Next: 4. Data View Editors Up: Toolkit for Conceptual Modeling Previous: 2. Document Editing

Subsections

3. Diagram Editing

3.1 Definitions

Diagrams that are made by some TCM diagram editor are a special kind of graph with a certain representation (layout, view). Each TCM editor is a specialization of the generic graph editor. A graph consists of a set of subjects. A subject is either a node or an edge that connects two subjects. Subjects are of different node types and edge types. For example, a TDFD data flow diagram graph consists amongst others of nodes of type data process and nodes of type data store, and edges of type data flow.

The graphs that are edited with TCM contain edges that connect exactly two (not necessarily different) subjects. So some kind of special relationship between more than two nodes, for instance the specialization relationship in an ER diagram (made by the TESD editor), is specified in TCM as a node (a taxonomy junction node in this case) that is connected by three or more different edges (generalization or is-a relationship edges in this case).

In general, edges connect two (not necessarily) different subjects. So in principle it is possible that an edge connects to another edge. This feature is available in the generic diagram editor (TGD) and also in a limited form in a few of the specific diagram editors. In the other editors an edge connects only nodes, it never connects edges.

In a representation of a graph, nodes and edges are shown as shapes. In the case of nodes the representing shapes are boxes, diamonds, ellipses etc. In the case of edges, shapes are lines, arrows etc. In general, there is a many-many relationship between a certain subject type and a certain shape type and also between a particular subject instance and a particular shape instance. It is possible that a subject type is represented by more than one shape type. For instance a class-diagram made by TSSD can contain different kinds of class boxes (e.g. boxes with three compartments that show the class name, attributes and operations, and boxes that only contain two compartment to show only the name and the attributes). Also, it is possible that a particular subject instance is represented by more than one shape instance. This feature can be used when you want to represent the same subject at different positions in the same diagram (duplication). But in most cases, the relationship between subjects and their representations is one-one and for this reason, subjects and their shape representations will often be mixed up.

A TCM diagram is composed of a graph, its representing shapes, and extra diagram information, such as the diagram name, the author and the creation date. A TCM diagram has an empty graph after creation. You can build a graph by interactively drawing it with the TCM user interface, node by node, edge by edge.

Some diagrams are hierarchic. This means that the underlying mathematical structure is a hierarchic graph rather than a graph. In a hierarchic graph, nodes are organized in trees: a node may be a child of another node. The parent-child relationship is represented by inclusion: a node shape contains the node shapes of its children. (In hierarchic documents, duplication is forbidden. This ensures that the parent-child relationship is acyclic.) In non-hierarchic diagrams, overlapping of nodes has no special meaning.

TCM documents should satisfy certain constraints. Most constraints are specific for the particular diagram technique supported by the editor. Many constraints are checked by TCM. The constraints that are checked, are classified according to how TCM deals with possible violations.

1.

Built-in constraints. These are constraints which are built-in and which can never be violated because there is no command in the user interface to achieve that. We give two examples of built-in constraints: 1. an edge connects exactly two (not necessarily different) existing subjects, 2. an entity type in TESD is always represented by a box or a double box.

2.

Immediately enforced constraints. These are constraints that are immediately enforced by TCM if that is possible. When you perform a command that would violate a constraint that is immediately enforced, this command is rejected immediately by TCM and a pop-up window with an error message is displayed. Examples of immediately enforced constraints are: 1. an entity type and a relationship node in TESD cannot be connected by a Generalization and 2. two entity types cannot have the same (non-empty) name.

3.

Soft constraints. These are constraints that can be violated. Soft constraints are needed because the TCM user interface demands that you draw only one node or one edge at the time, and only edit the label of an existing subject. So there are certain commands that produce diagrams that might violate a constraint and that can not be immediately enforced. Two examples of soft constraints are: 1. all entity types have a non-empty name and 2. a relationship node in TESD should be connected to entity type nodes by two or more different edges.

Soft constraints are checked by TCM when the Check Document command is issued by you. Check Document displays a list of error messages in a pop-up window and the nodes and edges that cause these errors are selected in the diagram. As opposed to the previous two classes of constraints, you are responsible for correcting the diagram.

It is possible that some immediately enforced constraints can still be violated by some user command. Such a constraint is then classified as both an immediately enforced and a soft constraint. Example: if a part of a diagram is copied and pasted into the same diagram, the unique name constraint is a soft constraint as it can be violated by the paste command. But when you have edited the text of one of the nodes in the diagram, and you issue the stop edit command, the unique name constraint can and will be immediately enforced.

3.2 Creating Nodes

The set of tiled buttons labeled Nodes, at the left side of the main window, contains radio buttons indicating the current type of node that can be created. You can change the current node type selection by clicking another tiled button with button-1.

For creating a node, first choose the desired node type from the node tiles. Click button-1 in the background of the drawing area, at the desired position. A new node shape appears, having its center at the position where is clicked and having the shape as indicated by the selected node type in the tiles. The node shape is surrounded by a set of selection handles which are black or grey rectangles, showing that the shape is selected. Selection handles of a node shape are also used for resizing the node.

3.3 Creating Edges

The set of tiled buttons named Edges contains radio buttons indicating the current type of edge that can be drawn. Furthermore there is a so-called check button labeled create curve. If the curve button is on, new edges that are created are drawn as a Bezier curve. If the curve button is off, a multi-segment line is drawn as a set of connected straight line segments. To create a new edge, first choose the desired edge type in the edge tiles. There are three sorts of edges that you can create:

Straight edges. Drag with button-2 from a node to a node. This means you push and hold down button-2 somewhere inside the source node and move the mouse with button-2 still pushed down. During dragging the mouse pointer turns into a $\epsfig{figure=p/sightP.ps}$ and a line from the source node follows the movement of the mouse. You can always abort creating an edge while dragging by clicking button-1. When you are arrived in the target node and release button-2, then a straight edge is created. The two points of the edge at the borders of the connected nodes are called the end points of the edge.
Segmented edges. These edges consist of more than one segment. The points where the segments are connected are called intermediate points. To create such a segmented edge, release button-2 in the background while you are dragging. Then the first intermediate point is created. After that, the line follows the mouse all by itself. Then you can add another intermediate point by clicking button-2 somewhere else in the background. You complete the edge by moving into the target node and clicking button-2.
Curved edges. You can also create an edge drawn as a Bezier curve ^3.1. For this, turn on the create curve toggle button in the tiled menu. The Bezier curve requires exactly two intermediate points, which makes four edge points in total. These intermediate points are added in the same way as creating normal multiple line segments as described for segmented edges. When the four points are determined, a smooth curve is drawn between the end points and the two intermediate points act as controlling points. You can toggle between the curved and segmented edge representation by invoking the Convert From/To Curves command from the Edit menu.

**Figure 3.1:** Three forms of edges.
$\includegraphics{p/createedges.eps}$

After creating the edge, the edge is selected which is visible by a selection handle on each line point. You can adjust the line points by moving the handles; see section 3.6.

When an edge is created and it is the only straight edge connecting this particular couple of nodes, it will go (virtually) through the center points of the nodes. When there are multiple straight edges connecting the same pair of nodes, they will be equally distributed by default over the opposing sides. Segmented and curved edges are not equally distributed when they connect the same pair of nodes. TCM tries to intersect the first and last segment of a segmented line orthogonally with the sides of the connected nodes. If this is geometrically impossible, the segment will be directed to the center of the node like with straight edges. For understanding how it works it helps when you try it out yourself.

Tip: When you want to enforce that a straight edge is always orthogonal to a node, then you can draw a segmented line of three points that looks like a straight line. You enforce then that the line is always connected orthogonally with both of the nodes (providing that this is geometrically possible). This gives often a more pleasant optical result especially when you want to connect two nodes that differ a lot in size.

It is possible to move the end points of an edge as well, in the same way as you move the intermediate points. When you drag with button-1 the handle of an end point to another position then the new end point will be the position at the border of the node shape that is the closest to where you released the handle. However, when you drag with button-1 the handle of an end point into a different node shape then the edge will be redirected to that node. This means that you can change the nodes that an edge connects.

After an edge is created you can add extra intermediate points. When you drag with button-1 on an existing line, a new intermediate point is created at the position where you stop dragging, and a new segment will be added to the line. Press Button-2 on a line handle to remove the line handle/segment from an existing line.

In TGD and some other editors it is possible to connect an edge with an edge like in figure 3.2. You create these kinds of edges in the same way as edges that connect nodes. Only you start to press button-2 while the mouse pointer is positioned at an edge and/or you release button-2, after dragging, when the mouse is on some edge.

**Figure 3.2:** An edge connects another edge.
$\includegraphics{p/edgeedge.eps}$

3.4 Selection Commands

The current selection is a subset of the displayed shapes. Shapes in the selection have visible selection handles. The shape in the selection that was first selected has black selection handles and the other shapes in the selection have grey selection handles.

Shapes have a selection area. For lines, it is a band around the segment(s) of the line. The selection area for node shapes changes when you toggle hierarchy on or off: In non-hierarchic documents, it is just the inside of the node shape. In hierarchic documents, the selection area is a band around the visible parts of the node shape (boundary lines, interior lines, text labels).

The following operations on the selection exist:

Select a single shape. Click button-1 in the selection area of an unselected shape (node shape, line). The selection will contain only this shape which will show black handles.
Add a shape to the selection. Click button-2 in the selection area of an unselected shape. The first selected shape will show black selection handles, the other selected shapes will show grey selection handles.
Remove a shape from the selection. Click button-2 in the selection area of a selected shape. The selection handles disappear.
Empty the selection. Click button-2 somewhere in the background. All selection handles disappear.
Select a part of the diagram. Drag button-1 starting in the background. The area that you are selecting, is enclosed by a stippled box, and its bottom-right corner is attached to the mouse. If you release button-1 then every shape that was (partly) inside the stippled box is selected and every other shape outside the box is unselected. One of these selected shapes will show black selection handles, the others will become grey.
Make selected shape first selected. You can enforce that a selected shape is the first selected by clicking button-1 in the selection area of an already selected (grey) shape. The selection handles of the old first selected shape will become grey and the clicked shape's handles will become black.
The Select All command from the Edit menu selects all shapes of the diagram. By clicking button-2 in the background you deselect all shapes.

3.5 Editing Text

For getting into edit mode make sure that the shape with the label that you want to edit is the only selected shape. When you move the mouse pointer into the single selected shape, the mouse pointer turns into a $\epsfig{figure=p/ibeamP.ps}$ . By clicking button-1 on the shape or typing the first character you go into edit mode. See section 2.5.1 for the different text edit commands that exist in TCM. These apply to all document editors including all diagram editors.

Every TCM diagram and tree editor has a special node type called Comment, which solely consists of a text label and is intended to add comment to the diagram. This node type cannot be connected by any other edge, except in the generic diagram editor, TGD. But comment nodes can be selected, moved and of course edited, like any other node type.

**Figure 3.3:** Example diagram with some text labels.
$\includegraphics{p/labelsexample.eps}$

Tip: If you want to edit the label of an edge, and you try to select it first with button-1 you could miss the edge and create an unwanted node. To avoid this, make the selection empty and select the edge by button-2 instead, because, when you miss no node is created.

3.6 Moving Shapes

You can reposition the following things in a diagram:

A node shape. You can move a node shape by positioning the mouse into that node shape and then drag it with button-1. The shape does not need to be selected. A grey node outline moves along with the mouse. The mouse pointer turns into a $\epsfig{figure=p/movingP.ps}$ . If you release button-1 then this node shape is redrawn at the new position and all its adjacent lines are redrawn too.
The selection. If you drag one of the selected node shapes, all other selected shapes are moved with over the same distance, including the line handles of the selected edges.
Intermediate line handle. You cannot move the line as a whole, but you can drag the individual handles of a line. When a line is selected and the mouse is in one of the handles (which is indicated by a $\epsfig{figure=p/movingP.ps}$ ), then you can drag with button-1 this handle and a grey outline of the new line follows the mouse. You can move the intermediate handles of a line to an arbitrary place in the drawing area.
When you drag a line outside a line handle with button-1, a new line handle will be created and the line is extended with an extra segment. When you hit a line handle with button-2 (instead of button-1), the line handle is deleted and consequently the line will be deprived of a segment.
End line handle. When you drag and release one of the end point handles with button-1 into the connected node or somewhere in the background, then the end point will be placed at the border of the connected node at a position that is the closest to the place where the handle was released. However, when you drag the handle into a different node then the edge will be redirected to that node. The result is that the edge now connects to a different node ^3.2.
The label of an edge. You can move the label of an edge by dragging it with button-1. The edge need not be selected. You cannot move the label of a node; when you move or resize a node, the label positions inside the node are recalculated, as well as the labels of the adjacent edges.
The Align nodes command in the Edit menu has six variants as you can see in the Align Nodes submenu of the Edit menu. See figure 3.4 for the various alignments. The alignment takes place according to the position and size of the first selected shape, which should be a node shape. Depending on the chosen alignment type, the rest of the nodes in the selection will be aligned to the first shape, either horizontally (same center y-coordinates), vertically (same center x-coordinates), to the top (same top y-coordinates), to the bottom (same bottom y-coordinates), to the left (same left x-coordinates) or to the right (same right x-coordinates).

**Figure 3.4:** TCM Node Alignment Dialog.
$\includegraphics[width=3in]{p/nodealignmentdialog.ps}$

When you want to abort a move command, while you are moving, you have to click button-2.

3.7 Resizing Shapes

Resizing a node shape in all directions is possible by dragging button-1 on the selection handles. The mouse pointer turns into one of the eight symbols that indicate to what direction the node shape can be resized: $\epsfig{figure=p/8sizeP.ps}$ . You can use button-2 to resize a node shape in the desired direction only. When you want to abort the command while you are resizing, you have to click button-2.

Unlike moving, resizing only works on one shape at the time i.e. it does not influence the other shapes of the selection. If you want to resize multiple shapes in one command then you could use the Same Size command in the Edit menu. The Same Size command makes the size of all selected node shapes the same size as the first selected shape (the shape with the black handles). The first selected shape should be a node shape of course. The edges in the rest of the selection are ignored by this command.

Tip: when node shapes have been moved or resized then their adjacent edges are redrawn and all their handles and labels are placed back to their default positions. This is sometimes undesirable ^3.3. To avoid that your effort gets lost, postpone moving edge labels and handles to their final positions until when you think that the connected nodes are OK. Draw the graph first and then you can bother about the layout.

3.8 Deleting Subjects

The edit command Delete (called from the Edit menu, or by the accelerator <Ctrl+D>), removes all subjects from the selection. If a node is deleted then all its adjacent edges are deleted too. If a node has duplicate shapes (see section 3.10) of which some but not all are selected then a question dialog is raised asking whether you want to delete all duplicates (and henceforth the node in the graph) or that you only want to delete the selected duplicate shapes.

The Edit menu command Delete All removes all subjects, selected or not. Before the command is executed, you are first asked by means of a question dialog if you really want to delete everything.

Fortunately, like all commands, deletion commands can be undone with Undo.

3.9 Cutting and Pasting Subjects

The diagram editors have a buffer for cutting, copying and pasting parts of a diagram (selected shapes plus the subjects that they represent) inside the same editor.

You can either Cut (<Ctrl+X>) or Copy (<Ctrl+C>) the selection to the buffer. When you perform a cut, all selected subjects are copied into the buffer, and the selection is removed from the diagram (including the unselected edges that are adjacent to the selected nodes). When you perform a copy, no subjects are removed from a diagram, but a copy of the selection is made and put into the buffer. Both Cut and Copy copy the unselected nodes that are connected by selected edges into the paste buffer.

Pasting means that the contents of the paste buffer is being copied into the diagram. When you perform a Paste (<Ctrl+Y>) command, a stippled box is shown that has the size of the pasted area which is attached to the mouse pointer near its top-left corner. When you click button-1, the subjects in the paste buffer are copied into the diagram. The Append Diagram command is also implemented as a paste command. When you want to abort pasting while you are moving the paste box, you have to click button-2.

You can paste the same contents of the paste buffer more than once because always a copy of the contents is made. Pasted nodes and edges are new nodes and edges, not duplicates of existing nodes or edges. The paste buffer remains intact when you load another diagram, so it is possible to cut and paste between different diagrams (but only in the same editor).

3.10 Creating and Deleting Duplicates of a Node

It is possible to represent one node by several instances of the same shape in non-hierarchic diagrams. The Duplicate command copies the node shapes in the selection and puts them into a box, lust like the Paste command. However, the duplicate command does not make use of the paste buffer. The copied node shapes can be positioned at an arbitrary place, just like the Paste command. Edges cannot be duplicated ^3.4. Unlike the Copy command, the new node shapes have the same node subject as the original. When a node has duplicate shapes then all these shapes have a small asterisk in the top-left corner.

When the label of a duplicate shape is updated then all its duplicate shapes are updated too and when a node is deleted (with Delete) then all the duplicate shapes of this node are deleted too by that command.

The duplicate command is intended to be used when the same node has to be represented at different places to avoid cluttering up the diagram. For instance, in a DFD, the same external entity or data store is often drawn at different positions in the same diagram, and TCM keeps track that these are in reality representations of the same subjects.

The Delete command can be used to delete all selected duplicate shapes, see section 3.8. When all the shapes of a node are deleted then the node itself is deleted too. But you can delete individual duplicate shapes as well.

3.11 Changing Shape Properties

All Update shape property commands are undo-able commands and they are called from the Properties menu of the main window.

The Update Line Style command from the Properties menu pops up a dialog window with a number of toggle buttons for the different line styles: solid, dashed, dotted, dual, invisible and wide dotted (see figure 2.9). Select the line style that you want and press Apply and then the line styles of all the selected shapes are updated to this style.
The Update Line Width command from the Properties menu shows a dialog window with a number of toggle buttons for the different line widths. The line width icons range from 1 pixel to 9 pixels wide, and are used to set the line widths of shapes. When you press Apply then of every selected shape the line width is set to the chosen line width.
Update Line Ends. This entry pops up a dialog window to set the line ends of a line. See figure 3.5. You can change the 'begin of a line' as well as the 'end of a line'; for each line end a set of tiled toggle buttons is displayed, each containing a bitmap symbol for the corresponding line end. When you issue a command to change the line ends, each selected line is redrawn according to this new line ends. This command is currently only available in the generic diagram editor (TGD).

Figure 3.5: TCM line ends dialog.

$\includegraphics[width=4.5in]{p/line_endsdialog.ps}$
The Update Text Font command from the Properties menu pops up a dialog window in which you can select a font (see figure 2.11). A font consists of the attributes font family, font style and point size. For each of the three attributes there is a list of toggle buttons. Also, each list of toggle buttons in the dialog has an extra check button called update attribute that indicates whether that font particular attribute should be updated or not. This makes it possible, for instance, to only change point sizes or font families of some texts but to keep the other font attributes the same.
The dialog also shows a preview of some text in the selected font so you can see how it will look in your diagram. When you press the Apply-button then the dialog is dismissed and of each selected shape, the font is updated to the selected font.
With the Default Text Font entry from the Default Properties submenu you get a similar dialog window. Here you can set the default text font. The text shapes of every newly created node or edge will get this font. The page headers and numbers are also drawn in this default font.
The Update Text Alignment command from the Properties menu shows a dialog window with three toggle buttons, left, center and right. Choose the alignment and press Apply. Then the text alignments of all the text shapes in the selected shapes are updated to the selected alignment. Note that this alignment determines only the alignment of texts of multiple lines, not the positioning of the text shape itself in or near its parent shape.
The Set/Unset Text Underlining command from the Properties menu sets/unsets (toggles) the text underlining of the selected shapes.
The Update Line Color command from the Properties menu pops up a color chooser dialog (see figure 2.13). The color chooser shows the names of all the available colors in a scrolled list. With clicking button-1 in the list you can select a color name. The associated color is made visible in a label below the scrolled list, named preview. When you press the Apply button then the line colors of all the selected shapes are updated to this color.
The Update Text Color command from the Properties menu pops up a color chooser dialog. This works the same as the Update Line Color command except that after pressing Apply the colors of all text shapes of the selected shapes are updated to the chosen color.
The Update Fill Color command from the Properties menu pops up a color chooser dialog. This works the same as the Update Line Color command except that after pressing Apply all selected shapes are filled with the chosen color.

**Figure 3.5:** TCM line ends dialog.
$\includegraphics[width=4.5in]{p/line_endsdialog.ps}$

3.12 Miscellaneous Edit Commands

The Node/Edge annotation command from the properties menu pops up a text edit dialog in which you can type arbitrary text to annotate the subject. See section 2.5 for using the text edit dialogs.
With the Find menu entry in the Search menu you call a dialog in which you can search for some text string in the diagram. The find dialog is described in section 2.5.3. From this dialog you can find the next text or find all texts that matches the string to find. In the first case, the first shape that is found is selected and the scrollbars of the main window are moved to center the shape in the main window. When you click Find Next again, the next shape is selected (round-robin). When you choose Find All, all shapes that contain a string that matches are selected.
With the Replace menu entry in the Search menu you call a dialog in which you can replace texts in the diagram. The replace dialog is described in section 2.5.3. It has a find next command that works the same as in the find dialog. Furthermore, the replace dialog has a replace next and a replace all button. Replace next means that for the next shape (the shape that is found with find next) all their text strings that match are substituted by the string to replace (that is the second string filled in in the dialog). In the case of Replace all this happens to all shapes in one command (global substitution).

Note that find and replace work on entire shapes at the same time. But keep in mind that a shape could have multiple text labels and each individual text label could match the string to find or the string to replace multiple times (at least when you look for a substring). When you want to find or replace within a single text label or table cell, you should load that text label first in the out-line text editor and then do a find or replace within that dialog. The out-line text editor only works on a single text shape at the time.

Note also that you can fill an empty string for the string to find in both dialogs. An empty string matches only the text shapes that are empty. You can also fill in an empty string in the replace with text field, providing that the string to find field is not empty.

3.13 Undo and Redo

The last issued edit commands can always be undone and, when undone, also redone again. Undo is the first entry of the Edit menu (in the menu bar and also under button-3). Undo has the accelerator <Ctrl+U>. Undo is multiple levels deep. Redo is the second entry of the Edit menu and has the accelerator <Ctrl+R>. You can redo the last undone commands. Redo is also multiple levels. However, when you have undone a command and then issued a new command then the undone command can not be redone anymore. Both undo as redo show in their menu entry the name of the command that can be undone respectively redone. In the current implementation the undo can be several hundreds of levels deep, but when a new diagram is created by Load or New, the undo history is deleted. When undo or redo is not possible, the corresponding menu entry is shaded and can not be issued.

While you are busy performing a command, like creating an edge, moving the paste box or resizing or moving some shapes, you can abort the command by clicking button-1 during edge creation or button-2 for the other commands. When you have aborted, nothing is changed and the undo or redo of that command is not possible nor necessary and the undo menu entry lists the command issued before the aborted command.

Figure 3.6 summarizes all atomic commands that are available in every diagram editor. These commands are either issued by the mouse, via the Edit or Properties menu, or via an accelerator. Each of these commands is undo- and redo-able.

**Figure 3.6:** All atomic diagram edit commands.
$\includegraphics{p/diagramcommands.eps}$

3.14 The Generic Diagram Editor (TGD)

The generic diagram editor is a diagram editor that has all the features that are described in this chapter.

3.14.1 Nodes and Edges

The generic diagram editor has three node types generic node, empty node and comment, and one edge type, generic edge. Empty nodes are nodes that don't have a name label. TGD has no immediately enforced and no soft constraints. A generic or empty node can be represented by a number of node shape types and a generic edge can be represented by a number of line types, see figure 3.8.

Each node shape and each line can be drawn in the line styles solid, dashed or dotted. You can set the default line style of the node shapes via an option menu below the tiled node buttons in the main window. The default line style for lines can be set via the same kind of option menu below the tiled edge buttons. The line style of a node shape or line can be changed via the Update Line Style command in the properties menu. This command sets the line style of all selected shapes to the chosen line style.

All node shapes in TGD, except comments, bulls eyes, black dots and solid horizontal and vertical bars can have an index text label. If you want to see these text labels select the create/edit index toggle button below the node tiled buttons. By default the index labels are numbers between 1 and the total number of nodes having an index. Each new node shape will get the lowest currently unused index number higher than 0. When you deselect the create/edit index toggle, all node indexes remain visible, but new node shapes will not get an index number. When the create/edit index toggle is selected you can edit the node indexes just like the name of a node. You can create a new index label for a node shape by editing the empty index label of the node shape. The node index can be made any string in TGD. If you issue the Renumber Indexes command from the Edit menu of TGD then the current indexes of the diagram are renumbered. Each node shape that has an index will get as index a unique number between 1 and the total number of node shapes with an index.

In TGD it is possible to convert between different node shape types. For this, select the shapes that you want to convert to some other node shape type and choose the desired shape type from the submenu of Convert Node Shape Type in the Edit menu. This will pop up a dialog window containing toggle buttons for the various node shape types you can convert to. See figure 3.7 for the available node shape types. This makes it possible to change a box in an ellipse or vice versa. Note however that only the representation is changed. The node itself is not changed.

**Figure 3.7:** Possible node shape types to convert to.
$\includegraphics[width=3.5in]{p/convertnodeshapetype.eps}$

In TGD, every node, including comment nodes, can be connected by any type of edge and given the facts that you can make comment nodes invisible by emptying its text string or make the line style of a node shape invisible and that you can connect edges with edges too, you can make very complex drawings in TGD. The arrows and other kinds of line endings of the lines in TGD can be changed via the Update Line End command in the Properties menu.

Figure 3.8: Generic diagram nodes and edges.

$\includegraphics{p/box.eps}$	Generic node	$\includegraphics{p/diamond.eps}$	Comment	$\includegraphics{p/ellipse.eps}$	Generic node
$\includegraphics{p/circle.eps}$	Generic node	$\includegraphics{p/square.eps}$	Generic node	$\includegraphics{p/blackdot.eps}$	Empty node
$\includegraphics{p/roundedbox.eps}$	Generic node	$\includegraphics{p/verticalbar.eps}$	Generic node	$\includegraphics{p/horizontalbar.eps}$	Generic node
$\includegraphics{p/triangle.eps}$	Generic node	$\includegraphics{p/hexagon.eps}$	Generic node	$\includegraphics{p/ellipsedbox.eps}$	Generic node
$\includegraphics{p/blackdotellipse.eps}$	Empty node	$\includegraphics{p/solidverticalbar.eps}$	Generic node		Empty node
$\includegraphics{p/stickman.eps}$	Generic node	$\includegraphics{p/folder.eps}$	Generic node	$\includegraphics{p/subfolder.eps}$	Generic node
$\includegraphics{p/cube.eps}$	Generic node	$\includegraphics{p/buildingblock.eps}$	Generic node	$\includegraphics{p/disk.eps}$	Generic node
$\includegraphics{p/bce_boundary.eps}$	Generic node	$\includegraphics{p/bce_controller.eps}$	Generic node	$\includegraphics{p/bce_entity.eps}$	Generic node
$\includegraphics{p/comment.eps}$	Generic node	$\includegraphics{p/notebox.eps}$	Empty node
$\includegraphics{p/line.eps}$	Generic edge	$\includegraphics{p/arrow.eps}$	Generic edge	$\includegraphics{p/doubleheadedarrow.eps}$	Generic edge
$\includegraphics{p/doublearrow.eps}$	Generic edge	$\includegraphics{p/generalizationarrow.eps}$	Generic edge	$\includegraphics{p/openarrow.eps}$	Generic edge

Footnotes

... curve ^3.1: See http://www.moshplant.com/direct-or/bezier for more information about Bezier curves.
... node ^3.2: The question is if it is still the same edge when it connects to a different node...
... undesirable ^3.3: But imagine the situation that labels do not move when nodes and edges are moved, that would be far more frustrating.
... duplicated ^3.4: Duplicate lines are meaningful too in our opinion, so we want to add them to TCM in the future.

Next: 4. Data View Editors Up: Toolkit for Conceptual Modeling Previous: 2. Document Editing

Henk van de Zandschulp
2003-01-20