WO2006061134A1

WO2006061134A1 - Device for positioning an object on an interface

Info

Publication number: WO2006061134A1
Application number: PCT/EP2005/012832
Authority: WO
Inventors: Laurent Baleydier
Original assignee: Kartoo Sa
Priority date: 2004-12-08
Filing date: 2005-12-01
Publication date: 2006-06-15

Abstract

The invention relates to a method for classifying and managing object-results obtained according to a request making it possible to generate a set of EC couples comprising proximity relations between the object of said set. The set is used for arranging at least a part of objects in such away that a convex-shaped array is formed in which the members of the same couple are placed substantially side-by-side along the array. Said proximity relation is used, mainly for displaying objects in an optimal manner minimising intersections or crossings of different links between objects and/or between attributes and objects. The inventive method makes it possible to reduce hardware resources such as storage and computing capacities. In addition, instructions to carry out are simple and easily performable with the aid of Flash interfaces.

Description

Method of positioning objects on an interface

The present invention relates to a method of positioning on an interface of object-results obtained following a request. This method makes it possible to generate a set of pairs comprising a proximity relation between the objects of said set. This assembly is used to dispose at least a portion of the objects so as to form a convex shape arrangement in which the members of the same pair are arranged substantially side by side along the arrangement.

Most search engines known to date can display the results in the form of lists, in which the objects are simply listed one after the other. This type of presentation is well suited for cases where objects are few. By cons, as soon as their number increases, this display mode quickly reaches its limits. Indeed, as soon as a list is extended, for example beyond the displayable content on a single screen, many users only consult the first lines presented. If the results have multiple pages, as is often the case when searching on a network such as the Internet, most of the results can be completely ignored by the user.

To overcome this type of situation, the applicant has developed and popularized a display mode of objects not in the form of lists, but in the form of spatial diagrams, in which the objects, and any associated themes or attributes, are distributed. in the space. To make such an interface even more ergonomic, the applicant has also planned to display the existing links, either between objects and associated themes, or between objects themselves from the same family.

The graphical possibilities of the interface are thus judiciously used, and make it possible to provide additional information about the displayed objects and themes. This type of display is often referred to as the "map interface". Given the significant increase in the number of searches for all kinds of information, whether on the content available on a network or on the user's station - which has more and more storage capacities- , this type of interface is particularly useful and more and more used.

However, following a search, we obtain data limited to the only result, that is, objects that respond favorably or that meet the criteria of a query. Thus, the data, or objects, that can be positioned in the interface do not include indications relating to their spatial positioning in such an interface type. Various principles have been developed to position the objects so as to best occupy the available surface of the interface. However, the principles developed are often heavy, and require significant hardware resources such as large memory capacity, a lot of processor time, and so on. Consequently, such applications are generally located on a server or remote station, rather than on the user's computer.

For example, some approaches used for placing objects on an interface use a recursive method:

- begin with an initial position of the objects, often established on an arbitrary basis;

interactions are defined between all the objects of the interface (for example attraction and repulsion forces which are a function of the distances between the objects);

the system is gradually modified by slightly modifying the coordinates of the objects (taking for example a small Delta t time interval) so as to take account of the interactions; these two last operations are repeated until the system obtains some stability. Thus, we must compare a plurality of times each object to all the others in order to study the relative positions of the system;

- The last operation is to check for any overlaps and to move some objects to avoid overlapping.

This method is used, for example, in 3-body systems (eg planetary attraction) that do not have a mathematical solution to calculate the coordinates of objects at a given moment.

This method, however, has some disadvantages:

- it takes a considerable number of operations to achieve a stable disposition (with weak or substantially zero interactions);

each iteration requires a large number of calculations since it is necessary to calculate for each object the interactions with all the others;

- there is no guarantee of finding a stable solution. For example, the system may oscillate, because there may be several minimum interaction points;

- the final solution depends on the initial positions of the objects. Thus, if the latter is random, one can, with the same elements, find endlessly different final provisions;

- a stable solution at the level of interactions does not necessarily imply an optimal arrangement of objects or a good readability of the map.

Thus most of the technical solutions made available to date involve a centralization of heavy technical functions such as carrying out research and / or the spatial organization of objects so as to be able to arrange them in the space. The heavy functions are thus carried out remotely, using appropriate tools, with the necessary resources. On the other hand, this centralization requires the provision of network connections allowing the exchange of data between the centralized station and the user stations. Such connections are not always available. In addition, for certain configurations or tasks, one may wish to perform all the tasks locally (on the workstation).

Moreover, with known type interfaces, little positioning optimization is achieved. Thus, it is common for the relative positioning of the objects to create zones of intersection or superposition of the displayed links. It also happens that a card has inconsistencies. For example, objects that should be positioned substantially close to each other remain distant because no position of proximity equilibrium is reached.

The interface loses its clarity. For these different reasons, the spatial representation of a series of objects on a working interface always poses many difficulties, these difficulties being increased according to the number of objects.

These numerous difficulties also explain the still widespread use - despite the drawbacks mentioned in the foregoing - of conventional displays, in the form of lists of objects, for example comprising a list containing the objects resulting from a request, presented one by one. after the others, according to an order established according to criteria that often escape the user himself.

To overcome these various drawbacks, the present invention proposes a method of positioning, on a graphical interface, result-objects associated with attributes obtained following a search carried out by a search engine, for a digital information system comprising a microprocessor, memory elements, a display device, said method comprising the steps of:

(a) for each of the attributes, the number of connected objects is established and the attributes are weighted to favor the attributes that connect the least number of objects;

(b) forming pairs of objects according to the attributes and for each pair of objects sharing at least one attribute, a score is established based on the set of weights of the attributes associated with that pair;

(c) generating a set of pairs EC constituted by at least a part of the pairs for which a score could be established;

(d) a proximity relation is generated between the different objects of said set so that the objects whose score is the most favorable (the most linked to each other) are given a strong proximity relation and the least connected objects to each other are assigned a weaker proximity relationship;

(e) producing and displaying an interface for displaying said result-objects on the display device, and disposing on said interface at least a part of the objects of the set so as to form a convex-shaped arrangement in which members of the same pair are arranged substantially side by side along this arrangement,

Thanks to this method, the number of operations is greatly reduced compared to a recurrence approach (it can be 100 to 1000 times smaller, and even more in some cases). Indeed, the most expensive calculation is that of the scores which is done only once and involves few elements;

Moreover, contrary to the known methods, one does not seek stability, but readability since:

-the most related objects are the closest ones (thus minimizing the number of lines or graphical links likely to cross all or part of the interface); the convex shape minimizes the crossings (by definition, the lines of the links can not pass over the objects);

the convex form also makes it possible to organize / arrange the objects in the space and thus to be able to study them successively without considering again an object already taken into account and without forgetting it.

The process is reliable, with great stability.

The method provides a tool that can be displayed in a web browser.

On the other hand, the number of operations to be performed is known at least approximately: therefore, the calculation time is known or at least can be estimated relatively reliably.

The method according to the invention also makes it possible to minimize hardware resources such as memory capacities and processor capabilities.

The generated proximity relation is used in particular when displaying the objects (results), in order to present the results optimally. The proximity can be of different nature depending on the type of score.

Such a spatial arrangement in the form of a convex shape arrangement makes it possible to eliminate or at least minimize the crossings between the links, whether the links between the objects themselves and / or the links between objects and objects. attributes, and thus optimize and clarify the interface. By convex shape arrangement is meant a provision in which any connection between two objects on this portion represented by segment or portion of the right, is contained within said provision. In addition, said segment or right-hand portion forms no other intersection with said disposition than that formed at the two linked objects. In other words, everything point between two points of the convex form is necessarily located within this form. The arrangement may be configured to provide a two or three dimensional representation.

Advantageously, the proximity data is generated by establishing a ranking / arrangement of the pairs in order according to the score (ascending / descending). For example, one can get a disposition based on the number of links (increasing / decreasing) between the objects of a couple.

Advantageously, the pairs are listed or arranged to form a list, (L1) according to the ascending or descending order of the score.

Advantageously, the score corresponds to the sum of the weights of the attributes associated with the pair.

Advantageously, the weights of the various attributes are established in a progressive manner, the attribute connecting the more objects having the most favorable weight, and the attribute connecting the least objects with the least favorable weight.

Advantageously, a second list (L2) is formed, considering, in order (increasing or decreasing), each of the pairs of L1 successively, until a stop condition is satisfied, for establish a second list L2 on the basis of the following criteria: ~ (i) if no object of the pair belongs to L2, add the pair to the list of L2, (without joining);

- (ii) if only one object of the pair belongs to L2, and this object is not an end of a sequence, there is no change for the object already positioned in the list L2; the object not belonging to L2 is put on hold; - (iii) if only one object of the pair belongs to L2 and this object is an end of a single sequence, the other object of the pair is added at the end of L2, next to the first object; - (iv) if the two objects of the pair belong to L2, and are ends of two different sequences of L2, we form a single sequence from the two sequences in which the objects appear.

Advantageously, one starts with the pair of L1 whose score is the highest.

Advantageously, the convex shape arrangement is generated from at least a part of the objects of the list L2, arranged in the interface with positions corresponding to the order established in the list L2. Thus, neighboring objects in list L2 remain adjacent once arranged along the convex shape arrangement.

Advantageously, the arrangement is made or created or produced at the interface with said objects, arranged in the order given in the list L2, and arranged in the space so that the two objects placed in end approach to form a loop, or a quasi-loop with L2 objects.

Advantageously, for any case other than one of the cases (i) to (iv) above, the list L2 remains unchanged.

According to a first embodiment, a stop condition is satisfied if the list L1 is empty.

According to another embodiment, a stop condition is satisfied if the list L2 contains all the elements.

According to yet another embodiment, a stop condition is satisfied if the list L2 contains all non-isolated objects. According to an advantageous embodiment, the result-objects are preferably derived from a search by a search engine. It can be a search of websites done on the Internet, a search for data of any type in an internal or external database, a search for favorites, etc.

According to an advantageous embodiment, the score is established by counting the number of links (NL).

The score can be established by comparing data, attributes, values, elements, whose nature depends on the type of activity or the field of application. For example, it may depend on the number of links between two objects, the semantic distance, the physical or geometric distance, the time difference, the difference between the number of letters, and so on. The score can also be the result of a mathematical operation whose data come from the objects to be compared. According to another embodiment, the user sets himself one or more scores.

According to a preferred embodiment, the score established is proportional to the number of links counted. A link between two objects can come from several sources, such as the number of common attributes between two objects. These can be common object qualifiers, common keywords, common letters, numbers or combinations of common letters and numbers, and so on.

Advantageously, attributes or themes relating to more than one object are disposed within said spatial arrangement containing the object related to that theme. Such a provision is particularly advantageous since it makes it possible to visually represent either the link between the members of a couple, or the link between the members of a couple and an attribute, by minimizing the intersections between the different segments, and by thereby improving the readability and clarity of the interface. In such a case, the attributes or themes are preferably arranged substantially at the center of gravity (with possibly slight shift towards the center).

Advantageously, a computing and generation module is loaded on a workstation, so as to locally prepare said convex shape arrangement. Said loading is preferably performed in a web browser.

The invention also provides computer software for implementing the method described above. The invention finally provides a software product comprising a computer readable recording medium in which is stored a computer software element for the implementation of the method described above.

The invention will now be described and presented in relation with FIGS. 1 to 6, presenting exemplary embodiments used in a nonlimiting manner for purposes of illustrating the invention, and in which:

FIG. 1 presents a functional diagram illustrating the various steps making it possible to go from the list L1 to the list L2; FIG. 2 presents a second functional diagram illustrating the various steps for processing the elements or objects not yet arranged in the interface;

FIG. 3 presents a first example of a convex shape arrangement according to the invention;

FIG. 4 shows a second example of a convex shape arrangement according to the invention, with the same objects as the example of FIG. 3;

FIG. 5 illustrates an exemplary spatial arrangement of objects made in a conventional manner;

FIG. 6 illustrates an example of spatial arrangement of the same objects as in FIG. 5, but realized according to the method according to the invention. In this document, a "sequence" is understood to mean a series of objects, for example (E1, E2) for a pair; (E1, E2, E3) for a trio; (E1, E2, ..., En for a sequence of n elements.

By "convex-shaped arrangement" is meant a provision in which any connection between two of the objects on that portion represented by a segment or a right-hand portion is contained within said provision. In addition, said segment or right-hand portion forms no other intersection with said disposition than that formed at the two linked objects. In other words, every point situated between two points of the convex form is necessarily situated inside this form. The arrangement may be configured to provide a two or three dimensional representation.

"Score" means the sum of the weights of an element, object, attribute or other.

In order to explain the principle implemented in the context of the method according to the invention, a simple alphanumeric example is used in what follows. We therefore consider a set of points P1, P2, ... P8, interconnected by letters A, B, C, ... G. In this example, the figures advantageously represent the objects that we want to have in the space, and the letters the attributes associated with these objects. However, the opposite can also be achieved without affecting the method according to the invention. Relationships are defined as follows:

A (1, 2) (which means that the points P1 and P2 are connected by the letter A, or share the same attribute, designated A).

B (1, 5.7)

C (6.7)

D (1, 4,6,7) E (2,4)

F (3.6.8)

G (3.8) Relationships can be translated into the following matrix;

Attributes are weighted by ensuring that the more links an attribute has, or the more an attribute is common to a larger number of objects, the lower its weight. In other words, the trend towards exclusive links increases weight. In the present example, the following weights are used:

- a letter that links 2 points to a weight of 3;

- a letter that links 3 points to a weight of 2;

- a letter that links 4 points or more to a weight of 1;

We then obtain:

Weight (A) = 3

Weight (B) = 2

Weight (C) = 3

Weight (D) = 1

Weight (E) = 3

Weight (F) = 2

Weight (G) = 3

We replace the letters or attributes in the previous table by their score. The score of an item, letter, attribute or other is the sum of the weights. Thus, the weight is associated with an element, an attribute, object, etc., while the score is associated with a couple.

We then consider the list L1 pairs of points or objects that are sorted by decreasing scores. An L1 list is obtained as follows: L1 "(3.8) (6.7) (1, 2) (1, 7) (2.4) (1, 5) (3.6) (5.7) (6.8) (1, 4) (1, 6) (4,6) (4,7)

This list contains objects that would be desirable to position side by side. However, the list is not directly usable, since it includes repetitions. We must now determine the order in which the objects can be arranged side by side, without repetition, so as to find, as far as possible, a single list with all the objects. To do this, we build a new list L2 from L1, considering the pairs in the order of the list L1, according to their score. The objects of the pairs of the list L1 should in principle be arranged side by side. It is now necessary to establish a kind of priority in order to integrate the objects that are repeated, always favoring the couples of higher scores.

The functional diagram of FIG. 1 explains the different steps of this part of the method, making it possible to obtain the list L2 from the list L1 so as to obtain a minimum number of sequences, and preferably a single sequence. This diagram includes five main cases or tests identified in Figure 1 using references i to v. In step 0, in the first line, we select a pair, the one with the highest score. If L1 was generated as in the previous example, the pairs are successively selected in L1 from left to right, corresponding to a selection of the highest score at the lowest score. At each of the steps i to v, the objects of the selected pair are tested or compared with respect to the objects of the list L2. Then, as the case may be, the action provided is applied to one or the other of steps i to v.

If none of the objects in the pair belong to list L2, case i (addition of torque, without junction) applies. In this case, the couple considered is added to the list L2, leaving it as a couple. At this stage, no junction or integration is made between the objects of this pair and those of one or other of the L2 suites. This step i makes it possible to reduce the list L1 of two objects, since the two objects of the couple considered are integrated in L2.

If only one of the two objects of the pair is already present in the list L2, and this object is not an end of a sequence, the case ii (putting on hold of an object) applies. Since an object is already in the list L2, then there is no change for this object nor for the corresponding list. The other object is put on hold, for example by placing it in a file "objects put on hold" for further processing, as described below. This step ii makes it possible to reduce the list L1 of an object, since one of the two objects of the pair considered is already integrated in L2.

If only one of the two objects of the pair is already present in the list L2, and this object is an end of a single sequence, the case iii (complement by the end) applies. Since an object is already in the list L2, then there is no change for this object nor for the corresponding list. The other object is added to one end of the L2 suite including this object, preferably next to it. This step iii makes it possible to reduce the list L1 of an object, since one of the two objects of the pair in question is already integrated in L2.

If the two objects of the pair are already present in the list L2, and each of the objects is an end of a distinct sequence, the case iv (junction) applies.

Since both objects are already in the L2 list, the overall contents of the L2 list do not change. On the other hand, the format of the list, or more precisely the layout respects of the suites is modified as follows: the two suites containing the two objects are joined to form one. The junction is preferably performed by keeping the objects of the two suites concerned at the same respective locations and placing the objects of the two suites one after the other in a single sequence. The two original suites then form a single suite.

In all other cases, normally concerning the object repetitions, the list L2 is not modified, and the couples are removed from the list L1.

For example, if the two objects of the pair are on the one hand already present in the list L2, and that these objects are each at an end of one and the same sequence; if the two objects of the pair are on the one hand already present in the list L2, and that these objects are at any position of one or more sequences, excluding the ends.

The example previously presented, used to produce the list L1 is now used to illustrate, by way of example, the implementation of the various previous steps.

We take the 1st couple: L2 = (3,8)

We consider the 2nd couple (6.7). This pair fulfills condition (i) of Figure 1. Since it has nothing in common with the 1st, it is placed side by side: L2 = (3,8) (6,7)

We consider the third pair (1, 2). This couple also fulfills condition (i) of Figure 1. It has nothing in common with the first two. It is also placed as follows: L2 = (3,8) (6,7) (1,2) The fourth element (1, 7) has a common point with each of the last two pairs and allows to connect these pairs of L2. This pair fulfills condition (iv) of Figure 1. Thus, each of its objects is in the list L2; since this pair makes it possible to connect two pairs, this link is materialized or made effective by joining said two pairs, so as to form a single sequence. In this example, we obtain: L2 = (3,8) (6,7,1,2)

The next pair (2,4) fulfills the condition (iii) of figure 1. The object not present, the 4, is added at the end next to the other object of the pair, the 2. We obtain: L2 = (3.8) (6.7, 2.4)

The pair (1, 5) fulfills the condition (ii) of Figure 1 II has a common point with (6,7,1,2,4) but can not integrate without breaking a stronger bond, the couples (7.1) or (1, 2) each having a score greater than the torque (1, 5) ι It is also noted that the common object is not positioned at one end of a sequence. Object 1 is already in the list. Object 5 is put on hold. Thus, L2 does not change: L2 = (3,8) (6,7,1,2,4)

The next pair (3,6) fulfills the condition (iv) of Figure 1. It has a common point with the last two pairs, and allows to bring together the two suites. We thus obtain: L2 = (8,3,6,7,1, 2,4)

The pair (5, 7) fulfills condition (ii) of FIG. 1. Object 5 is already on hold. So, no change.

All other pairs ((6,8), (1, 4), (1, 6), (4,6) (4,7)) satisfy condition (v) of Figure 1. In addition, all objects of these couples are already in the L2 list. These couples do not make any changes to the L2 list, There remains only one object, the 5, to be placed in the list, since all the others are already there, this object having also been set aside for further processing. This subsequent treatment is discussed below.

Once the previous tests have been carried out, the list L2 contains, in the simplest case, a single sequence with all the elements. However, there are more complex cases. The flowchart of FIG. 2 illustrates various possible cases and the corresponding actions, if one has, once the preceding tests have been applied, one or more distinct series, one or more elements put on hold (for example a pair of which one element belongs to L2 but is not an end of a series), or one or more isolated elements (unrelated to others). The following tests are performed. Consider the element (s) not positioned in the list L2:

For separate series, we can position them at random if they have no connection. But if they have one or more links (for example, an element put on hold because one of the elements belonged to a sequence without being an endpoint), we can consider the orientation for which the distance between the rejected pair of scores the most strong is the shortest.

As an illustration, let us take for example the following series:

Series l: (E1, E2, E3, E4, E5)

Series 2: (E6, E7, E8, E9) Couple rejected using the first tests: (E8, E2)

We have two possible provisions:

(E1, E2, E3, E4, E5) (E6, E7, E8, E9) which is equivalent to (E6, E7, E8, E9) (E1, E2,

E3, E4, E5) or (E1, E2, E3, E4, E5) (E9, E8, E7, E6) which equals (E9, E8, E7, E6) (E1, E2,

E3, E4, E5) The layout that offers the shortest distance between E8 and E2 is the first:

(E6, E7, E8, E9) (E1, E2, E3, E4, E5) The two series are thus grouped as follows; (E6, E7, E8, E9, EI ₁ Ea ₁ ES ₁ E ^ E5)

For items with links, but previously put on hold, they can be inserted between the two elements of the final series that have the least links. According to an advantageous variant, the highest score pair is considered. Insert the element of the couple that was rejected to the left or right of the element that belongs to a series of L2. To know which way to insert it, one can consider (if it exists) another pair of score immediately inferior and insert the rejected element on the side where the second element of this pair will be the closest. By way of illustration, let us take for example a series (E1, E2, E3, E4, E5, E6), the rejected element E7 belongs to: (E3, E7) of score 2; (E5, E7) of score 1; E7 will fit to the left or right of E3. Since E5 is on the right, we choose this side. We obtain: (E1, E2, E3, E7, E4, E5, E6).

For isolated (unrelated) elements, either insert them between the objects of the final series that have the least links (lowest score pair), or position them out of the convex shape arrangement, in order to illustrate the independence between these different elements.

The remaining point P5 can either be positioned outside L2, thus remaining isolated, or following the others, and we obtain: L2 = (8,3,6,7,1, 2,4,5)

That is advantageously between P7 and P1 since it is one of the pairs (1, 5) and (5.7). We thus obtain a list containing all the objects: L2 = (8,3,6,7,5,1, 2,4)

The figure obtained by positioning these objects substantially side by side, in the same order as in list L2, once arranged in the form of a circle, is illustrated in FIG.

These same objects can also be arranged in another convex form, as shown for example in FIG. 4.

It should be noted that the attributes or themes are determined conventionally for example using words, groups of words, or keywords common to a couple.

To avoid having too many themes or irrelevant themes, filters can be used, such as filters consisting of deleting words of less than three letters.

Claims

A method of positioning, on a graphical interface, result objects associated with attributes obtained following a search carried out by a search engine, for a digital information system comprising a microprocessor, memory elements, a device for display, said method comprising the steps of:

(d) generating a proximity relationship between the different objects of said set so that the objects with the most favorable score are given a strong proximity relationship and the least connected objects are assigned a proximity relationship weaker ; (e) producing and displaying an interface for displaying said result-objects on the display device, and disposing on said interface at least a part of the objects of the set so as to form a convex-shaped arrangement in which members of the same pair are arranged substantially side by side along this arrangement.

The method of claim 1, wherein the proximity data is generated by establishing a ranking of the pairs in order according to the score.

3. Method according to one of the preceding claims, wherein the pairs are listed or arranged to form a list, (L1) according to the ascending or descending order of the score.

4. Method according to one of the preceding claims, wherein the score corresponds to the sum of the weights of the attributes associated with the pair.

5. Method according to one of the preceding claims, wherein the weights of the various attributes are established progressively, the attribute connecting the most objects with the most favorable weight, and the attribute connecting the least of objects with the least favorable weight.

6. Method according to one of the preceding claims, wherein forming a second list (L2), considering, in order, each of the pairs of L1, in succession, until a condition of stop being satisfied, to establish a second L2 list on the basis of the following criteria:

- (i) if no object of the pair belongs to L2, add the pair to the list of L2;

- (ii) if only one object of the pair belongs to L2, and this object is not an end of a sequence, there is no change for the object already positioned in the list L2; the object not belonging to L2 is put on hold;

- (iii) if only one object of the pair belongs to L2 and this object is an end of a single sequence, the other object of the pair is added at the end of L2, next to the first object; ~ (iv) if the two objects of the pair belong to L2, and are ends of two different sequences of L2, we form a single sequence from the two sequences in which the objects appear.

7. Method according to claim 6, in which one starts with the pair of L1 whose score is the highest,

8. Method according to one of claims 6 or 7, wherein the convex-shaped arrangement is generated from at least a portion of the objects of the list L2, arranged in the interface with positions corresponding to the order established in list L2.

9. Method according to one of claims 6 to 8, wherein the arrangement is performed at the interface with said objects, arranged according to the given order in the list L2, and arranged in space so that the two objects placed at the end approach each other so as to form a loop, or a quasi-loop with the objects of L2.

10. Method according to one of claims 6 to 9, wherein for any case other than one of said cases (i) to (iv), the list L2 remains unchanged.

11. Method according to one of claims 6 to 10, wherein a stopping condition is satisfied if the list L1 is empty.

12. Method according to one of claims 6 to 10, wherein a stopping condition is satisfied if the list L2 contains all the elements.

13. Method according to one of claims 6 to 10, wherein a stopping condition is satisfied if the list L2 contains all non-isolated objects.

14. Method according to one of the preceding claims, wherein the attributes or themes relating to more than one object within said spatial arrangement containing the object related to this theme are available.

15. Method according to one of the preceding claims, wherein a computing and generating module is loaded on a workstation, so as to locally prepare said convex shape arrangement.

The method of claim 15, wherein the loading is performed in a web browser.

17. Computer software for implementing the method according to one of claims 1 to 16.

A software product comprising a computer readable recording medium in which a computer software element is stored according to claim 17.