Automatic target enlargement for simplified selection

AUTOMATIC TARGET ENLARGEMENT FOR SIMPLIFIED SELECTION

The present invention relates to a graphics system for providing simplified pointer selection of objects.

It is a well known problem that it is difficult to select small objects with a mouse or other pointer. This problem arises most often, in a cluttered environment, or in the presence of unexpectedly selectable objects. Examples are CAD and VLSI design systems, where selection can be difficult even for an experienced user. A similar problem arises when selecting a moving target. It can also arise for partially disabled users or users not accustomed to mice, when selecting small or thin objects.

An example of a prior art document which attempts to solve this problem is US Patent No. 5,808,601 which discloses an object attractor mechanism where a pointer is attracted to objects based on a calculated mass for the object.

The present invention on the other hand provides a graphics system as claimed in claim 1.

The invention operates by permitting the pointer to select small objects when positioned near but not over an object. In the preferred embodiment, objects are considered to lie within a target area (which will be referred to as an extended target area) larger than the object's visible area. Preferably, the smaller an object, the more the target area extends around the object.

Preferably, as the pointer moves, a weighting*is calculated for each screen object. The object with greatest weight, if it is over a predetermined threshold, is taken as the selectable object. The 'weight' metric is defined to decrease with increasing geometric distance between the pointer and an object, and to increase for small target objects.

It will be seen that, while multiple identification of objects can happen with objects that visibly overlap, or where their extended target areas overlap, using the invention, it is possible, where a small object lies inside or on the edge of a larger one, for a pointer positioned over the large one and near but not over the small one to identify and make selectable the small object.

Embodiments of the invention will now be described with reference to the accompanying drawings, in which:

Figure 1 illustrates the measurements for calculating a weighting for an object according to a first embodiment of the invention,Figure 2 shows the extended target areas of a number of closely spaced objects; and

Figure 3 illustrates the measurements for calculating a weighting for an object according to a second embodiment of the invention.

The embodiments of the present invention will be described in object oriented terms, nonetheless, the invention is not limited to this method of programming.

In Microsoft operating systems such as windows '95 a Graphical Device Interface (GDI) is an interface to a graphic system for representing graphical objects and transmitting them to output devices, such as monitors and printers. Another example of a graphics system is xwindow, developed at the Massachusetts Institute of Technology (MIT) , and upon which almost all UNIX graphical interfaces, including Motif and OpenLook, are based.

Conventional graphics systems communicate with a pointer driver, for example, a mouse driver, and are informed of any movements of the pointer as well as any button pushes. The graphics system is able to examine all screen controls and to detect which screen control is being targeted so that screen controls can be pre-selected as the pointer moves, for example, to provide roll-over emphasis, and the graphics system is also responsive to a button push, so that a screen object can be selected. It is therefore the graphics system which determines the screen control which is to be selected and which is ’to respond to a pointer selection.

When a pointer lies within an application window, for example, a CAD drawing or a Freelance presentation, the operation of the application for selecting application controlled screen objects is analogous to the operation of graphics sub-system within the operating system. While some application graphic systems do not implement pre-select, ignoring mouse movements, until a button is pushed, they still include means for examining screen objects and determining which screen object is being pointed at any given time and as such are easily adapted to incorporate the present invention.

The present invention is based on the assumption that a graphics system either at the operating system level for handling operating system screen controls or within an application for handling screen objects can examine screen objects and identify the screen control which is to be selected and which is to respond to a pointer selection

Table #1 shows a sample function enabling a graphics system according to the invention to apply a weighting to displayed screen objects and to enable the graphics system to then determine the targeted screen object which in turn may be selected or made selectable (where pre-select is provided).

Referring also to Figure 1, the function identifies the weight of a target object 1, for a pointer position X. x and y axes are considered independently to make it easy to select thin objects. maximumGrowth is the maximum 'growth' allowed for an object. A constant KK is used to prevent two fairly possible cases of unselectable objects:

a small box with 0 sized grab handles where the extended area of the grab handles fully encroach on the box; and narrow characters in text with zero sized selectable intercharacter gaps (the gaps encroach on the characters).

The weight is calculated for each dimension and summed. If the sum weight is positive the pointer is within the extended target area of an obj ect.

The growth factor for an individual object is computed to maximize at maximumGrowth, and to fall towards 0 for large objects. KK being less than 1 slows the rate at which the growth factor falls. If the pointer is over the object the growthFactor is taken as the weight, otherwise the weight falls away by distance from the object. Thus at distance growthFactor the weight is 0, and beyond that, distance becomes negative.

It will be seen that using this first embodiment, the invention can be best used to select objects which are rectangular in shape and which lie parallel to either the x or y axis.

Figure 2 illustrates the operation of such an embodiment. As will be seen, some of the boundaries of the extended target areas surrounding objects are quite irregular as the surrounding objects fight for space.

In the example, box 10 is used to control and display the maximum growth. This is set rather high in the present example for illustrative purposes.

Extended target area 21 (marked dotted) indicates the selectable area for line 22 surrounded by various small boxes (23 and others) .

Target area 21 has become divided into three regions as the potential selectable area for object 22 is encroached into by the growth of the small boxes selectable areas.

It can also be seen that because object 23 is relatively small, the extended target area for box 23 (marked with horizontal lines) even spreads to the right hand side of line 22.

A text area 30 shows the division of selectable areas for letters. The vertical spread immediately makes all letters more easily selectable. The exact vertical spread varies according to the neighbouring objects. Nonetheless, it can be seen that, for example, the blank between and 'W' has been considerably widened at the expense of the 'W', but not at the expense of the The '1' has also become wide as both surrounding characters are wide. The remains fairly narrow, it encroaches a little into the 't'; but as 't' is itself thin the is not allowed to encroach too much.

The invention can also assist the selection of inter-character gaps within text (not illustrated) . As above, a text stream is considered as a series of boxes, one per character, each the width of its character.

The inter-character gaps may be considered as selectable boxes of zero width (and text height) placed between the characters. The gaps automatically encroach into the characters without completely obliterating their selectability. Changing the value kk in the weight formula (table 1) changes the relative selectable size of the characters and the gaps.

The target area 40 (marked by diagonal stripes) of large box 41 is generally moved left. This is because to the right of box 41 are a variety of small 'ports' (42 and others) whose target areas (43 and others) encroach into the target area 40 of the large box 41. However, the left hand side of the large box 41 lies in an uncluttered region, so the target area 40 spreads to the left of the box. *As the box 41 is large, this spread is not very pronounced.

Box 50 represents a box that can be scaled by standard drag handles along the edges and at the corners. Following conventional practice, these handles are not visible until pre-selected when the mouse moves into their target area; or selected after a mouse click. The target areas for the edge handles (51 and others) are marked by diagonal stripes. The target areas for the corner handles (52 and others) are marked by vertical stripes.

In a conventional system, the handles have to be given a precise size, with the present invention, the edge handles have zero width, and · the corner handles zero size. Their target areas (51, 52 and others) spread automatically to make the handles easily grabbable. The spread is greater into the screen area outside the box 50 than the spread into the box 50 itself. This is because the weight of box 50 limits the encroachment of target areas of the handles into its target area 53.

This automatically ensures that box 50 itself remains easily selectable. A smaller box would limit the internal growth of its grab handles even more.

In an alternative embodiment, Figure 3, where is it desired to select objects not aligned to the x and y axes, dx and dy are replaced with measurements dl and d2.

In this embodiment, it is assumed that objects are rectangular or are bounded by a rectangle having vertices vl..v4, running anticlockwise. If a pointer lies within an object, ie. if the pointer position lies to the left of each of the vertices, then dl and d2 are 0 and the object has its maximum weighting, eg. point x. This weighting is of course limited by the size of the object, as smaller objects have larger growth factors.

If the pointer lies to the right of vertex vl or v3, eg point x', then dl is calculated as the distance from the pointer to the respective vertex. Similarly, if the pointer lies to the right of vertex v2 or v4, then d2 is calculated as the distance from the pointer to the respective vertex. In the case of a pointer position such as x, the object will have both a positive dl and d2 component, thus diminishing its weight twofold with respect to that pointer position.

The growth factor in the direction of dl is taken as the distance between vl and v3 and the growthfactor in the direction of d2 is taken as the distance between v2 and v4. Thus, a weight can be calculated in each direction by subtracting dl and d2 from their respective growth factors, and an overall weight can be calculated by adding tHese weights as before.

It will be seen that in spite of the present invention, there could be a problem where a large object is nearly covered in small objects.

All the small objects may grow to overlap, and leave none of the large object selectable. Some form of special override may be needed for this: for example selection with 'shift' key.

It will be seen that, for best use, the system according to the invention should have pre-select feedback. (The means for providing such feedback are providing anyway by continuously determining the targeted object.) Feedback can be provided, for example, by highlighting a · selectable object, or merely indicating status by changing the pointer icon.

It will be seen from the foregoing description that the invention can be implemented within any operating system or general purpose application, for example, within publishing, drawing, presentation, spreadsheet or word processing packages or within dedicated applications, that is anywhere where there is a requirement on the application or operating system to select an object from within a screen display it controls.

Table #1

ΚΚ = 0.33 • find geometric distance in x direction.

If x < l.Left Then dx = l.Left - x

Elself x > l.Left + 1.Width Then dx = x - l.Left - 1.width

Else dx = 0 End If ' find growth factor based on object size growthFactorX = (maximumGrowth * maximumGrowth) I (maximumGrowth + KK*1. width) ' thus find weight (in this axis) wx = growthFactorX - dx ' repeat .... for y ' find overall weight

1.Weight = wx + wy