolbox

1 2 3 4 5 6 7 8

shape main { // draw a dashed line noshadow=true; setlinestyle("DASH"); moveto(0,0); lineto(100,0); }

9 10 11 12 13 14 15 16 17 18 19 20

shape target { // draw an arrowhead at the target end rotatable = true; startpath(); moveto(0,0); lineto(16,6); lineto(16,-6); endpath(); fillandstrokepath(); }

This is the result in the toolbox diagram.

223

Bonus Chapter 3: Building your own toolbox

224

Screenshot 6-3

Regenerate the toolbox and reupload it in the resources panel. This is what you get when you use the connectors described above.

Screenshot 6-4

Trick 7: the shapescripting canvas If you want to work with shapescripts effectively, you will benefit from knowing the shapescripting canvas.

Bonus Chapter 3: Building your own toolbox

225

Screenshot 7-1

For drawing most elements, the bottom right part is where the action happens: when drawing for example lineto(X,Y), both X and Y are positive. For connectors, you will often need the topright part of the canvas too. Y then gets a negative value, like in the example below. We will draw a giant arrowhead.

Bonus Chapter 3: Building your own toolbox

Screenshot 7-2

This is the shapescripting window in action.

226

Bonus Chapter 3: Building your own toolbox

Screenshot 7-3

Code for a giant closed arrow 1 2 3 4 5 6 7

shape main { // draw a line noshadow=true; moveto(0,0); lineto(100,0); }

8 9 10

shape target {

227

Bonus Chapter 3: Building your own toolbox

11 12 13 14 15 16 17 18 19

228

// draw a giant arrowhead @ target end rotatable = true; startpath(); moveto(0,0); lineto(40,30); lineto(40,-30); endpath(); fillandstrokepath(); }

And this is the result when applied in a diagram:

Screenshot 7-4

Using the topleft and bottomleft quadrants of the shapescripting canvas In rare cases, you might need the left side of the quadrant. The shape, mostly a connector, will then overlap the element it points to. This hourglass shape is an example.

Code for an hourglass shaped connector arrow

Bonus Chapter 3: Building your own toolbox

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

229

shape main { // draw a line noshadow=true; moveto(0,0); lineto(100,0); } shape target { // draw an hourglass @ target end rotatable = true; startpath(); moveto(0,0); lineto(10,10); lineto(-10,-10); lineto(10,-10); lineto(-10,10); lineto(10,10); endpath(); fillandstrokepath(); }

This is what it looks like on a diagram. As you can see, the connector overlaps with the element it is attached to.

Bonus Chapter 3: Building your own toolbox

230

Screenshot 7-5

If you want to know which line is which, just comment it out by putting // in front of it and retest the shape.

Trick 8: some shapescripting limitations Shapescripting is a wonderful way of developing your own toolbox, but it has its limitations and it is good to know them up front, so you don’t spend hours to “debug” a script that is actually executing to specifications. In my experience, these are the limitations frequently met in practice: • initialisation attributes must come in front of the script, before the drawing commands • you cannot specify the font and the font color of text used in shapescripts

Bonus Chapter 3: Building your own toolbox

231

Initialisation attributes up front There are three commands that specify how the shape should be drawn: • noshadow: if true, the shape is drawn without a shadow. The default value is “false”. • h_align: specifies where label text should be put horizontally. Valid values: left, center or right. • v_align: specifies where label text should be put vertically. Valid values: top, center or bottom. An example script shows what I mean: 1 2 3 4 5 6 7 8

shape main { noshadow = "true"; h_align = "center"; v_align = "top"; rectangle(0,0,100,100); println(#NAME#); }

This is the result when applied to the existing application shapes:

Bonus Chapter 3: Building your own toolbox

232

Screenshot 8-1

What to do if you want some text, let’s say the name of the shape, to appear on the bottom of the shape, while another label, such as the stereotype, is put on top? Then you will have to work with subshapes. An example script shows you the way: 1 2 3 4

shape main { rectangle(0,0,100,100); addsubshape("stereotypeshape", 100, 80); addsubshape("nameshape", 100, 80);

5 6 7 8 9

shape stereotypeshape { h_align = "center"; println("#stereotype#"); }

10 11 12 13 14 15

shape nameshape { h_align = "center"; println("#name#"); } }

This is the result:

Bonus Chapter 3: Building your own toolbox

233

Screenshot 8-2

Trick 9: hiding connector labels Connector labels often make diagrams hard to read. By default, some labels like stereotype names are shown. How to hide them? There is a shapescripting command that hides the labels: hidelabel(“LabelName”). These are the labels you can hide: • • • • • •

middletoplabel middlebottomlabel lefttoplabel leftbottomlabel righttoplabel rightbottomlabel

Each of these labels are detached, meaning you can change their position on the diagram. Their default position is described by their name.

Bonus Chapter 3: Building your own toolbox

234

Screenshot 9-1

How can you know which labels are visible? Make a diagram where the connector appears. Right click the connector and choose “Visibility | Label Visibility…”. A dialog appears:

Bonus Chapter 3: Building your own toolbox

235

Screenshot 9-2

Let’s visit our hourglass shaped connector again and try to hide the MyHourglass stereotype label:

Code for an hourglass shaped connector arrow without the stereotype label 1 2 3 4 5 6 7

shape main { // draw a line noshadow=true; hidelabel("MiddleBottomLabel"); moveto(0,0); lineto(100,0);

8 9 10

} shape target

Bonus Chapter 3: Building your own toolbox

11

{ // draw an hourglass @ target end rotatable = true; startpath(); moveto(0,0); lineto(10,10); lineto(-10,-10); lineto(10,-10); lineto(-10,10); lineto(10,10); endpath(); fillandstrokepath();

12 13 14 15 16 17 18 19 20 21 22 23

236

}

Remember to put this drawing command after the initalisation attribute “noshadow”. The result is this.

Screenshot 9-3

Even when you try to make the label visible by checking the visibility checkbox, it will not reappear: the shapescript takes precedence.

Bonus Chapter 3: Building your own toolbox

237

Notes • There is a drawing command “showlabel” with comparable use but the opposite effect. • Depending on the version of EA, it is important to write the label names exactly as shown in the example (capital first letter for each word)

Trick 10: what to do when an MDG Technology is lost MDG Technologies (Model Driven Generation) are a fantastic way to customize Enterprise Architect to your needs. In short, you create your own package that contains: • several profiles, for example, the elements and connectors we created in this bonus chapter • a toolbox that hosts these profiles, so you can easily use them on your diagrams • one or more diagram types tied to the toolbox, so that the right toolbox opens every time you create a diagram of this type. The MDG Wizard helps you to create an MDG technology that contains these items (and possibly more). You can find it here: Tools | Generate MDG Technology file”. The wizard will result in an MTS file, that you import via “Tools | MDG Technology Import”.

Bonus Chapter 3: Building your own toolbox

238

Screenshot 10-1

From time to time, you will see that the full MDG technology gets lost. It appears to be beyond repair. Or is it? This little trick seems to work. Open the UML Types dialog (“Settings | UML Types…”). Then check the stereotypes you created. Probably you will see that the shapescripts got deselected. Reselect them, then regenerate the MTS file with the MDG Technology wizard and the MDG technology will work again!

Bonus Chapter 3: Building your own toolbox

239

Screenshot 10-2

Note Working with MDG Technologies and creating own diagram types, toolboxes and the like will be far easier in EA Version 10.x When this book gets updated once version 10 is released, we will talk about the new functionality in more detail.

Overview of shortcuts Copy the shortcut page, cut out the rectangle, fold along the vertical lines and glue along the dotted line so you have a tall pyramid with three sides. Put it next to your computer and use it until you know the shortcuts by heart. Don’t forget to make a copy for your colleagues too! This file is available for download at the website³. ³http://www.aboriginemundi.com/index.php/2010/12/shortcuts-forenterprise-architect-free-download/

241

Overview of shortcuts

shortcuts