WO1998059302A1

WO1998059302A1 - Method and data processing device for inputting an input sequence and method for accessing a storage unit and a corresponding storage unit

Info

Publication number: WO1998059302A1
Application number: PCT/DE1998/001754
Authority: WO
Inventors: Frank Karl
Original assignee: Frank Karl
Priority date: 1997-06-23
Filing date: 1998-06-23
Publication date: 1998-12-30
Also published as: AU8533498A; DE19726569C1

Abstract

The invention relates to a method for operating a data processing device, in which a short-term storage (48) is used. When the short-term storage (48) is read, both the content of the stored memory word and the time sequence of access to the memory word are taken into account.

Description

Method and data processing system for entering an input sequence and method for accessing a storage unit and associated storage unit

The invention relates to a method for operating a data processing system, in which a data input unit generates an input sequence from data words with a predetermined number of bit positions. Different characters are displayed on a display unit depending on the value of the data words. At least one predefined value characterizes separating data words which subdivide the input sequence into input sections. Several of these input sections are stored in a memory. In addition, depending on the new data words generated since the last separation data word, an input section belonging to the new data words is read at least once during the generation of the input sequence and is automatically appended or inserted at the end of the input sequence.

In the book "Word 97 - The Big Book" by U. BretSchneider and B. Matthies, Data Becker, 1997, on pages 462 to 464 an "AutoComplete" procedure is explained, in which letters are entered using a keyboard. As soon as more than four letters of a word have been entered, a memory is searched for a word which begins with the letters already entered or to which the letters already entered are assigned as an abbreviation. Such words belonging to abbreviations are also referred to as "AutoText" or "text module".

A disadvantage of the known method is that only four letters per word have to be entered until a word can be clearly selected from the memory, which is then inserted completely or partially into the input sequence. The known method, as indicated, for example, on page 69 of the book mentioned, can be perceived by some operators to be rather confusing than helpful, so that it is recommended that these persons switch off the "auto-complete" method. As confusing is felt especially when the automatically inserted letters are not confirmed and therefore have to be removed.

It is an object of the invention to provide a simple method for entering an input sequence, in which a high percentage of the insertions can remain in the input sequence.

This object is achieved by a method having the features of patent claim 1. Advantageous further developments are specified in the subclaims which refer back to this claim.

The invention is based on the knowledge that there are input sequences in which the probability that an input section repeats decreases with the number of data words entered since the input section was entered. In other words: there is a high probability that an input section repeats itself after the input of a few data words. If the input units e.g. generated when entering a text, words that already occurred when another font was entered do not appear as often in the current font as words that have already been entered in the current font. Words that occur in a paragraph of the current font that has already been entered do not appear as often in the paragraph that is currently entered as words that have already been entered in the paragraph that is currently entered. The same applies to different sentences.

In the method according to the invention, the input section or a part of the input section which was last accessed is therefore read from the memory from a plurality of input sections which belong to the new data words entered since the last separation data word. Access means reading the memory only if no input sections are written to the memory during the method. However, if additional entries are saved in the memory during the procedure, the saving is also an access.

Thus, when the stored input sections are accessed, the time of the last access to the input sections is taken into account in addition to the sequence of letters within the input sections. This alphabetical and chronological procedure allows you to automatically insert a word residue after entering the first letter or, in the case of an abbreviation, an entire word that is highly likely to be confirmed. The probability of wrong decisions also decreases disproportionately if two or even three letters have to be selected, taking into account the last access, if the suggestion made after entering the first letter or the first two letters of the currently entered word is not confirmed has been and had to be removed. The disproportionate decrease in the probability of wrong decisions includes to be attributed to the fact that with increasing number of letters of a word already entered there are fewer associated input sections. In addition, the number of data words entered between the associated input sections increases, so that the number of wrong decisions is reduced due to the above-mentioned temporal relationships.

In a development of the method according to the invention, the memory is only searched when the first new data word has a specific value from a predetermined value set. As a result of this measure, the memory takes several process steps to be accessed only for selected inputs, and the implementation of the method is accelerated. This is important for a method that is carried out, so to speak, in real time for the input of the input sequence, because otherwise there may be delays in the input. This measure can also increase the likelihood of the entries being confirmed. For example, only when entering nouns in start the German language with a capital letter, made a proposal by the method according to the invention, so very high confirmation rates can be achieved. When entering a program text in a programming language in which the program commands begin with uppercase letters, almost all automatically inserted insertions are confirmed.

The method according to the invention is further accelerated if the memory is searched only if the number of data words entered after the input of the last separation data word does not exceed a predetermined number. E.g. If only a maximum of two suggestions are inserted per word, the memory no longer has to be searched after entering the third letter. The copying processes for inserting the proposal and removing them in the absence of confirmation are also eliminated. This measure is also advantageous because, in the case of two or three unsuccessful suggestions within one word, the relationships explained above obviously do not apply.

In another development of the method according to the invention, after an entire input section has been entered, it is checked whether the input section is already stored in the memory. If this is not the case, the input section is automatically added to the memory. With this measure, the automatic entry of input sections takes place immediately after the first use of the relevant input sections. Without this measure, the entry is usually made manually and only after the input section, e.g. a word that has already been entered several times. In the case of automatic entry, the second time the entry section is entered, e.g. with a keyboard, keystrokes can be saved.

If, when the memory is supplemented, a predetermined number of input sections in the memory are not exceeded by overwriting the input section which has not been accessed for a long time, a short-term memory is created. This Memory has the advantage that it "forgets" input sections that have not been accessed for a long time, that is, automatically deletes them. This is advantageous, for example, for words that have been stored in the wrong spelling in the memory or if input sequences for new subject areas, ie also with a changing vocabulary, are entered. The method according to the invention with a short-term memory is very powerful in terms of the confirmation rate with a small memory requirement. Good results can be achieved with a memory for fifty or a hundred input units. Such a small memory can also be searched very quickly.

The method according to the invention can be used advantageously if input sections are repeated in the input sequence, e.g. is the case when entering legal or technical texts. The method according to the invention is also advantageously used in word processing programs for inputting programs in programming languages such as C, C ++, Java, etc., since there are only a predetermined number of upper and / or lower case keywords in the respective programming language, depending on which entered program still a relatively limited number of recurring variable names.

The time sequence of the accesses can advantageously be taken into account if the input sections form an input section sequence in which the input sections are arranged depending on the last access time. Alternatively, certain groups of input sections, e.g. all input sections beginning with the letter "A" can be contained in different input section sequences. Since the individual input section sequences do not contain as many input sections, the time for searching is shortened.

The invention also relates to a data processing system for carrying out the method according to the invention or its further educations. The technical effects mentioned also apply to the data processing system. The known access methods are used to realize the access to the memory, such as associative memory, linked data lists or hash search methods.

In a further aspect, the invention relates to a method for accessing a storage unit and this storage unit itself. This access method and the storage unit are used in particular when executing the inventive method for entering the input sequence. The special feature of the access method according to the invention and the memory unit according to the invention is that, in addition to the known assignment, e.g. via an address or, as in the case of an associative memory, via the content of the memory cells, the chronological order of access to the memory words is also taken into account.

In contrast to a known memory unit, several memory words can belong to one input word. On the basis of the chronological selection criterion, one is selected in the memory unit according to the invention in the case of a plurality of memory words which belong to an input word and is output completely or only partially. The selection is preferably made within the memory unit, so that essentially only the input word has to be entered from the outside and the output word has to be read. An access unit, such as a processor, is therefore no longer required than in conventional memory accesses. The time selection takes place, for example, according to the LIFO principle (last in first out) or according to the FIFO principle (first in first out). In the first case, one of the memory words belonging to the input word is selected which has not been accessed for the longest time. In the second case, the memory word that was last accessed is selected from the associated memory words. The storage unit according to the invention already allows an output word to be output when the input word has only been partially defined and entered into the memory. The access time when reading a memory word is reduced in cases in which a suitable output word is already read out, although the input word has not yet been completely entered. An example of an application for the storage unit and the access method is, for example, the input of a sequence of letters in which, depending on the part of the word already entered, a new word or a remainder of the word is proposed, which is confirmed or rejected by an operator.

If the storage unit contains an associative memory, the advantages of associative access are used, e.g. the short access time. Associative memories are commercially available memory modules which are also referred to as "content addressable memory". Depending on the input word, a memory word is determined which contains the input word. Then either the entire determined memory word or only a part of it is output.

Exemplary embodiments of the invention are explained below with reference to the accompanying drawings. In it show:

FIG. 1 essential electronic functional units for entering an input sequence,

2 shows an auxiliary file which is used when entering the input sequence, and

FIGS. 3A and 3B show a flow diagram with method steps which are carried out when the input sequence is input.

FIG. 1 shows essential electronic functional units for entering an input sequence with a commercially available keyboard 10. FIG. 1 shows five keys 12 to 20 of the keyboard keys, which are connected to a keyboard controller 22 contained in the keyboard 10 via connecting lines 24 to 30. The key 16 is used, for example, when entering the letters "a" and "A", in the latter case the key 12 must also be pressed.

The keyboard 10 is connected via a multi-core line 34 to a computer 40 which, among other things. includes a processor 42 and memory 44. The processor 42 processes the commands of a word processing program 46 stored in the memory 44, a text being entered into an input file using the keyboard 10. In other words: the processor 42 is controlled by the method steps stored as an instruction sequence in the memory 44 as an instruction sequence. The memory 44 also contains a short-term memory 48, in which input sections of an input sequence entered with the keyboard 10 are stored. The word processing program is e.g. the well-known program "WinWord" or "Word 97", which can be expanded with the associated macro programming language by the macro blocks explained below.

When each key 12 to 20 is pressed, a letter macro is processed, of which letter macros 50, 52 and 54 are shown in FIG. 1, which in this order belong to the letters "a", "b" and "Z". The letter macro 50, 52, 54 belonging to the just actuated key 12 to 20 transfers the ASCII code (American Standard Code II) of the entered character to a main macro 56 and then starts the execution of the main macro 56 explained below with reference to FIGS. 3A and 3B.

The computer 40 is connected via a connecting line 58 to a screen 60 on which the letters entered are displayed. The current end of the input sequence is shown on the Screen 60 is displayed by a cursor 62. It can be seen on the screen 60 shown in FIG. 1 that the letter "A" has already been entered. The word extension "bsatz" has then been inserted automatically and can be confirmed by the operator by entering a "7". Details of these operations are explained below.

FIG. 2 shows an auxiliary file 100 which is used when the input sequence is input into the input file. The lines of the auxiliary file 100 are labeled ZI to Z6. In line ZI the ASCII data word of the key just pressed is transferred to the main macro 56, compare FIG. 1. In FIG. 3 the key "7" was just pressed so that the transferred data word has the value "55" according to the ASCII code . This value leads to the display of the "7" when the auxiliary file 100 is displayed on the display unit 60 (see FIG. 1).

For explanation purposes, assume that the word "paragraph" is to be entered. The initial letter "A" was already entered before "7", which indicates the confirmation of an inserted proposal. The letters of a currently entered word that have already been entered are stored in line Z2 of the auxiliary file 100. This is the letter "A" in Figure 2. The method explained below with reference to FIGS. 3A and 3B automatically inserts suggestions into the input sequence that are read from the short-term memory 48, see also FIG. 1. The content of the short-term memory 48 is shown in lines Z5 and Z6 of the auxiliary file 100.

After entering the letter "A", the word "paragraph" was determined in line Z5 of the short-term memory 48 and the word residue "bsatz", which had not yet been entered, was automatically added to the current end of the input sequence in the input file. The appended word residue is stored in line Z4 of the auxiliary file 100 for further processing. In line Z3 of the auxiliary file 100, the value of a variable IMWORD is stored which is the current one Has value one. The value one indicates that a noun is currently being entered, ie for a German text a word that begins with an uppercase letter. If no noun is currently entered, the variable IMWORT has the value zero. The auxiliary file 100 also contains a movable text mark TM, which is moved within the short-term memory 48 in order to avoid identical suggestions during the input of a word.

If the "7" key is entered while entering a noun, this means that the inserted suggestion is accepted so that it can remain in the input sequence. If a "7" is entered outside of a noun, this is not counted as confirmation and the "7" is entered normally in the input file.

FIGS. 3A and 3B show a flowchart with method steps which are processed by the processor 42 when executing the instructions of the main macro 56, compare FIG. 1. In the explanation of FIGS. 3A and 3B, reference will also be made below to FIGS. 1 and 2 without express reference taken. The method begins in a step 150 in which the variable IMWORT is assigned the value stored in line Z3 of the auxiliary file 100. In addition, the currently entered data word, ie the currently entered letter or character, from line ZI of the auxiliary file 100 is stored in a variable b $. The previously entered data words of the currently entered word stored in line Z2 of the auxiliary file 100 are assigned to a variable w $. The appended word remainder stored in line Z4 is assigned to a variable r $. The method is accelerated if the assignments mentioned are only carried out in another exemplary embodiment when the variables are actually required. Step 150 is always carried out immediately after a letter macro 50 to 54 has been processed, see FIG. 1. In step 152 following step 150, it is checked whether the currently entered data word, ie the variable b $, has a value which identifies separating data words. Separation data words are generated when entering a space, a comma or a period. If it is determined in step 152 that the current data word b $ is not a separating data word, a step 154 follows in which it is checked whether the variable IMWORT is zero. This can only be the case if the first letter of a word has just been entered. If this is the case, then in a step 156 it is checked based on the data word b $ whether the first letter is an uppercase letter and thus a noun is currently to be entered. If the data word b $ belongs to an uppercase letter, the variable IMWORT is set to the value one in a step 158. Step 156 is omitted in another exemplary embodiment, in which suggestions can be generated when all the words entered are entered, so that, for example, repetitive verbs or nouns written in lower case, for example, are processed in the same way as in the exemplary embodiment in FIGS. 3A and 3B is the case for nouns capitalized in German.

If, on the other hand, it is determined in step 156 that the first letter of the word currently entered is not an uppercase letter, a step 160 follows immediately after step 156 and the variable IMWORT remains at the value zero. In step 160, the processing in the input file is continued. The execution of the main macro 56, compare FIG. 1, is thus ended for the currently entered letter, since a second part shown in FIG. 3B is only processed if the variable IMWORD has the value one.

If, on the other hand, it is already determined in step 154 that the variable IMWORT has the value one, this means that a noun is currently being processed. The procedure in this case immediately continued with the second part shown in Figure 3B.

If it is determined in step 152 that the currently actuated key belongs to a separating data word, a method step 162 follows immediately after method step 152. In step 162 it is checked whether the variable IMWORD has the value zero. If this is the case, no noun is currently being processed and step 164 follows, in which the input file is returned to. The second part shown in FIG. 3B is then not carried out.

If, on the other hand, the variable IMWORT has the value one when queried in step 162, a step 166 follows immediately after step 162. Step 166 is only carried out if the processing of a noun has ended with the current input. In step 166, the variable IMWORT is set to the value zero to mark the completion of the entry of the noun. A step 168 then follows.

In step 168, the bookmark TM, cf. Figure 2, placed at the beginning of the short-term memory 48, i.e. to the beginning of line Z5 of the auxiliary file 100. This is necessary because the bookmark TM is displaced during the process in order to avoid repetitive suggestions in the case of several suggestions per word.

In the following step 170, the short-term memory 48 is searched for the noun stored in the variable w $, i.e. after the last word entered. The short-term memory 48 is always searched during the search to the end of the file of the auxiliary file 100, i.e. towards the right end of line Z6.

It is then checked in a step 172 whether the noun contained in the variable w $ has been found. If this word was not found, step 174 follows immediately after step 172, in which the current processing position is set to the bookmark TM in the auxiliary file 100, which, as already mentioned, was placed in step 168 at the beginning of the short-term memory 48. At this point, the noun stored in the variable w $ is inserted, ie written into the short-term memory 48.

In a subsequent step 176, the processing position in the auxiliary file 100 is moved to the end of the auxiliary file 100. And in step 180, the word stored at the end of the auxiliary file 100 and thus also at the end of the temporary memory 100 is deleted, compare step 178. The entry of this word was a long time ago, so that it is relatively unlikely that this word will be used again soon is entered. In addition, the number of words stored in the short-term memory 48 remains constant if a word which has already been stored is deleted for each new word inserted into the short-term memory 48.

In step 180, the word stored in the variable w $ is deleted. In addition, in this step, a remainder of the word, possibly stored in the variable r $, which contains the last suggestion, is deleted. The variable r $ only contains a suggestion if such a suggestion was made while entering the last noun entered.

If, on the other hand, it is determined in step 172 that the last entered noun has been found when searching the short-term memory 48, this noun is removed from the short-term memory 48 in step 182, the words in the short-term memory 48 adjacent to the noun to be removed closing the gap that arises. This is due to the fact that the data words of the auxiliary file 100 are managed in a manner similar to a so-called linked list.

Then in steps 184 and 185 the removed noun, which is also contained in the variable w $, is inserted again at the beginning of the short-term memory 48. The number of in Short-term memory 48 words thus remains unchanged here. In the following step 186, as in step 180, the noun stored in the variable w $ and the remainder of the word possibly stored in the variable r $ are deleted. Step 186 is followed by step 164, already explained, in which the input of the input sequence in the input file is continued.

The short-term memory 48 stores the last words entered in line Z5 and the first words entered in line Z6. The last words entered are to the left of a line. So the word "Berlin" was entered before the word "bus", see Figure 2. The entry of the word "car" is furthest back. The words in the short-term memory 48 thus form a sequence in which the position of the respective word is determined by the access time at which this word was last accessed. If you access the access times on a timeline, the order of the times corresponds to the order of the words belonging to the times in the sequence.

As already mentioned, the short-term memory 48 is searched line by line from top to bottom. The search within the lines is from left to right. A search order is thus defined in the short-term memory 48. This search sequence takes into account the chronological order of access in the case of several words which contain the letter sequence sought, starting from the text mark TM. From the words stored between the bookmark TM and the end of the short-term memory 48 with the searched letter sequence, the word that was last accessed is found.

A second part of the method shown in FIG. 3B is only processed if it is determined in a step 190 that the variable IMWORT has the value one and thus a noun is currently being entered. Otherwise, editing the last entered letter ended in a step 192 explained below.

If it is determined in step 190 that the variable IMWORT has the value one, a step 194 follows, in which it is checked whether the currently entered character is a "7", i.e. see FIG. 1 to determine whether the key 20 was pressed. A “7” is generally never entered when the letters of a noun are entered correctly, so that the “7” that is easy to enter on the keyboard 10 when entering the input sequence was selected as the confirmation character , which indicates that a previous proposal has been accepted. If the proposal is accepted, a step 196 follows immediately after step 194. In step 196, the proposal stored in the variable r $ is added to the letter sequence of the noun currently entered in the variable w $. The variable w $ thus contains all previously entered or confirmed letters of the noun just entered, so that the current input status is always stored in the variable w $ when the noun currently entered is edited further.

In the following step 198, a return is made to the input file and any "7" that may have already been inserted into the input sequence is removed, since it was obviously only entered as confirmation, step 200. Then in a step 202 the current input position is added to the end of the one already inserted Proposal. An operator can now either continue entering the noun to be currently entered or end the input of this noun with a separating data word.

If, on the other hand, it is determined in step 194 that no "7" has been entered, it can be assumed that either no proposal has been made or that an already inserted proposal has not been confirmed. In both cases, a step 204 follows immediately after step 194, in which it is checked whether the currently entered noun already has more than two letters Has. This query makes it possible to suppress suggestions after the second letter, so that a maximum of two suggestions are made for each noun entered. The suppression takes place in that the current processing position in the auxiliary file 100 is set at the end thereof, and thus no word can be found when searching the short-term memory 48 starting at the current processing position, step 206.

If, on the other hand, it is determined in step 204 that the noun currently entered does not yet contain more than two letters, the processing position in the auxiliary file 100 is set in step 208 to the position at which the bookmark TM is currently located. During the first search within a word, the bookmark TM is located at the beginning of the short-term memory 48, where it has been moved to after the last noun has been entered, cf. Step 168. The bookmark TM is located at a different location, if it has already been moved during the input of the currently entered noun, compare step 216 explained below.

In another embodiment, the number of suggestions per noun entered is not limited. In this case, steps 204 to 208 are omitted and step 210 immediately follows step 210 if no "7" has currently been entered.

Step 206 is followed, like step 208, by step 210, in which the noun part previously stored in the variable w $ is supplemented by the currently entered letter, which is stored in the variable b $. Subsequently, in step 212, the short-term memory 48 is searched from the current position of the bookmark TM for a stored word which contains the letter sequence stored in the variable w $. As already mentioned, the search starts from the current position of the bookmark TM if step 208 has been carried out immediately before step 210. Otherwise it will be a positive one Search result suppressed by starting the search at the end of the short-term memory 48.

In step 214 it is checked whether the search was successful. If so, i.e. If a word was found in the short-term memory 48, the processing is continued in step 216. In step 216, the bookmark TM is placed behind the found word in the short-term memory in the search direction, so that this word cannot be found again in the next search when further letters of the currently entered noun are entered. This measure prevents the same suggestion from being made twice while entering a noun. After step 216, step 218 follows.

In step 218 it is checked whether the word found in the short-term memory 48 contains other letters in addition to the letters contained in the variable w $. If this is not the case, i.e. if the word found is identical to the sequence of letters stored in the variable w $, no new suggestion can currently be made, and the processing is continued in step 220. In an alternative exemplary embodiment, however, instead of step 220 and the steps following this step, step 212 is returned, and the short-term memory is searched again in order to eventually find another word which, in addition to the letters contained in the variable w $, contains further words Contains letters, see arrow 221.

In step 220, the variable r $ is used to check whether an old suggestion was inserted into the input sequence when the method steps shown in FIGS. 3A and 3B were last processed. If this is the case, the letters in the variable r $ must be removed, since the old proposal has obviously not been accepted. These letters are deleted in step 222. The input file is then returned in step 224. In step 226, the old proposal, which is no longer valid, is removed from the input sequence. After that The processing of the last letter entered is ended in step 192.

If, on the other hand, it is determined in step 220 that when the process steps shown in FIGS. 3A and 3B were last processed, no old suggestion was made, step 228 follows immediately after step 220, in which a return is made to the input file. The processing of the letter currently entered is then ended in step 192.

If it is already determined in step 218 that the word found in the short-term memory 48 contains, in addition to the letters contained in the variable w $, further letters which are then used as a new suggestion, step 230 follows immediately after step 218. In step 230 the variable r $ is used to check whether an old suggestion is contained in the input file which was obviously not accepted, since otherwise steps 194 to 202 would have been branched to in step 194. If an old suggestion is present, the letter sequence of the old suggestion in the variable r $ is overwritten by the letter or the letter sequence by which the word found in the short-term memory 48 differs from the content of the variable w $, i.e. through the new proposal, compare step 232.

In the following step 234, the input file is returned and the old proposal is overwritten with the new proposal stored in the variable r $, step 236. The processing of the letter currently entered is then ended in step 192.

If, on the other hand, it is determined in step 230 that no old proposal is contained in the input file, step 238 follows immediately after step 230, in which the new proposal is stored in the variable r $. The input file is then activated in step 240 and the new proposal is inserted at the current input position, step 242. After Step 242 is immediately followed by step 192, in which, as already mentioned, the processing of the letter currently entered is ended.

However, if it is already determined in step 214 that no word has been found in the short-term memory 48 that contains the part of the noun that has already been entered or is identical to this part, a step 220 ′ follows the step 214 immediately after step 214 220 corresponds, but was shown again for reasons of clarity. Steps 222 ', 224', 226 'and 228' following step 220 'also correspond in this order to steps 222, 224, 226 and 228 already explained above. After processing steps 222', 224 'and 226' or Step 228 'is immediately followed by step 192, in which the processing of the letter currently entered is ended.

At the end of step 192, the current values of the variables w $, r $ and IMWORT are stored in the auxiliary file 100 so that they can be used when processing the next letter or character entered.

The method shown in FIGS. 3A and 3B is carried out when the input sequence is input for each input letter and for selected or all input characters. The entered letters and characters are processed immediately after entry, similar to what is known as real-time processing. This is the only way to insert or remove suggestions in the input sequence without delay if they are not confirmed.

The procedure for automatically generating suggestions when entering an input sequence can also be carried out without macros. The method steps explained with reference to FIGS. 3A and 3B are contained in the instruction sequence of the word processing program 46. Instead of the auxiliary file 100 is in this If a memory area is used in the computer 40 according to FIG. 1. Switching between auxiliary file 100 and input file is then also omitted, since the task of auxiliary file 100 is carried out by the memory area. In another exemplary embodiment, the auxiliary file 100 is not a so-called document that can be edited like the input file, but a storage file in the memory 44 of the computer 40 or on its hard disk memory.

In a further exemplary embodiment, corrections made are also taken into account when entering a noun. This results in the possibility of accepting changes in the short-term memory 48 after confirming a proposal, e.g. if the ending of the proposed noun is changed.

When the method according to the invention was tested, a short-term memory with fifty words was used. The correctly proposed noun additions accelerated the input of the text considerably, since a large number of keystrokes was omitted, especially for the long words. A pleasant signal tone generated by the computer 40 for each suggestion allowed the operator to use the advantages of the method according to the invention without having to constantly monitor the display unit 60. Even with over 250 keystrokes per minute and a processor 42 with a clock rate of 66 MHz, there were hardly any delays due to the method steps explained above.

Instead of managing the words in the short-term memory 48 via a linked list, in another exemplary embodiment a table is used for administration. As with the linked list, the words retain their position within the sequence of their memory cells in memory 44 and 48, respectively, so that copying operations are eliminated. The table contains the address values of the memory cells in which the words are stored as well as the position of the words within the sequence assigned to these address values. Using the table are preferably both assigned certain addresses to positions in the sequence as well as vice versa to determined positions in the sequence addresses in the memory.

Access to the memory 44 can be simplified by using a table called a mirror table. While the table for consecutive memory addresses stores the positions that are generally not consecutive within the sequence, the mirror addresses for consecutive positions within the sequence store the memory addresses that are generally also not consecutive.

Claims

claims

1. Method for operating a data processing system (40),

in which a data input unit (10) generates an input sequence from data words with a predetermined number of bit positions,

Different characters are displayed on a display unit (60) depending on the value of the data words,

characterizes at least one predetermined value separating data words, which divides the input sequence into input sections,

a plurality of input sections are stored in a memory (44, 48),

at least once during the generation of the input sequence depending on at least one new data word generated since the last separating data word from the memory (44, 48), an input section belonging to the new data word is searched for which contains the new data word or is assigned to the new data word (step 212 ),

and in which a found input section or a part thereof is read from the memory (44, 48) and appended to the end of the input sequence (steps 236, 242),

characterized in that several input sections in the memory (44, 48) belong to the new data word,

and that of the input sections belonging to the new data word, the input section which was last accessed is selected.

2. The method according to claim 1, characterized in that in

Memory (44, 48) is only searched if immediately after the last generated separating data word, a data word is entered with a value that belongs to a predefined set of values that only comprises part of the values of the data words (step 156).

3. The method according to claim 2, characterized in that the

Set of values contains values to which capital letters are preferably assigned according to the ASCII code (step 156).

4. The method according to any one of the preceding claims, characterized in that the memory (44, 48) is only searched for when the number of data words generated after the last separating data word does not exceed a predetermined number, which is preferably two, three or four data words (Step 204).

5. The method according to any one of the preceding claims, characterized in that after appending the input section

(Steps 236, 242) it is checked whether the next data word generated has a predetermined value which indicates an acknowledgment of the appending (step 194),

and that if the value deviates from the predetermined value, the appended input section is removed (steps 226, 236, 226 ').

6. The method according to any one of the preceding claims, characterized in that the input units contain data words, the value of which results in a representation of letter characters on the display unit (60).

7. The method according to any one of the preceding claims, characterized in that after the input of an input section is checked whether this input section is already stored in the memory (44, 48) (step 172), and that the input section is written into the memory (44, 48) if the input section is not yet stored (step 174).

8. The method according to any one of the preceding claims, characterized in that when a predetermined number of input sections stored in the memory is reached, an input section is deleted, preferably an input section which has not been accessed for the longest time (step 178).

9. The method according to any one of the preceding claims, characterized in that it is used when generating an input sequence with repetitive input sections in a word processing program (46).

10. The method as claimed in one of the preceding claims, characterized in that the associated input sections or all input sections in the memory (44, 48) form an input section sequence,

that the position of an input section in the input section sequence is determined by the last access to the respective input section,

the order of the input sections in the input section sequence corresponds to the order of the last access to the respective input sections,

and that prior to reading an input section, the input section sequence is searched in a search direction in which, starting with a defined input section towards the end of the input section sequence, the input sections are searched for input sections which contain the new data words (step 212).

11. The method according to claim 10, characterized in that the search is terminated at the first search success and the input section found is the selected input section

(Step 214).

12. The method according to claim 10 or 11, characterized in that in the event of a search success, the input section found receives the position at the beginning of the input section sequence (step 184, 185).

13. The method according to any one of claims 10 to 12, characterized in that an input section to be stored again in the memory (44, 48) is stored at the beginning of the input section sequence (steps 168, 174).

14. The method according to any one of claims 10 to 13, characterized in that an input section to be removed from the memory (44, 48) is removed from the end of the input section sequence (steps 176, 178).

15. The method according to any one of claims 10 to 14, characterized in that the input sections remain in their memory cells of the memory (44, 48) when changing their position within the input section sequence,

and that the change in position is noted in a position management list, preferably in a table, which addresses memory cell addresses of the memory (44, 48) assigns to positions in the input section sequence.

16. Data processing system (40), in particular for carrying out the method according to one of the preceding claims,

with a data input unit (10) for generating an input sequence from data words with a predetermined number of bit positions, a display unit (60) for displaying characters as a function of the value of the data words generated, at least one value identifying separating data words which divides the input sequence into input sections,

a memory (44, 48) for storing a plurality of input sections from data words,

and with a control unit (42) which reads an input section or part thereof belonging to the new data word at least once during the generation of the input sequence depending on at least one new data word generated since the last separating data word, and automatically on End of the input sequence (steps 236, 242),

characterized in that the control unit (42) selects, in the case of a plurality of input sections belonging to the new data word, the input section or which was accessed last.

17. Method for accessing a storage unit (44, 48),

in which memory words with the same number or a different number of data words from a predetermined number of bit positions are stored in a write mode (steps 176, 185),

and in a read mode (steps 212) an associated memory word is output to an input word consisting of at least one data word, which memory word contains or is assigned to the input word, characterized in that, in the case of a plurality of associated memory words, a selection is made depending on the chronological order of access to the memory words.

18. The method according to claim 17, characterized in that the associated memory word is output, which was last accessed,

or that the associated memory word is output which has not been accessed for the longest time.

19. The method according to claim 17 or 18, characterized in that the associated memory words or all memory words in the memory (44, 48) form a memory word sequence,

that the position of a memory word in the memory word sequence is determined by the last access to the respective memory word,

and that before the readout of a memory word, the memory word sequence is searched in a search direction in which, starting with a fixed memory word towards the end of the memory word sequence, the memory words are searched for memory words which contain the input word (step 212).

20. The method according to claim 19, characterized in that the

Search is terminated at the first search success and the memory word found is the selected memory word (step 214).

21. The method according to claim 19 or 20, characterized in that in the event of a search success, the memory word found receives the position at the beginning of the memory word sequence (steps 184, 185).

22. The method according to any one of claims 19 to 21, characterized in that a memory word to be newly stored in the memory is stored at the beginning of the memory word sequence

(Steps 168, 174).

23. The method according to any one of claims 19 to 22, characterized in that a memory word to be removed from the memory is removed from the end of the memory word sequence (steps 176, 178).

24. The method according to any one of claims 19 to 23, characterized in that the memory words remain in their memory cells when changing their position within the memory word sequence,

and that the change in position in a position management list, preferably in a table, is noted which addresses of memory cells in the memory assign to positions in the memory word sequence.

25. Storage unit (44, 48) for storing data words and in particular for performing the method according to one of claims 17 to 24,

with memory cells for storing memory words consisting of at least one data word with a predetermined number of bit positions,

with an input unit for inputting an input word from at least one data word, and with an output unit for outputting a memory word associated with the input word which contains the input word or is assigned to the input word,

characterized in that, in the case of a plurality of associated memory words, one of these memory words is selected depending on the chronological order of reading and / or writing the memory cells.

26. Storage unit (44, 48) according to claim 25, characterized by a control unit for selecting the storage word depending on the chronological order of access to the associated storage words.

27. Memory unit according to claim 25 or 26, characterized in that the memory cells are contained in an associative memory.