FR2746524A1 - Man-machine interface for telephone - Google Patents

Abstract

The man-machine interface combines digits on a conventional telephone key pad to allow alphabetic characters to be formed with various key combinations. In addition various control functions can be generated by other key combinations. A single key is used to handle direct entry of a digit, double presses or press with extended hold down. These different modes can be identified and associated with control functions. Two keys are pressed to generate a character, according to a defined pattern of groups. The face of the key pad is marked to show which letters are associated with different key combinations. Upper and lower case letters are determined by the order in which the two keys are pressed.

Description

 This document describes a human-machine interface for entering alphanumeric characters from a 12-key telephone keypad (the ten Arabic digits plus * and #).

 Such an interface is of course less practical than a real alphanumeric keyboard (like that of a computer or a typewriter) but it finds its justification in the context of short messages. It is a telecommunications service allowing a generally mobile subscriber (paging or additional mobile telephone service) to send a message of a few sentences. On the one hand, due to the brevity of the message, the lack of comfort of the interface is more tolerable. On the other hand, the interface only uses the keys that the terminal already has, and therefore authorizes a smaller terminal (which is critical if it is a mobile), and less expensive.

The interface can have two applications (A1) and (A2). The first, (A1) is to allow a subscriber to a mobile telephone service to compose a short message on his mobile telephone before transmitting it. The second application (A2) would be to provide an interface from any phone via voice keys to a server acting as a post office for short messages. It would then be necessary that the interface described here
becomes a standard, because to be usable it supposes that indications are printed on the terminal. Note that what characterizes the first application (A1) with respect to (A2) is not that the terminal of the sender is mobile. It could be any terminal of a telephone network which is digital up to the subscriber (ISDN for example), provided that the short message service is offered there. What characterizes it is that the short message is in digital form from the terminal, while in the second application it is only in this form from the server center.

It is this distinction that allows more flexibility in the definition of
The interface.

First we will examine how the already existing interfaces can be improved. The solution generally adopted by these interfaces is illustrated in FIG. 1. There are three letters per key, and the selection is generally made by pressing the key more or less long. This
interface has several faults: - The first of them (D1) is that the time required for each letter is not the same depending on the letter, and this in an entirely arbitrary manner.

- The second (D2), and undoubtedly the worst is that unless you know the position of the letters on the keyboard by heart, you have to look back and forth between the keyboard and the terminal screen for each letter. This is not very natural: that one observes typing on the keyboard of a computer, and one will note that the glance does this back and forth not with each letter but by four or five letters . This back and forth is tiring because it forces the user to find the position on the screen (application Al) of the last letters typed much more often.

- The third defect (D3), is that this interface does not generally respect the position of the letters on an ordinary alphanumeric keyboard, this is due to the fact that one cannot have more than 9 letters per line, unless doing bring to the same key 4 letters, which makes the interface even slower to use.

- Finally the last defect (D4) is that one cannot distinguish capital letters and small letters, if indeed such a distinction is admitted by short messages.

The interface that we propose has the difference that the generation of a letter is not based on the action of a single key but on the action of two adjacent keys. We limit ourselves to adjacent pairs of keys because this facilitates the printing of indications of possible combinations and their effect. The notion of adjacent that we use deserves to be explained. By adjacent in the strict sense is meant what is commonly called adjacent, that is to say distinct and whose columns and rows are the same or adjacent, for example in FIG. 1 the 5 is adjacent to the 1, 2, 3 , 4, 6, 7, 8, and 9. By adjacent to the border we mean that the keys would only be adjacent in the strict sense if the keyboard was toric, that is to say if the highest line came below from the bottom line, and if the leftmost column, came to the right of the rightmost column. For example, in Figure 1 the 1 is adjacent to the border of the 3, *, and #. Finally by adjacent, or even adjacent in the broad sense if need be to insist, we mean that the keys are adjacent to the strict sense or at the edge
We will first describe the different ways of representing the effect of adjacent key combinations on the keyboard. By effect is meant either the generation of a character or the execution of a function. The second possibility is indeed interesting, in particular in the case of the application (A2) where functions such as the cancellation of the last character entered, or the sending of the message are necessary. Also note that the representations described below do not necessarily apply to the whole keyboard, but a pair of adjacent keys, that is to say that each time there are at most only two effects to represent, each corresponding to one of the two possible key press orders. The same keyboard allows several representations to coexist. The representations are classified into three more or less equivalent sets (R1), (R2) and (R3). Figures 2 and 3 give examples of (R1) and (R2) respectively. Only representations from one of these three sets should appear on the same keyboard.

- The first representation (RI .1) applies to the case where the order of pressing the keys is not important, that is to say that the two orders produce the same effect. A line connects the two adjacent keys to show that their combined action is necessary, and the effect is printed on this line. In cases where the keys are adjacent to the border, the line does not really connect the keys to each other, but rather one of them with the position of the counterpart of
The other in a fictitious keyboard identical to the real keyboard and juxtaposed to it in one of the eight lateral or diagonal positions. In this case the line and the effect can be printed once or twice, it's all a question of location.

- The representation (RI .2) applies to the case where the order matters, and there are two effects each corresponding to an order. The representation is the same as (RI .1) except that the two effects are printed on the line, each on the side of the first key pressed in the combination that generates it. Subsequently, this key is called the pre-key of the effect, the other being called post-touch. In the case of adjacent keys on the border, the line can be made to appear twice with only one of the two effects on each, the one whose pre-touch is the key from which the line starts. We can also do as in the case of strict adjacency, except that the effect which is on the side where there is no key has as post-key the key from which the stroke starts.

- (RI .3) applies to the case where the order matters, but only one of the two orders has a defined effect, the effect of the other being reserved for future use (relevant in the case of (A2) ). In this case the representation is as in (RI .1) except that the line is an arrow pointing to the post-key of the defined effect. It would seem more judicious to use (RI .2) of which one of the two effects would be a symbol whose meaning would be defined later, but the advantage of (RI .3) is to be more readable.

- (R1.4.1) applies to the case of alphabetic letters, when we want to be able to generate both upper and lower case. In this case it is advisable that the same pair of adjacent keys is used for a given letter, and that the distinction between upper and lower case is made by order.

Only the capital letter is printed on the line, which allows not to overload. The line is an arrow pointing to the capital post-key.

Formally it is the same as (RI .3), it is up to the user to know
The effect of the reverse order.

- (R1.4.2) is the same spring as (R1.4. 1), the difference is that there is no arrow. Its use supposes that the set of keys on the keyboard is ordered. When the pre-key is the smallest of the keys for this order, then it is the capital letter which comes out, when it is the largest, it is the small letter. Examples of possible order are given below.

Formally (RI .4.2) resembles (Rî .1), it is up to the user to know.

 - (R2.1.1) applies in the same cases as (R1.1), that is to say when the two possible combinations have the same effect. As for (R1) the keys are connected by lines to materialize the relations of adjacency. It is also possible to play on the shape of the keys as illustrated in FIG. 3.

The case of adjacent keys on the border is treated in the same way as in (R1). The difference is that the effect is printed on one of the two keys, on the side where the line starts, rather than on the line itself. The one of the two buttons that is chosen to indicate the effect is the smaller of the two when
I 'the set of keys of the keyboard is ordered, we will see examples of orders further - (R2.1.2) idem as (R2.1.1), but the effect is on the two keys. (R2.1.1) is more readable than (R2.1.2), but we can afford the latter for the adjacent keys on the border.

- (R2.2) is identical to (RI .2) except that each of the two effects is printed on its pre-key on the side where the line starts.

- (R2.3) is identical to (R1.3) except that the line is not an arrow, and that the effect is printed on its pre-key.

- (R2.4) is identical to (RI .4.1) except that the line is not an arrow, and that the capital letter is printed on its pre-key, while the small letter is printed nowhere - (R3.1.1) , (R3.1.2), (R3.2), (R3.3) and (R3.4) are similar to the different representations of type (R2) above defined, except that the keys are not linked by lines . The position of the effect in the key is sufficient to indicate
The other touch of the combination. The relevance of these representations can be assessed from Figure 3, imagining that the lines are absent.

 In the definition of certain representations of the application of key combinations on their effects, in particular (R1 .4.2) to and (R2. 1.1), mention is made of a classification of the keys on the keyboard. Such a classification is especially important when the interface must allow to distinguish upper and lower case. Even if the correct typing order is materialized by an arrow, as in (R. 1.4.1), it is preferable that there is an easy-to-remember classification in order to minimize the amount of information to be read on the keyboard. Here are two examples of classifications meeting this requirement: - (01) is the order given by the meaning of the writing in French. A key a is smaller than a key b if the line of a is above that of b, or if a and b are on the same line, and that a is to the left of b.

- (02) is a partial order, unlike (01) which allows any pair of keys to be compared with each other, (02) is only defined for adjacent keys, which is sufficient here. The idea is the same as (01), but the keyboard is considered as a torus, namely the columns (resp. Rows) follow from left to right (resp. Top to bottom) and the first follows the first. a and b being adjacent, a is smaller than b if it is on the line preceding that of b, or if it is on the same line and its column precedes that of b. Note that this is the same order as (01) except for the adjacent border keys where it is the reverse order.

We are now going to see different possible modes for actuating a pair of keys, all or part of which can be implemented. In their descriptions which follow, the term characters means 0 to 9, plus * and #, that is to say in general the character which is associated only with a single key, and not with the combination of two adjacent keys, which will be called letter: - The mode (M1) is that of the simultaneous action: it consists in pressing the pre-key and without releasing the finger on the post-key then. The advantage of (M1) is that it alone is enough to distinguish between letters and numbers for which you just press a key, then release it before pressing a new one. In addition it allows an additional functionality, automatic repetition: by keeping the two keys pressed (resp. The key pressed), after a certain the letter (resp. The number) repeats. This feature exists on most computer keyboard. The disadvantage of (M1) is that the keyboard must be able to detect the simultaneous pressing of two adjacent keys, which moreover makes it unusable for the application (A2). On the other hand, this mode requires greater skill on the part of the user, in particular for combinations where the two keys are one above
The other.

- The other modes, which we call globally (M2), consist in pressing the two keys sequentially. The distinction between numbers and letters can be done in two ways, (M2.1) and (M2.2).

- In the case of (M2.1), press the same button twice in a row
key to generate a number.

- In the case of (M2.2) after a key has been pressed and then released
there is a time delay after which the machine
no longer waits for a second key is considered to be the number which
just entered.

Note that (M2.1) and (M2.2) can be active at the same time, so either (M1) or (M2) must be selected. Finally, the last subdivision of (M2) relates to the processing of the sequences of two nonadjacent keys, the pressures of which are in the case (M2.2) spaced a shorter time than the time delay. Thereafter M2.x means M2.1 or M2.2.

- In the case (M2.x.1) when the second key is pressed, the number
carried by the first is generated, and the machine then waits for pressure
of a key adjacent to the second.

- In the case (M2.x.2) the pressure of the second key is considered
as an error, and is therefore ignored. It would also be advisable if
this mode is active a beep warns the user.

 Examples: from Figures 2 and 3: the a is obtained by combination of 1 and 4, the p by combination of 3 and 4.1 'space by combination of * and 0.

 Finally we will examine how this interface reduces the defects that we have listed for the usual interfaces used: - (D1), the time, that is to say the complexity of typing, is a priori the same for each letter , perhaps it differs from numbers to letters.

- (D2) the selection of the letters is no longer pressed, it is no longer necessary to look at the screen for each letter.

- (D3) as we can put more than 9 letters per line (we can put 12 at most), it is possible to use the layout of a classic alphanumeric keyboard, for example azertyu as in Figures 2 and 3.

- Finally (D4), we can distinguish upper and lower case letters according to the order of support as we saw above.

Claims (7)

 1) Human-machine interface, characterized in that pressing adjacent combinations of keys on a keyboard, the keys of which are arranged in an orthogonal grid, makes it possible to generate the characters of a given alphabet, and / or functions (such as the correction of the last character entered, the sending of the message thus typed.), the adjacency being defined by considering the keyboard as a torus (the first line follows the first, ditto for the columns), and by considering the keys as adjacent to themselves.  2) Human machine interface according to claim 1, characterized in that the character, word or ideogram representing the effect of the key combination is printed on the keyboard: - (i) in the center of the key if the combination consists of an action on the same key (double press, or long press, or a press followed by a delay without further press, depending on the keyboard action mode), - (ii) between

 two keys, on the side of the first key pressed, if the

<tb> consists of an action on two distinct keys (successive presses by releasing, or without releasing the first before pressing the second, depending on the keyboard action mode), this premising the user to deduce I ' action it must take for the effect to be produced. <tb> corn  3) Human machine interface according to claim 1, characterized in that when two possible effects are the production of two different caligraphies of the same letter (for example the capital letter and the small letter) then on the one hand these effects are produced by action on the same two keys, the distinction is made on the order of pressure, and on the other hand a single character representing the two effects is printed on the keyboard between these two keys.  4) Human machine interface according to claim 1, characterized in that the combination of two separate keys is done by successively pressing the two keys without releasing the first key before pressing the second (overlapping clicks), while the combination of a key with itself is done by pressing and releasing this key (single click).  5) Human machine interface according to claim 1, characterized in that the combination of two separate keys is done by pressing and releasing a key and then the other (separate clicks), while the combination of a key with itself is done by pressing and releasing this key, then by repeating this action once (double click).  6) Human machine interface according to claim 5 characterized in that the combination of a key with itself made by a press-release, followed by a time configurable by the user during which no key is pressed (simple click followed by a silence), and conversely in the combination of two distinct keys the two clicks are spaced a time shorter than this same delay (distinct clicks closer together).  7) Human-machine interface according to claim 5 characterized in that the combination of a key with itself made by pressing and releasing, in which the time between pressing and releasing is greater than a time configurable by the user ( single long click), and in that, in the combination of two distinct keys, the first click is shorter than this same delay (first short click).