GB2470585A

GB2470585A - Using a predictive text module to identify an application or service on a device holding data to be input into a message as text.

Info

Publication number: GB2470585A
Application number: GB0909143A
Authority: GB
Inventors: Dean Ezra; Dileep Chalana Kaluaratchie
Original assignee: NEC Corp
Current assignee: NEC Corp
Priority date: 2009-05-28
Filing date: 2009-05-28
Publication date: 2010-12-01
Also published as: GB0909143D0

Abstract

Disclosed is a method for inputting data into text messages. The messaging module has an input analyser that analyses the text input by a user and identifies a service or application on the device eg smart phone or PDA, that can provide relevant dynamic user data to be input into the text message. The messaging module then uses its predictive text module to generate input text for insertion into the message using the data from the identified service. The services may include an address book, calendar, position locating system, media player and/or internet browser. The predictive text module may store data identifying any prediction in a message and change the data in the message if the data changes before the message is sent. Also disclosed are devices with messaging systems that use their predictive text module to generate static and dynamic text inputs., wherein the dynamic text is marked so that changes to the data can be identified before the message is sent.

Description

Messaging System The present invention relates to a messaging system and in particular to a text messaging system that provides text prediction to perform automatic completion of text messages.

The invention can be used in text messaging systems used on portable computer devices, such as on mobile telephones, PDAs, laptop computers, and also on general computers such as PCs and the like.

Predictive text messaging systems are well known and are predominately used in mobile telephones for creating messages using the Short Message Service (SMS), Multimedia Message Service (MMS) and email. These predictive text systems include the well known "T9" system that predicts the word that the user is typing from the first few characters of the word that the user types using their keypad. These systems use a static dictionary of words and the corresponding sequence of key presses required to enter each word.

Some systems allow users to add new words to the static dictionary, so that names and other missing words can be predicted in subsequent text predictions.

The present invention aims to provide an alternative text prediction system. Embodiments of the invention improve the existing text prediction systems by providing sentence completion that can add real time dynamic information into the sentence. In the event that the dynamic information changes, then the predicted text can be updated automatically at the time the message is sent. Embodiments of the invention can also perform the prediction on the basis of related text in addition to or instead of text input by the user.

The related text may include, for example, the text of the message to which the user is replying, the text of an article on which the user is commenting or the text of a meeting invitation.

The present invention provides an electronic device (such as a computer or cellular telephone) comprising: a messaging module for generating and sending messages to other devices; a plurality of user services for providing different services to the user and each service providing dynamic user data; a text input for receiving input text for generating a message using said messaging module; an input analyser operable to analyse the input text and to identify from the input text one or more user services that can provide dynamic user data relevant to the input text; and a predictive text module operable to generate predictions for the input text for insertion into the message using dynamic user data from said one or more user services.

In one embodiment, the input analyser splits the input text into sentences and identifies one or more sources of dynamic user data that may be relevant for each sentence. The input analyser may perform a pattern matching operation between the input text and stored data relating probable text to sources of dynamic user data. The pattern matching may be performed by a word or phrase spotting comparison between the input text and the text of the stored data or by performing a natural language analysis of the input text and using results of the natural language analysis and the stored data to identify the one or more sources of dynamic user data that may be relevant to the input text.

The user services may include one or more of: an address book, a calendar, a position locating system, a media player and a browser service for accessing on-line services.

Other service can, of course be included as well. Indeed new services can be added when they become available.

The text input may be received from user input (such as via a keypad, touch screen, hand writing recognition, speech recognition etc) or from related text. The related text can include the text of a message to which the user is replying and/or the text of an article on which the user is commenting.

The predictions generated by the predictive text module can include one or more of: text data, image/video data and audio data.

Typically, the messaging module will generate a message using static text such as text input by the user and one or more predictions made by the predictive text module. In this case, the predictive text module preferably maintains data identifying the one or more predictions that have been included in the message. The marking of the predictions in this way can help the predictive text module update the predictions should they change between the time that the predictions are inserted into the message and the time that the message is sent. The predictive text module may update predictions in this way automatically or in response to user confirmation.

The predictive text module can make use of services that are provided both on the electronic device and in a remote device, such as in a remote server. A communications module (such as a browser) is provided to facilitate the connection to such remote devices.

According to this aspect, the invention also provides a method of generating a message on an electronic device that has a plurality of user services stored thereon, the method comprising: receiving input text for use in generating the message; analysing the input text to identify one or more sources of dynamic user data associated with said user services that may be relevant to the input text; and generating predictions for the input text for insertion into the message using information from said identified one or more sources of dynamic user data.

The invention also provides an electronic device comprising: a messaging module for generating and sending messages to other devices; a text input for receiving input text for generating a message using said messaging module; and a predictive text module operable to generate predictions for the input text for insertion into the message; wherein the input text comprises related text that is not entered by the user. The input text may include both related text and text input by the user. The related text may comprises text extracted from a message to which the user is replying or text of an article on which the user is commenting.

The invention also provides an electronic device comprising: a messaging module for generating and sending messages to other devices; a text input for receiving input text for generating a message using said messaging module; and a predictive text module operable to generate predictions for the input text for insertion into the message; wherein said predictive text module is operable to generate predictions comprising static text and dynamic text that can change and is operable to mark dynamic text so that changes to the dynamic text that are made before the message is sent can be identified.

These and other various aspects and features of the invention will become clear from the following detailed description of exemplary embodiments that are described with reference to the accompanying drawings in which: Figure 1 is a schematic overview of the messaging system embodying the present invention; Figure 2a and Figure 2b illustrate the dynamic nature of the prediction used in the embodiment; Figures 3a to 3e illustrate various lists and data structures used by the predictive system embodying the invention; and Figure 4 is a block diagram illustrating the main components of a mobile telephone used in the system shown in Figure 1.

Overview Figure 1 schematically illustrates the operation of a messaging system embodying the present invention. In the illustrated embodiment, the messaging system is implemented on a mobile telephone 3, although it may of course be implemented on any computer system such as a PC or the like.

Initially, at step Si, the user will start the messaging service (for example an SMS/MMS/email application that is on the mobile telephone 3) and will start to input text, for example via a keypad or touch screen (not shown). The input text may also come from the message to which the user is replying (if appropriate). For example, the text in an incoming message asking Where are you?" could be used as input when predicting the text in the reply message. An input analysis module 4 then analyses the input text in step S2, by splitting the text into sentences and then by performing a pattern matching operation that compares the input text with a list 6 of probable words and phrases to identify if the mobile telephone 3 has or has access to any sources of information (ie other services on the telephone 3 or connected to the phone network that can be accessed) that could be used to form a relevant reply. Figure 1 illustrates some sources that may be used. As shown, the predictive text module 5 may obtain information from the address book 7 that can receive updates from on-line social networking sites, such as a "Twitter" update 9 or a "Facebook" update ii. The predictive text module 5 may also obtain information from the entries of the calendar that are stored in the calendar database 13.

The predictive text module 5 may also obtain information relating to the current location of the user. This location information 15 may be obtained from, for example, a GPS module 17 (and may be converted from a longitudellatitude position into a street address using an on-line address service 19), from the calendar 21 (that might identify that the user is currently in a meeting at a specified location etc), from the phone profile 23 (which may identify when the user is in a meeting or outdoors etc), or from a "twitter" update 25 that identifies, for example, that the user is on holiday is the USA. The predictive text module 5 may also obtain information from the media player 27, for example identifying the current music track 29 being played to the user.

In step S3, the predictive text module 5 uses the results of the input analysis step S2 to generate a list 31 of context based information for the input text that can then be passed to the identified source to generate a prediction. For example, the input text may include the sentence what restaurants are in reading', which the analysis module 4 determines is best dealt with by the address book 7. Instead of passing the whole input text string to the address book 7, the predictive text module 5 derives context information from the input text -in this example the predictive text module 5 would derive context information of Restaurant' and Reading', which can be passed to the address book 7 for making its prediction in step S4. As those skilled in the art will appreciate, a number of predictions may be available and not all of them may be relevant or desired by the user. Therefore, in this embodiment, before passing the context information to the selected source, the predictive text module 5 generates a predictions list 48 that lists all the predictions that have been or can be made and inserted into the message. For the above example, therefore, the predictive text module would generate an entry in the predictions list 48 that said "Insert names of restaurants from address book in Reading?". If the user selects this item in the predictions list 48, then the predictive text module 5 passes the relevant context information 31 to the address book 7 so that it can retrieve the names (and perhaps also addresses/telephone numbers) of all restaurants in Reading. If the mobile telephone 3 is aware of an on-line service that can provide this service such as Yellow Pages, then the predictive text module 5 may add the entry "Insert names of restaurants in Reading from Yellow Pages?" into the predictions list 48. In this case, if the user selects this option, then the predictive text module would pass the context information 31 to the remote Yellow Pages service and add the results to the message.

The predictive text module 5 may also use, in step S5, a conventional "T9" predictive text algorithm 35 for text input at step Si. As illustrated in Figure 1, the T9 algorithm can generate predictions by comparing the input text with a static database 37 that includes words from a dictionary 39, a custom words list 41 and a list of common words 43. The different predictions generated by the predictive text module 5 are then output, in step S6, for display to the user on the display of the mobile telephone 3.

Some of the dynamic information obtained from the services on the telephone 3 (or remote services to which the telephone 3 has access) is highly dynamic in nature -in other words it is changing frequently; and may change between the time that the user enters the text into the message and the time that the message is sent to the intended recipient. In some cases, it may be desirable to update this information at the time that the message is sent. To help achieve this, the predictive text module 5 keeps a record 47 of text that has been inserted into a message that comes from a dynamic source.

Then, at the time that the message is to be transmitted, the predictive text module 5 updates the information in the records 47 and compares the text in the current message with the text in the updated record 47. If the updated text is different, then in some cases, the predictive text module 5 will automatically update the text in the message; in others the predictive text module 5 will keep the original text and in others the predictive text module 5 will prompt the user to confirm if a change is required.

Examples of scenarios of how this updating may or may not be performed will now be given: Completing a message with my location (Figure 2a): 1. The user starts typing: Hi James, I am currently at 2. The predictive text module 5 recognises that the user is about to give a location, and so gives the end user an inline pull down list of the following dynamic data: * GPS location (E.g. Street name, postcode) * Location based on categories (Location is NEC, NEC is my contact 3. User selects one of the options and the text is added to the message being typed.

4. If the user's location changes whilst the user is writing the message, then the location will be automatically updated.

Completing a message with my current music track (Figure 2b): 1. The user starts typing: Hi James, I am listening to...

2. The predictive text module 5 recognises that the user is about to give the name of the music track being played on the phone 3, and so gives the user an inline pull down list of the following dynamic data: a. Current track: "Englishman in New York" by Sting.

3. User selects the option; and the text about the track is added to the message.

4. In contrast to the first example, this information should not be changed automatically if the current music track ends and another starts to play, because the user's message may be specifically related to the first track and not the second. Therefore this data will not change automatically. Instead, the predictive text module 5 may disregard the change given the current context or it may prompt the user to determine whether or not the track should be changed.

Thus, in this embodiment, the predictive text module 5 tracks the dynamic text that has been inserted into the message and decides whether or not to update the message or to prompt the user to make the decision, depending on the context of the dynamic information. The predictive text module 5 may be programmed to operate so that text from some sources are always or sometimes updated automatically; so that text from other sources are never updated; and so that the user is prompted to decide whether or not to update the text from other sources.

Detailed description

As those skilled in the art will appreciate, there are various ways in which the system described above may be implemented. The following description provides an example of how the system can be implemented. In this example implementation, it will be assumed that the user is replying to the following message: "Hi Dean, Do you know how to get to Baker Street from Paddington? James." Initially, the user selects the previous message and chooses reply to'. In response, the input analysis module 4 reads the previous message and parses the message to split it into sentences using end of sentence punctuation (full stop, question, comma and exclamation marks). Each sentence is then stored in an element of an array or list 51 that has a unique ID associated therewith, as illustrated in Figure 3a. The input analysis module 4 then processes each sentence in the array 51 in turn, to look for any matches of text in the input sentence with any of the words or phrases contained in the list 6 of probable words or phrases. Figure 3b illustrates part of the list 6 of probable words and phrases together with associated data. As shown, each entry of the list 6 includes: an entry ID (column 0); the probable word or phrase (column 1); the ID of one or more of the dynamic sources of information that will provide the prediction (column 2); a human readable description of the match that will be output to the user (column 3); a string that will be passed to the messaging application for inserting into the message (column 4); and a definition of the parameters that will be passed to the one or more sources identified in column 2 (column 5). The list 6 shown in Figure 3b only shows four probable phrases. In practice, the list 6 will contain a large number of probable words and phrases together with their associated data and will be pre-programmed into the mobile telephone 3 and can be updated whenever a new source of dynamic information becomes available.

The matching between the input text from list 51 is performed against the text in column 1 of the list 6 of probable words and phrases. This comparison can be done using regular expression pattern matching. When creating the regular expression, the input analysis unit 4 will replace any fields (i.e. words surrounded by < and > symbols) with regular expression wild cards (a wild card in the regular expression language is ** symbol). For example, the <destination> and <source> fields in the following original Column 1 string would be replaced: Original string: "Do you know how to get to <destination> from <source>?" String converted to regular expression: "Do you know how to get to * from The comparison between the input text and the converted string can then be performed to see if there is a match and if so to identify the destination and source included in the input text.

If a match is found, then the input analysis module 4 adds an entry (row) to a temporary list 53 of matches that is illustrated in Figure 3c. As shown, each entry in the list 53 of matches includes: in column 0 an entry ID for the match; in column 1 a pointer to the entry ID in the list 6 of probable words corresponding to the match; in column 2 a pointer to the entry ID in the list 51 of input text sentences corresponding to the match; and in column 3 the parameter values that will be passed to the corresponding dynamic source. The parameter values (column 3) are determined from column 5 in the list 6 of probable words and for any variadic parameters using the results of the match to identify the values to be used. For example, for the sentence "Do you know how to get to Baker Street from Paddington?", the input analysis unit 4 will find a match with entry 10003 in the list 6 of probable words and from the regular expression wild cards, will identify "Baker street" as the destination and "Paddington" as the source. Thus, when generating the entry in the list 53, the input analysis unit 4 will add the parameters "GetDirections", "Baker Street", "Paddington" to column 3 (as shown in Figure 3c).

Each input sentence may match with more than one entry in the list 6 of probable words and an entry is added to list 53 for each match that is found. The same processing is performed for each sentence in the input text. Once all the sentences in the input text have been processed in this way, the predictive text module 5 processes each entry in the generated list 53 by following the reference in column 1 back to the list 6 of probable words and then adding the value in column 3 (Human readable description of match) from list 6 as a new row in the predictions list 48 (shown in Figure 1). A pointer is also added to the predictions list 48, that points back to the corresponding entry in the list 53 of possible matches. This pointer of course is not displayed and is used to find the corresponding entry in the list 53 if the user selects that prediction. The predictive text module 5 then updates the user interface displayed to the user on the display of the mobile telephone 3 to show the updated entries in the predictions list'.

Once the initial processing of the related text has been performed, the predictive text module 5 monitors the user's key presses when he/she enters text into a displayed message' text box. For each character entered by the user, the predictive text module 5 checks if the key press represents the end of a new word (i.e. any of the following symbols: space, question mark, exclamation or comma). If the entered character does represent an end of word, then the predictive text module 5 will copy the text starting from the current character and the preceding characters until the beginning of the sentence (i.e. a full stop, comma, question mark, exclamation or first character of text entered is found) into a list 55 of sentences derived from user typing, which in this example is illustrated in Figure 3d.

The predictive text module 5 then repeats the matching process described above, to find any matches between the text input by the user (entered in the list 55 shown in Figure 3d) and the words and phrases in column 1 of the list 6 of probable words and phrases shown in Figure 3b. For any matches that are found, the predictive text module 5 extracts the information associated with the match from the list 51 and builds another temporary list 57 for the matches, which is shown in Figure 3e. As can be seen from Figure 3e, the type of information determined for each match from the list 6 of probable words and phrases is the same as the information determined when generating the list 53 of matches found in the related text. Once the new input sentence has been processed in this way, the predictive text module 5 processes each entry in the generated list 57 in the same way as for the entries in list 53 sO that the corresponding Human readable description of the match is added to the predictions list 48 and displayed to the user. Again, a pointer to the corresponding entry in the list 57 is added to the predictions list 48 sO that if that entry in the predictions list 48 is selected by the user, the system can trace back and find the corresponding entry in the list 57.

The above processing is repeated until the user stops entering text and/or selects an item from the predictions list 48 that is displayed to the user on the display of the mobile telephone 3. If the user clicks on an item in the predictions list 48, then the predictive text module 5 identifies the pointer for the selected item and uses it to find the corresponding entry in list 53 or 57. The predictive text module 5 then uses the pointer in column 1 of the entry in list 53/57 to identify the entry in the list 6 of probable words, from which the predictive text module 5 identifies the dynamic source to which the parameters in column 3 of list 53/57 will be passed to perform the prediction. For example, if the user selects the "Insert directions" entry in the displayed predictions list 48, then the predictive text module 5 will identify that this corresponds to entry 20001 in the list 53 shown in Figure 3c. The predictive text module then uses the pointer in column 1 of list 53 to identify that this entry corresponds to entry 10003 in the list 6 of probable words and phrases, which itself is associated with the GPS module 17 (identified from column 2 of list 6). The predictive text module 5 then passes the parameters listed in column 3 of entry 20001 in list 53 to the GPS module 17 and awaits the results. The GPS module 17 will receive the query and will generate the require directions in accordance with its conventional programming and return the results to the predictive text module 5. In response, the predictive text module 5 generates text to be inserted into the message using the results received from the GPS module to complete the fields of the text string stored in column 4 of entry 10003 in the list 6. The predictive text module 5 then inserts the resulting predicted text into the current cursor position of the message' box displayed to the user on the display of the mobile telephone 3.

Matching Process When performing the matching process discussed above, the input analysis module 4 may perform an exact or Boolean match that requires an exact match between a word or phrase in the input text and the word or phrase in the list 6 of probable words or phrases.

However, more sophisticated pattern matching techniques may be used that look for similar words or phrases or that allows for typing errors etc. In a more advanced implementation, Natural Language processing (often referred to as Linguistic Statistical Analysis') may be used to find possible predictions. Such an approach would score an input sentence/phrase or word based on the number of occurrences of a word associated with a given meaning. For example, the input analysis module 4 would understand all the possible different ways that a notional meaning could be written in a given language and then score the given sentence against how closely it matches the prediction. This approach would match a higher number of different ways of writing the same meaning.

However this approach will be more computationally intensive for mobile devices such as the mobile telephone 3 compared to simple string comparisons.

Updating predicted text As mentioned above, in this embodiment, the part of the predicted text that has been determined from a dynamic source may be updated when the message is to be sent and in order to allow this to happen in an efficient manner, the text added to the message is marked so that the system knows which parts may need to be updated. This is achieved by using dynamic field markers (<NAME>) that surround the text that may change. The messaging module understands that text surrounded by < and > symbols are not to be displayed to the end user. For example, If the string added to the message was derived from entry 1002 in list 6 and the text obtained from the dynamic source (in this case from the calendar 21) is "Career Review Meeting" then the text passed to the messaging module would be: "Yes, I am currently in <meeting> Career Review Meeting <\meeting>" The messaging module would ignore the text within the < and > symbols and would therefore display: "Yes, I am currently in Career Review Meeting" Just before the message is sent, the predictive text module 5 will walk through the text that was sent to the messaging module and will update values in the dynamic fields (for example, in the above scenario the text Career Review Meeting' might be changed).

Updating the Sources of Predictions One of the advantages of the embodiment described above is that the predictive text module 5 can call a number of different dynamic sources of information (using generic API queries) to provide information that can be added to a message as a prediction based on text that the user has input or based on related text (eg the text of a message to which the user is replying). As more sources of information become available to the mobile telephone 3, the system can be updated to allow the predictive text module 5 to obtain relevant information from the new source. In order to update the system so that it can work with a new service, all the system has to do is to update the list 6 of probable words to include appropriate sentences and parameter definitions that can be used with the new source. Such an update can be provided by a mobile telephone network operator, for example, and requires no user programming.

Benefits of the Embodiment The system described above has the following advantages: The system will use any available applications on the device that can provide information relating to the text being entered into the message or that has been previously entered. Current predictive text only uses a dictionary of words. This allows predictions to be more accurate due to a wider range of data that can be analysed.

The system can add real time dynamic information in a message that has previously been entered into other areas of the device (e.g. appointments in the Calendar application) without the user having to manually duplicate the information. What's more, the invention can suggest a human readable form to show the information, allowing the end user simply to select the prediction and add it without having to rewrite it. Prior art predictive text systems do not provide this functionality. For example, a plumber may enter all their house bookings into the Calendar application 21, but they would still have to manually look at the calendar 21 to find a free time if a new customer asked to book them in an email message. The described system would be able to predict a response to that email along the lines of "I am next available at 10:30am until 12:00am on 12/10/2009" by looking at the calendar database 13.

The system can use as input for analysis: information received by the device in addition to or instead of text entered directly by the end user. Prior predictive text systems only perform analysis on text that is directly entered by the end user in the current message.

Again, this allows predictions to be more accurate due to a wider range of data to analyse as input.

If the end user has selected to use a piece of information predicted by the system, and the information changes before the user sends the message, the system can elect to update the information. For example, if the system adds the user's current location and then the user moves, the system could update the location before the message is sent.

Prior predictive text systems do not change information once it has been inserted into the message.

Mobile Telephone Figure 3 schematically illustrates the main components of the mobile telephone 3 shown in Figure 1. As shown, the mobile telephone 3 includes transceiver circuitry 63 which is operable to transmit signals to and to receive signals from a mobile telecommunications network via one or more antennae 65. As shown, the mobile telephone 3 also includes a controller 67 which controls the operation of the mobile telephone 3 and which is connected to the transceiver circuit 63 and to a loudspeaker 69, a microphone 71, a display 73, and a keypad 75. The controller 67 operates in accordance with software instructions stored within memory 77. As shown, these software instructions include, among other things, an operating system 79, a messaging module 81, an API module 83, a browser module 85 and the above described input analysis module 4, predictive text module 5, address book 7, calendar 21, GPS module 17 and media player 27. The operating system 79 controls the general operation of the mobile telephone 3 and controls communications with the telephone network using the transceiver circuitry. The messaging module 81 provides the above described messaging functionality to the user (such as SMS/MMS/email). The messaging module 81 is operable to provide the user interface via which the user can enter text or other inputs and to use the predictive text and input analysis modules to generate the predictions in the manner discussed above.

The API module 83 is operable to provide the predictive text module with access (using a generic API) to the other services on the mobile telephone 3, such as the address book 7, the calendar 21, the GPS module 17 and the media player 27. The browser module 85 provides the predictive text module 5 with access to on-line sources of information that can be used for the prediction.

Modifications and Alternatives A detailed embodiment has been described above. As those skilled in the art will appreciate, a number of modifications and alternatives can be made to the above embodiment whilst still benefiting from the inventions embodied therein.

In the above embodiment, a mobile telephone based messaging system was described.

As those skilled in the art will appreciate, the message prediction techniques described in the present application can be employed in other computing systems such as, for example, personal digital assistants, laptop computers, web browsers, etc. Various sources of dynamic information were described above. As those skilled in the art will appreciate, various other sources may be used. The dynamic information that could be inserted could be extendable to any information available to the device (text, image data, video data, audio data, links or a mixture of these); whether it is information derived from the user's information inside other applications or information remotely accessible via services (e.g. Web services). Examples of different sources of dynamic information that may be provided by the prediction system are given below. It should be noted that this list is not an exhaustive or complete list of data sources.

Locations derived from GPS/AGPS module: Currently at: "14 high street, Reading" "The Beeches restaurant, London, SW1 4GF" "15 metres to the left of you" (information about the other user's position and orientation may be provided in the message to which the user is responding or may be provided by an on-line service that track subscribing users locations and orientations.) "Walk 10 metres along high street, turn Right at Bond Street." Location derived from the handset's active profile: E.g. work', home'.

Calendar derived information: "Currently in meeting with John Smith, Jane Doe" "Will be free at: 14:3Opm" (in reply to can we have a meeting at 14:3Opm on 12/12/08?" : "Meeting confirmed" (calendar entry can be automatically added).

Contact derived information: User starts typing John All John's contacts are offered to the end user.

Data from Online sources: In reply to "what was that movie with James Dean?", would look up James Dean in online database.

"I have an Amazon wish list with the following item: "Links could be added" -Amazon wish list of items you want for Christmas.

Predictions based on intelligence gained from data providers/services: Time to destination: I will arrive at 10:30am based on received traffic information.

Automatic calculator: End user starts typing a sum (5+1), text auto completes with answer.

Current music track being played, auto insert their most popular genre, artist etc. Insertion of end user's Mood, presence etc. Insertion of media (pictures, music, url links etc): End user mentions key words relating to a photo they took on their mobile telephone, system identifies the keyword and offers to add the photo (which it found by searching photo filenames and/or meta data associated with the photos).

Derived data via the vicinity of other handsets: If a person is writing the minutes of a meeting, the system can identify other users that are present (for example by performing a Bluetooth scan for other devices that are within range of the mobile telephone 3) and then obtain their names from the corresponding contact details given in the address book 7.

Checking of files being attached If an end user writes Enc.' or attached at the end or in the middle of an email message, the system can check to ensure that the file is attached and prompt user to add one if it is not.

In the above embodiment, the predictive text module 5 received input text either from the user or from the message to which the user is replying. Other sources related text for the input can be used. For example, the user may be responding to an on-line article. In this case, the predictive text system may use the text of the article as input text for determining predictions. As a further alternative, the input text may be obtained from a meeting request issued by the calendar and to which the user is responding.

In the above embodiments, a number of software modules were described. As those skilled in the art will appreciate, the software modules may be provided in compiled or un-compiled form and may be supplied as a signal over a computer network, or on a recording medium. Further, the functionality performed by part or all of this software may be performed using one or more dedicated hardware circuits. However, the use of software modules is preferred as it facilitates the updating of the computer system.

Additionally, the functionality of one or more of the software modules described above may be combined into a single module or split into a number of different modules if desired. For example, the novel predictive text functionality described above may be provided as a standalone software module or it may be incorporated with an existing predictive text application.

Various other modifications will be apparent to those skilled in the art and will not be described in further detail here.

Claims

Claims: 1. An electronic device comprising: a messaging module for receiving messages from other devices and for generating and sending messages to other devices; a plurality of user services for providing different services to the user and each service for providing dynamic user data; a text input for receiving input text for generating a message using said messaging module; an input analyser operable to analyse the input text and to identify from the input text one or more user services that can provide dynamic user data relevant to the input text; and a predictive text module operable to generate predictions for the input text for insertion into the message using dynamic user data from said one or more user services.
2. A device according to claim 1, wherein said input analyser is operable to split the input text into sentences and to identify one or more sources of dynamic user data for each sentence.
3. A device according to claim 1 or 2, wherein said input analyser is operable to perform a pattern matching operation between the input text and stored data relating probable text to sources of dynamic user data.
4. A device according to claim 3, wherein said pattern matching operation performs a word or phrase spotting comparison between the input text and the text of the stored data.
5. A device according to claim 3, wherein said pattern matching operation is operable to perform a natural language analysis of the input text and to use results of the natural language analysis and said stored data to identify the one or more sources of dynamic user data.
6. A device according to any of claims 1 to 5, wherein said user services includes one or more of: an address book, a calendar, a position locating system, a media player and a browser service for accessing on-line services.
7. A device according to any of claims 1 to 6, wherein the text input is operable to receive input text from user input or related text.
8. A device according to claim 7, wherein the related text includes one or more of: the text of a message to which the user is replying and text of an article to which the user is commenting.
9. A device according to any of claims 1 to 8, wherein the predictive text module is operable to generate predictions comprising one or more of: text data, image data and audio data.
10. A device according to any of claims 1 to 9, wherein the messaging module is operable to generate a message using text input by the user and one or more predictions made by the predictive text module and wherein the predictive text module is operable to store data identifying the one or more predictions that have been included in the message.
11. A device according to claim 10, wherein the predictive text module is operable to identify changes in the predictions so that the predictive text module can change the one or more predictions.
12. A device according to claim 11, wherein the predictive text module is operable to change the one or more predictions when the message is to be sent.
13. A device according to claim 11, wherein said predictive text module is operable to select whether or not to update the one or more predictions in dependence upon the source of the dynamic user data.
14. A device according to any of claims 1 to 13 that is a cellular telephone, a personal digital assistant, a computer or the like.
15. A device according to any of claims 1 to 14, wherein one of said user services provides access to one or more remote services that maintain dynamic user data that can be used for making said predictions.
16. A method of generating a message on an electronic device that has a plurality of user services stored thereon, the method comprising: receiving input text for use in generating the message; analysing the input text to identify one or more sources of dynamic user data associated with said user services that may be relevant to the input text; and generating predictions for the input text for insertion into the message using information from said identified one or more sources of dynamic user data.
17. An electronic device comprising: a messaging module for receiving messages from other devices and for generating and sending messages to other devices; a text input for receiving input text for generating a message using said messaging module; and a predictive text module operable to generate predictions for the input text for insertion into the message; wherein input text comprises related text that is not entered by the user.
18. A device according to claim 17, wherein said input text comprises said related text and text input by the user.
19. A device according to claim 17 or 18, wherein said related text comprises text extracted from a message to which the user is replying or text of an article on which the user is commenting.
20. An electronic device comprising: a messaging module for receiving messages from other devices and for generating and sending messages to other devices; a text input for receiving input text for generating a message using said messaging module; and a predictive text module operable to generate predictions for the input text for insertion into the message; wherein said predictive text module is operable to generate predictions comprising static text and dynamic text that can change and is operable to mark dynamic text so that changes to the dynamic text that are made before the message is sent can be identified.
21. A device according to claim 20, wherein said predictive text module is operable to update automatically dynamic text that changes before the message is transmitted.
22. A device according to claim 20, wherein if said predictive text module identifies that the dynamic text has changed, then the predictive text module is operable to query the user if the dynamic text should be updated and is operable to update the dynamic text if the user selects that the text is to be updated.
23. A computer implementable instructions product comprising computer implementable instructions for causing a programmable electronic device to become configured as the device of any of claims 1 to 15 or 16 to 22.