Molecular_Interaction_Maps/LayoutRules

<<TableOfContents: execution failed [Argument "maxdepth" must be an integer value, not "[2]"] (see also the log)>>

Summary

This is perhaps not so much a particular use case, but an explanation of layout constraints that cut across several use cases.

Step-by-Step User Action

• N/A

Visual Aides

Multiple Edge Segments: An edge may be broken into any number of segments.

Vertical and Horizontal Edges: The primary paths for all edges should align vertically and horizontally. The last edge segment that connect to a node, may be at an angle that deviates from the horizontal or the vertical, as in this example:

Crossing Edges: The edges may cross one another at right angles, but may not overlap nor cross where there is a node.

Minimum Edge Length: To avoid visual confusion with the symbols for ../TranscriptionInhibitionShorthand and ../TranscriptionShorthand, there should be a minimum segment length, that is longer than the hitches in these symbols.

Other Angles: There are several places were acute angles as used a part of the notation. These include ../BooleanLogic, ../PathBranching, ../ShowCompetition, and ../StoichiometricConversion. In each of these cases, the user interpretation of the meaning of the acute angle is dependent on the context in which the angle occurs.

Requirements for Cytoscape

TBD

Importance

The layout rules are important for minimizing clutter and reducing ambiguity.

Other Examples

Comments

AllanKuchinsky - 2007-01-23 05:12:16

This may be challenging in terms of changes needed for the renderer.  It might require some support for rudimentary constraint handling in the EdgeDetails class.  One issue is how general to make such constraint handling.  Should we keep things simple and implement only that constraint handling that is necessary to support MiMs?  Should it be more general to support other, unanticipated usage?  One hybrid measure to consider is to support a "manhattan" style set of constraints, as is done in some CAD layout systems.  This all requires further study of the renderer internals.

AllanKuchinsky - 2007-05-25 10:58:27

One simplifying assumption is that if the acute angles are always at one end of the edge, then they could be implemented via new arrowhead types.