EP0992947B1 - Apparatus and method for the storing of data concerning the usage of an end-user device - Google Patents

Description

The invention relates to an arrangement and a method for storing data on a use of a terminal according to the preamble of claim 1 and of claim 7. The arrangement and the method are particularly suitable for users of franking machines.

Postage meter machines at least requested the service of printing a postage value on a print carrier (adhesive tape, letter). Further known services include, for example, postage calculations for determining the postage value of a most favorable carrier or freight forwarder (EP 747 864 A2). On the one hand, each terminal thus provides the service for which it is programmed.

On the other hand, it is known to connect the terminals to a remote data center (DE 195 49 305 A1) or to a remote personal computer (DE 195 17 557 A1) in order to be able to use their services in interaction with the remote facility.

Already EP 493 948 B1 discloses a postage meter machine which can print the delivery type as a separate optional printing stamp for the franking stamp or integrated with it and which is equipped with a plurality of registers in a secure module for storing accounting data relating to the use of the franking machine relate to the franking of objects. A first set of registers refers to a specific first service and a second set of registers refers to a specific second service, whereby the special services are selectable via the input means and the accounting data of the respective selected service is updated. However, only certain services will be collected for billing purposes. These statements do not give the carrier sufficient information about customer behavior.

Now, some postal authorities / mail carriers are required or favored by discounts that the user makes printouts on postage meter / freight accompanying document / bill mail stored in the postage meter or in a period of time, ie billing, statistics or receipts upon reloading Credit increase. Predicting customer behavior would also be of interest to future postal carriers, who will offer new postal services, which will then have to be billed separately. According to EP 285 956 B1, a franking machine is equipped with a special operation memory and with a connection for an external printer. The user is required to select and print only certain of the stored periodic data collection.

For this, the user not only needs to provide a separate printer, but may also need to reserve a lot of time to search out and print the data.

The applicant's new franking machines use digital printing units. For example, the franking machines T1000 and JetMail by the applicant Francotyp Postalia AG & Co. have, for the first time, a thermal transfer printer or an inkjet printer worldwide. This makes it possible in principle to print on a filled letter in the area of the postage stamp also addresses and other information, which are in a corresponding context with a service. For example, optional prints are also printed on the franking in order to mark the shipping type or form.
The form of storage may be adapted to the needs of a plurality of users of the same postage meter. For example, some machines assume class formation in the form of cost centers that are assigned to individual user groups. For printing out corresponding reports even without a separate external printer, DE 42 24 955 A1 proposes a method and arrangement for internal cost center printing. The printouts that can be generated for each cost center contain subdivided listings of the post office use of the postage meter machine according to optional printouts. The saving of the additional printer is beneficial to the users. However, it is unreasonable for users to print out lists of data on the use of the postage meter machine at the request of the mail carrier or data center. Because of course, during the printing of lists of the cost center data by means of the franking machine-internal printhead can not be franked.

From FR 2 665 003 B1 a device for tracking the post office use of an electronic franking system is known, which communicates with a central office, which exchange electronic messages. Besides the bytes, the message format includes Information on identification, register status and credit status as well as a number of additional bytes for the statistical sequence. The bytes for the statistical sequence contain a breakdown by postage value, by weight or by destination of the franked postal items. Part of the statistical order instructions may be modified during communication with the control center. Modification of a statement must, however, be done in advance, before a statistical entry is started. Communication takes place via telephone line or via electronic module, which must be physically transported to the communication partner.

From EP 717 376 A2 a statistical program franking machine is known, which performs the transmission of statistical data to the data center by modem and the changing of parameters in the meter for the statistical data acquisition by downloading new parameters from the data center to the definition of the class boundaries. However, the possibilities for performing statistics are limited to postage classes.

Due to the large number of services, it is not always clear for which of the services the provider capacity would have to be expanded. In some cases, outdated telecommunications systems do not allow high data throughput. Now, in a non-prepublished German Patent Application 19731304.3-53, it has been proposed to generate selectable link instructions for class formation in the data center and to transmit them to the postage meter machine in the form of new statistical instruction data. The formation of a new statistics mode takes place in the franking machine on the basis of the new statistical instruction data and before the use of functions and services and their statistical recording in the memory of the postage meter machine. An advantage is the pre-compression of the data by the link and that a transmission of pre-compressed data to the data center does not interfere with the process, as their transmission takes little time.

The object of the invention is to make the storage of data in a terminal and its communication with a remote data center so that the manner of statistics can be subsequently defined. The terminal should not be blocked by keeping statistics. Unaffected by the storage and transmission of usage data, the use of the terminal for services is to be enabled in a service mode. Services for a franking device should also be statistically detectable, which are used in interaction with the data center.

The object is solved with the features of claims 1 and 7.

The invention is based on the availability of low-cost memories with a very high storage capacity in order to collect usage data of a system. The usage data relates to information about user-selected properties of a system or features of a device that are capable of rendering services at least statistically detectable. The mailing or the mailing of items by a public or private mail carrier is only one possible service type of a service company. Device of the shipping system is understood to be a terminal remote from the service provider or a terminal device of a franking system standing at the customer's. The characteristics of the shipping system or (customer / end) device are to be understood as specific user inputs and specific automatically made entries that are related to the order that a customer places with the service provider. Advantageously, on the one hand, the memory capacity of a user memory can be optimally utilized by a space-saving storage space management of usage data, ie, be set lower than in the case of a known pure historical storage of usage information. On the other hand, when the available storage capacity is fully utilized, the number of storages can be increased, ie the terminal or device of a franking system can have a longer storage life Be operated before communication with the data center is required.
Shortly before or during the memory overflow of the user memory, the stored data is loaded into a memory of the data center. Based on the improved possibilities of communication at a high baud rate, communication requires only a relatively short period of time to transfer all the data from the terminal to the data center. Upon successful data transfer from the user memory to the remote memory of the data center, the originally available memory capacity can be restored by reinitialization, optionally including deleting the user memory.
The terminal is in particular a franking machine, a device of a franking and / or mail processing system, a PC franking machine or any peripheral device of a system which itself provides a service or permits the use of the service of a third party. A record of individual usage events is created in a corresponding service mode. The usage information preferably identifies the postal use of the franking machine or of a franking system. The franking mode according to the invention comprises a storage mode for the cumulative storage of the current usage information together with the previous usage information. It is contemplated that a postage meter, while operating or as a result of operating in the postage mode, will store usage information sparingly on the one hand and in a form that reflects the historical order of events on the other hand. Advantageously, however, no finished or pre-compressed statistics must be stored in the terminal franking machine. According to the invention, in the remote data center, the generation of an on-demand statistics on a use of a terminal after loading those data from the user memory in the remote memory, which were stored non-volatile during use of the terminal in the user memory.

It is envisaged that a space-saving historical storage method will be used and data compression will be performed by modem during transmission. According to the invention, a storage of such usage data of an event is dispensed with, which is equal to the previously stored usage data of an earlier event. Equality exists in the way that existing differences can not be grasped. For example, certain data are only available in a connected scale and are not transmitted to the franking machine. Equality is also present if existing differences are not recorded in the parameter memory or a corresponding information is missing. It is possible that a device of a franking system remains unused or is not connected (for example, a separate static balance).

According to the invention, information which relates to the properties or features which have already become active or which are bound to become effective and which are capable of rendering services statistically detectable, is analyzed for qualitative and quantitative changes. An event of qualitative change of a property is historically stored in which new memory locations are occupied for such usage data. However, a quantitative change only leads to the correction of the data in the occupied memory locations with respect to the relevant data of the result of a previous analysis. The assignment to the other data remains.

The result of the analysis is stored in a historical order of records or in the form of historically listed event code in an event list. This results in a space-saving memory usage for a historical storage of usage data regardless of the form of storage, ie, regardless of whether the information is stored in the form of data in a record or alternatively, whether a storage in the form of event code in an event list he follows.

In this context, an event is a determination of a future change of state of postal data or shipping information in a device that is inevitably or inevitably occurring, which is caused by an input or setting made automatically or by the user. A feature which can change the state of a machine or system in connection with a service is, for example, a stored (special) setting of a franking machine or of another device of a franking system, which has an effect on the franking, because in the impression correspondingly a comparable characteristic feature appears or because at least in the billing of the service has a corresponding user input or because the user behavior affects. The stored (special) setting of a franking machine or a franking system is also referred to as a parameter. Such parameters are listed in a parameter list and remain non-volatile stored in a parameter memory until the next change. The change of a parameter preceding the franking, which relates to a qualitative property, is according to the invention an event to be documented in the use memory 16, which occupies a state change. With a large number of such parameters, the space-saving storage method is of particular importance. It has been found that for a given user behavior, it is particularly advantageous if the storage of events relating to qualitative and quantitative properties takes place as listed specific code in an event list. The microprocessor is inventively programmed to organize the allocation of memory locations for the data of a record in the current process according to the occurring events, wherein the data of the record include at least one parameter value and a number of pieces. An associated reference list stores a specific code for events relating to qualitative properties, wherein the code can be assigned to a property in a dynamic process of a property that can be freely defined within predetermined limits.

Figur 1a,

Blockschaltbild einer Frankiermaschine,

Figur 1b,

Variante mit separaten Benutzungsspeicher und mit OTP in der Steuereinrichtung der Frankiermaschine,

Figur 1c,

Blockschaltbild eines Meters einer Frankiermaschine,

Figur 2,

Gesamtablaufplan für eine Frankiermaschine.

Figur 3a und b,

Speicherformate in erster Variante mit einem Stückzähler und in der zweiten Variante mit mehreren Stückzählern,

Figur 3c-1 bis 3c-5,

Darstellung von Speicherzuständen nach einer dritten Variante mit separat verwaltet Stückzählern,

Figur 3d-1

bis 3d-5, Darstellung von Speicherzuständen nach einer vierten Variante mit nur einem Stückzähler,

Figur 4,

Flußplan zur Abspeicherung von Benutzungsdaten gemäß Fig.3c,

Figur 5,

Darstellung eines Speicherplatzes der Parameter-Liste,

Figur 6,

Darstellung eines Speicherplatzes der Bezugs-Liste,

Figur 7a,

Suchroutine für Flußplan nach Figur 7b,

Figur 7b,

Flußplan zur Abspeicherung gemäß Fig.3d,

Figur 7c,

Subroutine zum Flußplan gemäß Figur 7b,

Figur 8,

Erweiterung auf eine beliebige Anzahl der Ereignisse,

Figur 9,

Darstellung einer Routine mit Speicher-Neuinitialisierung.

The inventive space-saving historical storage method differs from a known per se statistics with class formation in that storage in classes with respect to qualitatively different use is eliminated and thus the historical order of storage of user data can be subsequently retrieved in principle again. In contrast to a historical memory method known per se, a memory-space-saving method is used, which also leads to a shortening of transmission times in a data transmission. The manner of the final statistics need not already be fixed beforehand, but is defined if necessary at the time of the query or later in the generation of on demand statistics in the data center. The transmitted data can thus be used in an advantageous manner subsequently, possibly even at greater intervals after a query, to generate a respectively desired statistics, if necessary. The on-demand statistics have the advantage that a complete conversion of the statistics for a later evaluation is possible.
Advantageous developments of the invention are characterized in the subclaims or are presented in more detail below together with the description of the preferred embodiment of the invention with reference to FIGS. Show it:

FIG. 1a,

Block diagram of a franking machine,

FIG. 1b,

Variant with separate user memories and with OTP in the control device of the franking machine,

FIG. 1c,

Block diagram of a meter of a franking machine,

FIG. 2,

Overall schedule for a franking machine.

FIGS. 3a and b,

Storage formats in the first variant with a piece counter and in the second variant with several pieces counters,

3c-1 to 3c-5,

Representation of memory states according to a third variant with separately managed parts counters,

Figure 3d-1

to 3d-5, display of memory states according to a fourth variant with only one piece counter,

FIG. 4,

Flow chart for storing usage data according to FIG. 3c,

FIG. 5,

Representation of a memory location of the parameter list,

FIG. 6,

Representation of a memory location of the reference list,

FIG. 7a,

Search routine for flow chart according to FIG. 7b,

FIG. 7b,

Flow chart for storage according to FIG.

FIG. 7c,

Subroutine to the flowchart according to FIG. 7b,

FIG. 8,

Extension to any number of events,

FIG. 9,

Representation of a routine with memory re-initialization.

FIG. 1a shows a block diagram of the franking machine according to the invention with a printer module 1 for a fully electronically generated franking image, with at least one input element 2 having a plurality of actuators, a display unit 3, a MODEM 23 producing the communication with a data center, further input means 21 or balance 22 which are coupled to a controller 6 via an I / O control module 4 and to nonvolatile memories 5a, 5b and 9, 10 and 11, respectively, for data and programs, respectively, which include the variable and constant portions of the franking image.

In the German patent application DE 19534530 A1 closer explanations are made to individual functions of the means. A character memory 9 supplies the necessary pressure data for the variable parts of the franking image to a volatile working memory 7. The control device 6 has a microprocessor μP connected to the input / output control module 4, with the character memory 9, with the volatile random access memory 7 and nonvolatile random access memories 5a, 5b (with internal user memory) or (dashed line) with an additional user memory 16, with a program memory 11, with the motor of a transport or feed device, if necessary with strip release 12, an encoder (coding disk) 13 and with a calendar or clock / date module 8 is in communication. The memory module which comprises the non-volatile main memory 5b can be, for example, an EEPROM, which is secured against removal by at least one additional measure, for example sticking to the printed circuit board, sealing or casting with epoxy resin. In the non-volatile memory 5a, the current parameters are stored in a designated area, which are printed at least the postage as variable data on the mail piece. This memory area can also be executed as a separate parameter memory. The corresponding parameters listed in a parameter list and remain non-volatile stored in the parameter memory until the next change. The usage memory can also be realized separately or, for example, within the nonvolatile memory 5a by providing special memory areas. In a particular service provided by another service provider, separate storage areas, not shown, are provided for usage data or separate storage devices. The individual memory can be realized in several physically separate or not shown combined in a few modules. The required usage information is stored in at least one separate memory area including a plurality of memory locations 16-01 through 16-0n. In the embodiment the memory module 16 is assigned to the service provider and its memory area is assigned to all cost centers. In the case of a multiplicity of memory areas, each memory area can each be assigned to one cost center. Per memory space, a number of memory cells corresponding to the number of information units to be stored bits are addressed by the microprocessor uP. The following pieces of information bits, such as number, postage, date, promotional cliché number, weight, format, shipping information (shape, type, destination) and certain errors can be mapped to the aforementioned information bits.
The available storage capacity in the usage memory is for example 20 kbytes. It is optimally used by a space-saving storage space management, which nevertheless allows a reconstruction of the order of usage information for a service. The use of the franking machine in accordance with the franking service for the dispatch of mail pieces results in at least one specific user information. The usage information required for the subsequent creation of any statistics in the data center is transmitted via modem 23 to the data center.
A storage area is selected by a cost center entered by the user in order to store the postage value in a first storage location 16-01 as the first usage information in a first use "franking". Further memory locations for values or numbers are provided as user information. Usage information is assigned a counter which is incremented if the subsequent use gives similar usage information. However, if the subsequent use results in uneven use, a correspondingly different usage information is stored in one of the following memory locations. In the second memory location 16-02, at least one allocated number of memory cells exist as counters for storing a subsequent similar use. The counter requires only a small number of memory cells in the second memory location 16-02. The storage space saving thus results from the piece counting in an immediately consecutive similar use the franking machine. In one variant, a group of usage information is assigned a single counter. This is advantageous in the case of a stack of mail pieces, whereby each mail piece can be assigned this same group. The individual usage information of the group remains unchanged. In another variant, the usage information advantageously consists of at least one code or an identifier, which likewise occupies only a small number of memory cells of the memory location.

FIG. 1b shows details of the block diagram of another variant of the electronic postage meter with one-time programmable (OTP) processor in the control device. In the embodiment of a franking machine proposed in EP 716 398 A2, a suitable OTP processor with nonvolatile memory for billing and with an ASIC as an interface to the base is connected. Details of this embodiment have been disclosed in German patent application DE 19534530 A1 entitled: Method for securing data and program code of an electronic franking machine, as well as in more detail in the German patent application DE 19731304.3-53 entitled: Statistics Mode Reload and Statistics Statistical Classification the storage of a record, explained.
The block diagram shown in FIG. 1b also applies in principle to any other electronic control unit to which a printer (not shown) and a modem (not shown) can be connected. According to the invention, it is provided that the electronic control unit provides at least one separate memory area for usage data, which can be transmitted at intervals to the data center via modem. Alternatively, an electronic postage meter machine in a variant not shown here can also be realized by a personal computer with a connected commercial printer, if the throughput of mail in mail processing may be low. Such an arrangement is also referred to as a PC meter. The printer can also be used for high-speed mail processing be interpreted. A mail processing system with a controlled by personal computer printing machine base station is described in the non-prepublished German patent application DE 19711998.0.

FIG. 1c shows a block diagram of the meter of an electronic postage meter which can be detached from a base which is also suitable for mixed mail processing. Part of the meter is a special security module 40, which carries out the billing of the frankings and stores. The security module 40 is protected by physical and software security measures.
For non-volatile storage of usage data, a read / write memory 16 is connected to a microprocessor 6 of the meter. This memory has memory areas with a larger number of memory cells for data records. The microprocessor 6 is either coupled in a conventional manner with a volatile random access memory RAM 7, which stores the working variables, pixel data and forms the stack area for the various tasks, or equipped with an internal RAM. The microprocessor 6 is connected to a program memory ROM 11 or equipped with an internal Rome, with which the microprocessor 6 is programmed accordingly, so that according to the use of records in the non-volatile random access memory 16 arise. The non-volatile random access memory 16 is, for example, an NV-CMOS RAM or an E ² PROM. The microprocessor 6 is correspondingly programmed to manage at least one of the memory areas, wherein a certain number of similar data sets can be stored in the aforementioned memory area. On the microprocessor 6, a cliché memory 10, a battery-backed clock / date module 8, a postage rate storage 13 and input means and output means are also connected via a bus 44. An input means is the keyboard 2. Another - not shown input means - may be a modem or a smart card write / read unit and optionally also acts as an output means. A Display unit 4 is, for example, an LCD display with associated controller and a further output means is a printer, in particular in a base with a machine-internal print head 1, which is not moved during printing and with which at least one franking stamp imprint can be printed on a piece of mail. Billing in the security module 40 can take place, as is explained in more detail in the European application EP 789 333 A2 (entitled: Franking Machine). The hardware-based billing is fast and tamper-proof, so that no billing errors can arise. The security module 40 includes a One Time Programmable (OTP) processor 50, a power-on reset device 48, and a user-defined device 66 (ASIC) having a computing unit 60 and a connected CMOS-based lithium battery 42. RAM memory 41 on. The billable data are transmitted via a bus 44 in the base to the interface 64 of the user-specific module 66 and from there to the Abrecheneinheit 60, which performs the billing hardware. The print data are transmitted from the microprocessor 6 either via an interface 64 and printer monitoring unit 62 of the user-specific module 66 to the print register DR 15 of the print controller DS 14 of the printer or output directly via an I / O unit 4. The security module 40 has, in a manner known per se, a secured housing and at least one additional physical security measure which prevents a successful entry into the security module and its exploration. Under the title: Arrangement for a security module, 198 16 572.2 physical security measures were described in more detail in the non-prepublished German patent application. Under the title: Arrangement for access protection for security module, 198 16 571.4 further physical security measures were described in more detail in the non-prepublished German patent application. The security module prevents the misuse of programs or data stored in the meter, ie externally from the security module. suitable Security measures are also the European patent applications EP 660 269 A2 (method for improving the safety of franking machines) and EP 762 227 A2 (method and arrangement for increasing the security against manipulation of critical data) and EP 762 338 A2 (method for securing data and program code a electronic postage meter) removable.
The microprocessor 6 is preferably arranged in meters of the franking machine and communicates with a franking machine-internal interface to the base in particular with an actuator / sensor controller 17 and with an encoder 13 for determining the transport speed of the mail in combination, as already in principle in the EP 716 398 A2 was proposed. One of the actuators is a letter sensor, which determines the reaching of the printing position of a letter or other mail, with letter thicknesses up to 20 mm are possible. At higher thicknesses of the mail can be worked with a - not shown - strip encoder for franking strips, which is also connected via the actuator / sensor controller 17 to the microprocessor 6. In addition, the I / O unit 4 is provided with corresponding interfaces for modem and / or chip card read / write unit and with system interfaces for coupling further different external devices, such as an external balance, an automatic mail item feeder, a mail piece or a personal computer PC.

FIG. 2 shows an overall flow chart for a franking machine with steps 417 to 430 according to the invention for a storage mode within a franking mode 400. The flow has, after a start routine 101, a system routine 200 with a point s. It is envisaged that after switching on the franking machine in step Start 100 within a start routine 101, a functional check with subsequent initialization takes place. A program code in the non-readable internal OTP-ROM now allows several advantageous start-up security check routines, as described in the German patent application DE 19534530 A1 with the title: Method for securing data and program code electronic postage meter were set forth in more detail. The terminal is preferably a postage meter having a microprocessor 6 programmed to enter into an input / display routine 209 and a metering mode (400) within a system routine 200, the input routine 209 being for acquiring current usage data by the microprocessor the franking machine in the non-volatile memory 5a, prompting steps for inputting a usage information concerning at least the franking value, the weight or the shipping. The process continues after the input / display routine 209 with a branch to a communication mode 300 and to the modified franking mode 400. In the modified franking mode 400, the steps 401 to 407 are preferably executed in a known manner as explained in the foregoing. One difference, however, is that the OTP processor 50 of the security module 40 handles these steps, with the step 406 performed by the hardware abort unit 60 of the ASIC 66. The billing data is stored in the Li-battery based NV-RAM 41 in the security module. Following are a number of steps (not shown) for forming 415 a checksum and decrypting DES 416 into a MAC (Message Authentication Code).

• zum Vergleich aktueller Benutzungsdaten mit den kumulativ abgespeicherten Benutzungsdaten,
• Zur Abspeicherung eines neuen Datensatzes mit Benutzungsdaten im Benutzungsspeicher in historischer Reihenfolge entsprechend der Benutzung, wenn Benutzungsdaten gegenüber den im vorangehend abgespeicherten Datensatz befindlichen Benutzungsdaten verändert oder ungleichartig sind,
• zur Bildung eines Zählers für Benutzungsdaten,
• zur Inkrementierung des Zählers und Ersetzen entsprechender Daten eines bereits abgespeicherten Datensatzes durch neue Daten, welche den inkrementierten Zählerstand wiederspiegeln, wenn die aktuellen Benutzungsdaten gegenüber dem bereits abgespeicherten Datensatz unverändert oder gleich sind. Für die nichtflüchtige Speicherung der aktuellen Benutzungsdaten ist ein Parameterspeicher 5a vorgesehen. Im Benutzungsspeicher 16 werden die aktuellen und vorherigen Benutzungsdaten kumulativ nichtflüchtig gespeichert. Im Unterschied zur Behandlung der Abrechnungsdaten dürfen die Be-nutzungsdaten nach deren Übermittlung an die Datenzentrale vom OTP-Prozessor 50 des Sicherheitsmoduls 40 oder vom Mikroprozessor 6 des Meters überschrieben oder gelöscht werden. Es ist vorgesehen, daß der Mikroprozessor programmiert ist, zur Wiederherstellung der ursprünglichen Speicherkapazität im Benutzungsspeicher 16 durch Überschreiben oder Löschen des Speicherinhaltes im Rahmen seiner Neuinitialisierung.
  Die Figur 9 zeigt eine Routine mit Neuinitialisierung des Benutzungsspeichers 16. Nach dem Kommunikationsmodus 300 (Fig.2) wird der Punkt b und damit der Anfang der o.g. Routine erreicht. Im Abfrageschritt 211 wird abgefragt, ob bei der Kommunikation Benutzungsdaten zur Datenzentrale übermittelt wurden. Es ist vorgesehen, daß nach der Übertragung und Speicherung von Benutzungsdaten im entfernten Speicher im Abfrageschritt 211 einer Systemroutine 200 des Endgerätes festgestellt wird, daß die Daten übermittelt worden sind, daß im Schritt 213 der Benutzungsspeicher 16 Neuinitialisiert wird und anschließend im Schritt 215 ein Anzeigetext generiert wird, bevor zum Punkt s (Fig.2) der Systemroutine 200 zurückverzweigt wird. Vom Abfrageschritt 211 kann alternativ auf einen Schritt 214 zur Auswertung der Kommunikation verzweigt werden. Zwischen den Punkten b und d des in der Fig.2 gezeigten Flußplanes können weitere Abfragen liegen, bevor ein Dienstleistungsmodus erreicht wird, der den erfindungsgemäßen Speichermodus einschließt. Ist ein Benutzungsspeicher zu voll kann vom Dienstleistungsmodus auch direkt zum Punkt g verzeigt werden, um automatisch in einen Kommunikationsmodus II einzutreten. Es ist vorgesehen, daß kurz vor oder beim Speicherüberlauf die im Benutzungsspeicher des Endgerätes gespeicherten Benutzungsdaten an die Datenzentrale übermittelt werden. In Verbindung mit Fig.7c wird das weiter unten noch näher erläutert, wie ein voller Speicher erkannt wird. Im Rahmen der Neuinitialisierung 213 wird der Listenendecode auf den Speicherplatz N = 0 gesetzt ( Fig.9 ).
  Die Maschine kann feststellen, daß die Laufvariable n = N = 0 gesetzt ist. Im Rahmen der Neuinitialisierung kann auch ein Löschen des Speicherinhaltes oberhalb des Listenendecodes erfolgen. Der Speicherbereich wird dadurch wiederbeschreibbar bzw. frei für neue Benutzungsdaten und erreicht den - in der Figur 3d-1 gezeigten - ursprünglichen Status.
  Alternativ kann die Routine nach Fig.9 ggf. inclusiv mit dem Löschen von Speicherbereichen als Bestandteil des Kommunikationsmodus 300 bzw. 350 erfolgen - die in der oben genannten deutschen Patentanmeldung DE 19731304 A1 erläutert worden sind.
  Gemäß Figur 2 laufen im Speichermodus folgende Schritte ab: Ein Parametervergleich im Schritt 417 ergibt beispielsweise, daß ein neuer Parameterwert in die Frankiermaschine eingegeben wurde. Im Abfrageschritt 418 wird abgefragt, ob ein neuer Portowert in die Frankiermaschine eingegeben wurde, um zum Schritt 424 zu verzweigen. Im Schritt 424 wird eine neue Zeile in eine Liste eingetragen, wenn der neue Parameterwert der Portowert ist. Danach wird zum nächsten Abfrageschritt 419 verzweigt, denn es könnte zusätzlich auch ein weiterer Parameterwert verändert worden bzw. erstmalig eingegeben worden sein. Anderenfalls wird zum zweiten Abfrageschritt 419 verzweigt, wenn die erste Abfrage im Abfrageschritt 418 ergibt, daß der Portowert nicht verändert worden und nicht erstmalig eingegeben worden ist. Im Abfrageschritt 419 wird abgefragt, ob ein neuer Gewichtswert in die Frankiermaschine eingegeben wurde, um zum Schritt 425 zu verzweigen. Dabei wird zusätzlich zum Abfrageschritt 419 eine weitere Abfrage durchlaufen. Die zusätzliche Verzweigungsbedingung zum Schritt 425 ist, daß zuvor noch keine neue Zeile in dieselbe Liste eingetragen wurde. Andernfalls, falls im Schritt 424 eine neue Zeile in dieselbe Liste eingetragen wurde, dann wird im Schritt 422 in die bereits gelistete neue Zeile der neue Gewichtswert eingeschrieben, wobei. der Gewichtswert einen Gewichtsbereich repräsentiert. Im Schritt 425 wird eine neue Zeile in dieselbe Liste eingetragen, wenn dies noch nicht geschehen ist. Bezugnehmend auf die Darstellung in der Figur 3a wird die Information W1 aus der Zeile J = 1 überarbeitet, wenn eine neue Zeile in derselben Runde bereits in die Liste eingetragen worden ist. Danach wird über die Abfrageschritte 418, 419 zum dritten Abfrageschritt 420 verzweigt. Wenn die Abfragen ergeben, daß die Werte nicht verändert worden und nicht erstmalig eingegeben worden sind, weil eine neue Zeile in derselben Liste bereits gelistet ist, dann wird der dritte Abfrageschritt 420 erreicht. Im dritten Abfrageschritt 420 wird abgefragt, ob ein neues Datum in die Frankiermaschine eingegeben wurde, um zum Schritt 426 zu verzweigen. Dabei wird zusätzlich zum Abfrageschritt 420 eine weitere Abfrage durchlaufen. Die zusätzliche Verzweigungsbedingung zum Schritt 426 ist, daß zuvor noch keine neue Zeile in dieselbe Liste eingetragen wurde. Andernfalls wird in die bereits gelistete neue Zeile das neue Datum eingetragen. Im Schritt 426 wird eine neue Zeile in eine Liste eingetragen, wenn der neue Parameterwert das neue Datum ist. Danach wird zum nächsten Abfrageschritt 423 verzweigt. Anderenfalls wird gleich zum nächsten Abfrageschritt 423 verzweigt, wenn die dritte Abfrage im Abfrageschritt 420 ergibt, daß das Datum nicht verändert worden und nicht erstmalig eingegeben worden ist. Wird einerseits nun die Änderung eines Parameters festgestellt, dann wird zum Schritt 427 verzweigt, um eine neue Zeile in die Liste einzuspeichern, falls dies in derselben Runde noch nicht geschehen ist. Andererseits kann eine Vielzahl an nächsten Abfrageschritten durchlaufen werden, ohne daß eine Änderung eines Parameters festgestellt wird. In einem solchen Fall wird zum Schritt 428 verzweigt, um einen Stückzähler Z:= Z + 1 zu inkrementieren. Bezugnehmend auf die Darstellung in der Figur 3a wird der Zählerstand im Bereich N1 im Datensatz der Zeile J = 1 entsprechend korrigiert, ohne daß eine neue Zeile gespeichert werden muß, wenn die Benutzung gleichartig erfolgt bzw. wiederholt wird. In einem nachfolgenden Schritt 429 wird der noch verfügbare Speicherplatz im Speicherbereich des Benutzungsspeichers 16 überprüft. Ist die verfügbare Speicherkapazität für eine Speicherung eine nächsten ungleichartigen Benutzung noch ausreichend, dann wird zum Schritt 431 mit der an sich bekannten Frankierdruckroutine verzweigt. Andererseits, wenn die Liste fast voll ist, wird zum Schritt 430 verzweigt, um eine Warnung zu generieren und das Zeitintervall der Speicherung von Benutzungsinformationen zu schließen, bevor zum Schritt 431 mit der Frankierdruckroutine verzweigt wird. Die Frankierdruckroutine wird in an sich bekannter Weise vorzugsweise vom Mikroprozessor 6 durchgeführt. Da der Mikroprozessor in der Systemroutine die Druckanforderung 405 ständig und schneller abfragt, als ein zu frankierendes Poststück nachgeliefert werden kann, ist das Durchlaufen des Abfrageschrittes 349 gesichert. Somit kann auf indirekte Weise eine Kommunikation mit Übermittlung der Daten zur Datenzentrale und einem anschließenden Speicherbereichslöschen ausgelöst werden.
  Die Figur 3a zeigt ein Speicherformat in einer ersten Variante. Der Mikroprozesssor bildet mittels einem Speicher einen Stückzähler Z. Nur bei einer erstmaligen Speicherung von einer der Benutzungsdaten wird Z auf den Wert 1 gesetzt und die Abspeicherung des Datensatzes von neuen bzw. veränderten und unveränderten Benutzungsdaten erfolgt in einer neuen Zeile J = 2. Bei einer wiederholten Speicherung von unveränderten Benutzungsdaten genügt es, wenn im Bereich N1 der ersten Zeile J = 1 der inkrementierte Zählerstand eingetragen wird, weil eine offensichtlich gleichartige Benutzung des Endgerätes vorliegt.
  Jede Zeile J speichert Daten für folgende Informationen, wie Anzahl Nj, Portowert Pj, Kalenderstand Cj (Datum), Werbe-Klischee-Nummer Aj, Gewichtswert Wj, Format Fj, Versandinformationen Dj (Form, Art, Ziel) und ggf. bestimmte erfaßte Fehler Ej.
  Die Speicherzeilen J = 1, 2,..., 5, ..., sind im Speicherbereich in einer historischen Reihenfolge gelistet. Einem Poststückstapel an gleichartig frankierten Postenstücken kann eine bestimmte Speicherzeile zugeordnet werden, wenn die Postenstücke des Stapels hintereinander frankiert wurden, so daß ungleichartig frankierte Postenstücke nicht vorkommen. Einem ersten Stapel gleichartig behandelter Postenstücke entspricht dann eine erste Zeile J = 1, einem zweiten Stapel gleichartig behandelter Postenstücke entspricht dann eine zweite Zeile J = 2, ..., einem fünften Stapel gleichartig behandelter Postenstücke entspricht dann eine fünfte Zeile J = 5. In jeder der Speicherzeilen J sind dann Informationen für eine gleichartige Benutzung gespeichert. Nur bei ungleichartiger Benutzung, d.h. bei Änderung mindestens einer der Informationen muß eine neue Zeile in den Benutzungsspeicher 16 mit entsprechenden neuen und den übrigen vorangegangenen Benutzungsdaten eingeschrieben werden.
  Die Figuren 3a und 3b zeigen zwei verschiedene Speicherformate für eine historische Speicherung von gleichartigen oder ungleichartigen Benutzungsinformationen einer aufeinanderfolgenden Benutzung entsprechend einer Dienstleistungsart. Während in der Variante gemäß Figur 3a die Speicherplätze in einer Speicherzeile so neben- bzw. nacheinander in einer Reihe angeordnet sind, so daß sich der Wert für eine spezifische Information aus der Platzierung jedes der Speicherplätze in der Reihe ablesen läßt, sind in der Variante gemäß Figur 3b die Speicherplätze in einer Gruppe angeordnet, welche eine Kennung aufweist, wobei die Kennung die Platzierung in einer Reihe ersetzt. Die Variante nach Figur 3b erlaubt eine Platzierung einer Dreiergruppe innerhalb einer Reihe von Speicherplätzen bzw. innerhalb einer Zeile, welche vorteilhaft nun beliebig erfolgen kann. Der Mikroprozessor ist programmiert, nach der aktuellen Kennung eines Parameters zu suchen, welche eine Information über die Reihenfolge einschließt. Er liest dann eine interessierende Dreiergruppe aus, die im Ergebnis eines letzten Abspeichervorganges in Speicherplätzen abgelegt wurde. Eine Dreiergruppe hat im erläuterten Beispiel drei Speicherplätze. Das schließt aber nicht aus, daß eine Gruppierung eine Vielzahl an Speicherplätzen umfaßt.
  Eine erste Dreiergruppe hat die Bits B₁₁, die Bits B₁₂ und die Bits B₁₃. Die Bits B₁₁ des ersten Speicherplatzes 16-01 betreffen den Wert eines Parameters. Die Bits B₁₂ des zweiten Speicherplatzes 16-02 betreffen die Anzahl an Poststücken mit dem gleichen Wert des Parameters. Die Bits B₁₃ des dritten Speicherplatzes 16-03 betreffen Kennung für die Art des Parameters. Es ist vorgesehen, daß der Teil der Kennung, der etwas über die zeitliche Abfolge der Abspeicherung von Dreiergruppen aussagt, durch einen inkrementierbaren Zählerstand gebildet wird.
  Die Bits B₁₁ des ersten Speicherplatzes 16-01 der ersten Dreiergruppe betreffen einen ersten Wert des Parameters von der Art Portowert. Die Bits B₂₁ des entsprechenden ersten Speicherplatzes einer dritten Dreiergruppe betreffen einen zweiten Wert des Parameters von der Art Portowert, wobei der zweite Wert des Parameters historisch später eingestellt wurde. Zwischen der ersten und dritten Dreiergruppe liegt mindestens eine weitere Dreiergruppe. Im für die Figur 3b gewählten Beispiel existiert eine zweite Dreiergruppe mit Informationen betreffend eine Kenn-Nummer des Werbeklischees, welches zusätzlich zur Frankierung auf ein Poststück des ersten Stapels aufgedruckt wurde. Die Kennung umfaßt wieder einen Teil für die Art des Parameters und einen Teil für die zeitliche Abfolge der Abspeicherung der zweiten Dreiergruppe.
  Der Benutzungsspeicher 16 weist je Dreiergruppe einen jeweils dritten Speicherplatz 16-03 auf, in welchem zugeordnet zur jeweilig gewünschten Benutzungsinformation eine Kennung gespeichert vorliegt. Bei einer Abfrage des Speichers sucht der Mikroprozessor nach der aktuellen Kennung eines Parameters und liest dann eine interessierende Dreiergruppe aus, welche der nächstfolgenden Kennung eines Parameters vorausgeht. Der Wert oder die Nummer einer Benutzungsinformation wird im entsprechend dafür eingerichteten vorgesehenen jeweils ersten Speicherplatz 16-01 bzw. die Stückzahl gleichartiger Benutzungsinformationen jeweils im dafür vorgesehenen jeweils zweiten Speicherplatz 16-02 der Dreiergruppe gespeichert. Gegenüber der in der Figur 3a gezeigten Variante müssen allerdings mehrere Stückzähler separat verwaltet werden.
  Die Figur 3 c verdeutlicht ein Speicherformat am Beispiel einer listenartigen Darstellung von Speicherzuständen (Figuren 3c-1 bis 3c-5) im Verlauf einer historischen Speicherung von ungleichartigen oder gleichen Benutzungsinformationen bei einer aufeinanderfolgenden Benutzung entsprechend einer Dienstleistungsart. Die vorgenannte Dreiergruppe an Speicherplätzen wird nachfolgend als Liste bezeichnet.
  Die Anzahl der Listen ergibt sich aus der Anzahl der Benutzungsinformationen. Letztere sind im Ausführungsbeispiel nur der Einfachheit halber der Portowert, die Gewichtsstufe und das Datum. Tatsächlich können Listen für eine viel größere Anzahl an Benutzungsarten gespeichert werden. Jede Liste wird bei jeder Frankierung vom Mikroprozessor im innerhalb des Frankiermodus 400 abzuarbeitenden Speichermodus (Schritte 417 bis 430) bearbeitet.
  Am Anfang ist jede Liste noch leer. Ein Parametervergleich im Schritt 417 ergibt beispielsweise, daß ein Brief mit einem Portowert = 1,10 DM frankiert werden soll, daß sich das Gewicht des Briefes in eine erste Gewichtsstufe GW1 einordnet und daß der Kalenderbaustein 8 das Datum 31.08.98 ausweist. Der Mikroprozessor erzeugt im Speichermodus (Schritte 417 bis 430) den in der Figur 3c-1 gezeigten Zustand im Benutzungsspeicher 16.
  Anschließend erfolgt eine gleichartige Benutzung der Frankiermaschine bei der Frankierung des nächsten Briefes. Der Mikroprozessor erhöht in diesem Fall lediglich den Stückzähler in den drei Listen und erzeugt so im Speichermodus (Schritte 417 bis 430) den in der Figur 3c-2 gezeigten Zustand im Benutzungsspeicher 16.
  Nun werden weitere acht Briefe der gleichen Gewichtsstufe GW1, mit einem gleichen Portowert von 1,10 DM mit dem gleichen Datum 31.08.98 frankiert. Damit ergibt sich der in Figur 3c-3 gezeigte Zustand im Benutzungsspeicher 16. Es wurde lediglich der Stückzähler auf Z = 10 erhöht. Der erste Stapel umfaßte somit zehn gleichartige Briefe.
  Anschließend werden bei einer ungleichartigen Benutzung zwei Briefe der zweiten Gewichtsstufe GW2, mit einem zweiten Portowert von 3,00 DM, jedoch mit dem gleichen Datum 31.08.98 frankiert. Damit ergibt sich der in Figur 3c-4 gezeigte Zustand im Benutzungsspeicher 16. In der Datums-Liste wurde lediglich der Stückzähler auf Z = 12 erhöht. In den beiden anderen Listen für Portowert und Gewichtsstufe wurde jeweils eine neue Zeile eingetragen. Aus der Liste ist ersichtlich: Der frankierte erste Stapel umfaßte somit zehn gleichartige Briefe und der historisch nachfolgende frankierte zweite Stapel umfaßte somit zwei gleichartige Briefe.
  Am Folgetag am 01.09.98 wird noch ein Brief der ersten Gewichtsstufe GW1, mit einem ersten Portowert von 1,10 DM frankiert. Damit ergibt sich der in Figur 3c-5 gezeigte Zustand im Benutzungsspeicher 16. Da der Stückzähler in der ersten Zeile der Datums-Liste auf Z = 12 steht, bezieht sich das Datum auf die Einträge in den ersten beiden Zeilen der Listen für Portowert und der Gewichtsstufe, aber nicht auf den Eintrag in der letzten Zeile. Andererseits ist klar, daß ein weiterer Brief des ersten Stapels erst am Folgetag frankiert wurde, denn in der Datums-Liste wurde eine neue Zeile eingetragen, die der jeweils zuletzt eingetragenen Zeile in den beiden Listen für Portowert und der Gewichtsstufe zuzuordnen ist.
  Eine solche Daten-Speicherung ist platzsparend und erlaubt dennoch später nach Abfrage der Daten von der Datenzentrale eine beliebige Auswertung mit entsprechender Zuordnung der Daten zueinander.
  Die Figur 3 d verdeutlicht ein weiteres vorteilhaftes Speicherformat am Beispiel einer listenartigen Darstellung der Speicherzustände (Figuren 3d-1 bis 3d-5). Der Unterschied zum Format nach der ersten Variante gemäß Figur 3a besteht in der Vermeidung einer Protokollierung von gleichen Ereignissen in einer neuen Zeile. Es werden vielmehr nur die qualitativ veränderten bzw. ungleichartigen Ereignissen in der Reihen-folge ihres Auftretens protokolliert. Das erfolgt in Form einer Ereignis-Liste 500 mit zugeordneter Bezugs-Liste 610. In der Ereignis-Liste 500 werden Code eingetragen, deren Bezug zu qualitativ veränderten bzw. ungleichartigen Ereignissen aus der zugeordneten Bezugs-Liste 610 hervorgeht. Aus der Ereignis-Liste 500 ist der Verlauf einer historischen Speicherung von qualitativ veränderten bzw. ungleichartigen Ereignissen (Benutzungsinformationen) ersichtlich. Die Bezugsliste 610 speichert Referenzeinträge für die Code für qualitativ veränderten bzw. ungleichartigen Benutzungsinformationen.
  Die Referenzeinträge benötigen weniger Speicherplatz als die Einträge der Ereignisse, weil derselbe Referenzeintrag nicht wiederholt eingetragen wird, wenn sich das gleiche Benutzungsverhalten wiederholt und dabei gleiche Dienstleistungen in Anspruch genommen werden.
  Es ist bei einer bevorzugten Ausführung vorgesehen, daß der nichtflüchtige Speicher 5a ein Parameterspeicher ist, in welchem durch den Mikroprozessor 6 der Frankiermaschine bei der Erfassung von aktuellen Benutzungsdaten die Eigenschafts-Art und Wert der aktuellen Benutzungsinformation als Daten erfaßt werden. Der Benutzungsspeicher 16 der Anordnung weist Bereiche für die Speicherung einer Ereignis-Liste 500 und einer Bezugs-Liste 610 auf. Ein Programmspeicher 11 ist mit dem Mikroprozessor 6 verbunden und enthält ein Programm für den Speichermodus 417b - 430b, wodurch der Mikroprozessor 6 programmiert ist,
• daß Code in der Bezugs-Liste 610 und der Ereignis-Liste 500 gespeichert werden, wobei jedem Code Daten zur Beschreibung der Art und des Wertes einer Eigenschaft zugeordnet in der Bezugs-Liste 610 gespeichert werden,
• daß einer der Code zum Vergleich der aktuellen Benutzungsdaten mit den kumulativ abgespeicherten Benutzungsdaten der Ereignis-Liste 500 entnommen wird,
• daß die zugeordneten Daten betreff der Art aufgefunden und mit den dem Parameterspeicher 5a entnommenen jeweilig aktuellen Daten zu Eigenschaftsarten aktueller Benutzungsdaten verglichen werden, wobei wenn die aufgerufenen Daten zur Art ungleich sind, ein jeweils weiterer Code der Ereignis-Liste 500 solange entnommen wird, bis alle diejenigen Code der Ereignis-Liste 500 abgefragt worden sind, die sich auf qualitative Eigenschaften beziehen und ins Verhältnis mit jeweilig aktuellen Daten zu Eigenschaftsarten aktueller Benutzungs-daten gesetzt werden, wobei der Mikroprozessor 6 bei Ungleichheit der Daten zur Art einen neuen Code generiert und in die Bezugs-Liste 610 und die Ereignis-Liste 500 einschreibt, wobei auch die Daten zur Art und zum Wert der neuen Eigenschaft dem Code zugeodnet in der Bezugs-Liste 610 gespeichert werden, sowie wobei bei Gleichheit der Art der Mikroprozessor 6 noch den Wert anhand der in der Bezugs-Liste (610) gespeicherten Daten mit den auf den Wert der aktuellen Eigenschaft bezogenen Daten auf Gleichheit überprüft und bei Ungleichheit zum Wert der neuen Eigenschaft einen neuen Code generiert, die Daten zur Art und zum Wert der neuen Eigenschaft dem neuen Code zugeodnet in der Bezugs-Liste 610 speichert und den neuen Code sowie einen zugehörigen Ein-Stück-Code in die Ereignis-Liste 500 aufnimmt, welche mit einem Listen-Ende-Code abgeschlossen wird,
• daß der Mikroprozessor 6 jedoch bei Gleichheit lediglich einen Stückzähler Z inkrementiert und einen entsprechenden aktuellen Stück-Code in der Ereignis-Liste 500 mit dem aktuellen Stück-Code überschreibt.

The step 209 for the input / display routine of properties of a system or features of the device, the franking mode, together with a statistics mode already in the German patent application DE 19731304.3-53 with the title: method for Statistical Mode reload and for statistical recording statistics classes in the Storage of a data record, explained in detail. According to the task, instead of the statistics mode of the above-mentioned German patent application DE 19731304.3-53, a special storage mode according to the invention is now used. The latter is equipped with a correspondingly high number of memory locations for the entirety of the user information used to carry out any statistics in the data center needed. The statistics are therefore not created in the franking machine.
The memory mode according to the invention (steps 417 to 430) is called within a franking mode 400 by the OTP processor 50 of the security module 40 or by the separate microprocessor 6 of the meter. However, the storage of the usage data always takes place outside the security module, in contrast to the billing data. Such an arrangement for storing data on a use of a terminal consists of at least one memory and a microprocessor, wherein a non-volatile user memory 16 is formed for a cumulative storage of previous usage data, which is connected to the microprocessor 6. The microprocessor 6 is programmed to enter a memory mode and a communication mode. The microprocessor 6 is programmed in the communication mode for data transmission from the user memory 16 to a remote memory 31, wherein the data transmission is carried out to carry out a statistical evaluation of the usage data away from the terminal. The microprocessor 6 is programmed to restore the original memory capacity after the data transmission has occurred. The microprocessor 6 is programmed in memory mode:

• for comparison of current usage data with the cumulative stored usage data,
• For storing a new data record with usage data in the usage memory in historical order according to use, if usage data are changed or dissimilar to the usage data contained in the previously stored data record,
• for forming a counter for usage data,
• for incrementing the counter and replacing corresponding data of an already stored data record with new data which reflects the incremented counter reading if the current usage data are unchanged or the same as the data record already stored. For the non-volatile storage of the current usage data, a parameter memory 5a is provided. In the usage memory 16, the current and previous usage data are stored cumulatively non-volatile. In contrast to the handling of the billing data, the usage data may be overwritten or deleted after it has been transmitted to the data center by the OTP processor 50 of the security module 40 or by the microprocessor 6 of the meter. It is contemplated that the microprocessor is programmed to restore the original memory capacity in the user memory 16 by overwriting or erasing the memory contents as part of its reinitialization.
  FIG. 9 shows a routine with reinitialization of the user memory 16. After the communication mode 300 (FIG. 2), the point b is reached, and thus the beginning of the above-mentioned routine. Query step 211 queries whether user data has been transmitted to the data center during the communication. It is envisaged that after transmission and storage of usage data in the remote memory in query step 211 of a system routine 200 of the terminal it is determined that the data has been transmitted, that in step 213 the usage memory 16 is reinitialized and subsequently generates a display text in step 215 is branched back to the point s (Fig.2) of the system routine 200. The query step 211 may alternatively branch to a step 214 for evaluating the communication. Between the points b and d of the flow chart shown in Fig. 2 further inquiries may lie before a service mode is reached, which includes the storage mode according to the invention. If a user memory is too full, the service mode can also direct the user to the point g in order to automatically enter a communication mode II. It is envisaged that shortly before or during the memory overflow the usage data stored in the user memory of the terminal will be transmitted to the data center. In connection with Fig.7c will be explained in more detail below, as a full memory is detected. As part of the reinitialization 213, the list end code is set to the memory location N = 0 (FIG. 9).
  The machine can determine that the run variable n = N = 0 is set. As part of the reinitialization can also be done deleting the memory contents above the list end code. As a result, the memory area is rewritten or freed for new usage data and achieves the original status shown in FIG. 3d-1.
  Alternatively, the routine according to FIG. 9 may optionally be implemented with the deletion of memory areas as part of the communication mode 300 or 350, which have been explained in German patent application DE 19731304 A1 mentioned above.
  According to FIG. 2, the following steps take place in the storage mode: A parameter comparison in step 417 results, for example, in the fact that a new parameter value has been entered into the franking machine. In query step 418, a query is made as to whether a new postage value has been entered into the franking machine in order to branch to step 424. In step 424, a new line is added to a list if the new parameter value is the postage value. Thereafter, the next query step 419 is branched, because it could also be a further parameter value changed or first entered. Otherwise, the second query step 419 branches, if the first query in the query step 418 shows that the postage value has not been changed and has not been entered for the first time. In query step 419, a query is made as to whether a new weight value has been entered into the franking machine in order to branch to step 425. In this case, in addition to the query step 419 another query is run through. The additional branch condition to step 425 is that no new row has yet been entered in the same list. Otherwise, if in step 424 a new line has been entered in the same list, then in step 422 the new weighted value is written to the already-listed new line, where. the weight value represents a weight range. In step 425, a new line is entered in the same list, if not already done. Referring to the diagram in FIG. 3a, the information W1 from the line J = 1 is revised when a new line in the same round is already entered in the list has been. Thereafter, the query steps 418, 419 branches to the third interrogation step 420. If the queries indicate that the values have not been changed and have not been entered for the first time because a new row in the same list is already listed, then the third query step 420 is reached. In the third interrogation step 420, a query is made as to whether a new date has been entered into the franking machine in order to branch to step 426. In this case, in addition to the query step 420, another query is run through. The additional branch condition to step 426 is that no new row has been previously entered in the same list. Otherwise, the new date will be entered in the already listed new line. In step 426, a new line is added to a list if the new parameter value is the new date. Thereafter, the next query step 423 is branched. Otherwise, the next query step 423 branches immediately if the third query in the query step 420 shows that the date has not been changed and has not been entered for the first time. If, on the one hand, the change of a parameter is detected, then a branch is made to step 427 in order to store a new line in the list, if this has not already taken place in the same round. On the other hand, a large number of next inquiry steps can be run through without a change of a parameter being detected. In such a case, a branch is made to step 428 to increment a piece counter Z: = Z + 1. With reference to the illustration in FIG. 3a, the count in the range N1 in the data record of the line J = 1 is correspondingly corrected without a new line having to be stored if the use is carried out in the same way or repeated. In a subsequent step 429, the still available memory space in the memory area of the user memory 16 is checked. If the available storage capacity for storage next disparate use is still sufficient, then branched to step 431 with the known Frankierdruckroutine. On the other hand, if the list is almost full, a branch is made to step 430 to generate a warning and close the time interval of storage of usage information. before branching to step 431 with the franking print routine. The franking printing routine is preferably carried out by the microprocessor 6 in a manner known per se. Since the microprocessor queries the print request 405 constantly and faster in the system routine than can be replenished mail piece to be franked, the execution of the query step 349 is secured. Thus, indirectly, communication may be initiated with communication of the data to the data center and subsequent memory area erasure.
  FIG. 3a shows a memory format in a first variant. The Mikroprozesssor forms a piece counter Z by means of a memory Z is set to the value of 1 only for a first-time storage of one of the usage data and the storage of the record of new or changed and unchanged usage data is carried out in a new line J = 2. In a repeated storage of unchanged usage data, it is sufficient if the incremented count is entered in the area N1 of the first line J = 1, because an obviously similar use of the terminal is present.
  Each line J stores data for following information such as number Nj, postage value Pj, calendar status Cj (date), advertising cliché number Aj, weight value Wj, format Fj, shipping information Dj (form, type, destination) and, if necessary, certain detected errors ej.
  The memory lines J = 1, 2,..., 5,..., Are listed in the memory area in a historical order. A mailpiece stack of identically franked item pieces can be assigned a specific memory line if the batch items of the stack have been franked one behind the other, so that unevenly stamped item pieces do not occur. A first stack of similarly treated post pieces then corresponds to a first line J = 1, a second stack of similarly treated post pieces then corresponds to a second line J = 2, ..., a fifth stack of similarly treated post pieces then corresponds to a fifth line J = 5 Each of the memory lines J is then stored information for a similar use. Only in case of unequal usage, ie if at least one of the information changes, a new one must be created Line are written into the user memory 16 with corresponding new and the other previous usage data.
  Figures 3a and 3b show two different storage formats for historical storage of similar or dissimilar usage information of consecutive use according to a type of service. While in the variant according to FIG. 3a the memory locations in a memory line are arranged next to one another or consecutively in a row, so that the value for a specific information can be read from the placement of each of the memory locations in the row 3b, the memory locations are arranged in a group which has an identifier, the identifier replacing the placement in a row. The variant of Figure 3b allows placement of a group of three within a series of memory locations or within a row, which can now advantageously be arbitrary. The microprocessor is programmed to search for the current identifier of a parameter, which includes information about the order. He then reads out a triad of interest, which was filed as a result of a last storage process in memory locations. A triad has three memory locations in the example explained. However, this does not exclude that a grouping comprises a plurality of memory locations.
  A first triplet has bits B ₁₁ , bits B ₁₂ and bits B ₁₃ . The bits B _{11 of} the first memory location 16-01 relate to the value of a parameter. The bits B _{12 of} the second memory location 16-02 relate to the number of mailpieces having the same value of the parameter. The bits B _{13 of} the third memory location 16-03 relate to identifier for the type of parameter. It is envisaged that the part of the identifier which says something about the time sequence of the storage of groups of three is formed by an incrementable meter reading.
  The bits B _{11 of} the first memory location 16-01 of the first group of three relate to a first value of the parameter of the type postage value. The bits B _{21 of} the corresponding first memory location of a third Triplet affects a second value of the parameter postage type, with the second value of the parameter historically set later. Between the first and third group of three is at least one other triple. In the example chosen for FIG. 3b, there is a second group of three containing information relating to an identification number of the advertising cliché which was printed in addition to the franking on a mail item of the first batch. The identifier again comprises a part for the type of the parameter and a part for the time sequence of the storage of the second triplet.
  The usage memory 16 has a respective third memory location 16-03 per triple group, in which, assigned to the respective desired usage information, an identifier is stored. When querying the memory, the microprocessor searches for the current identifier of a parameter and then reads out a triplet of interest, which precedes the next identifier of a parameter. The value or the number of a user information is stored in each provided for this purpose respectively first memory location 16-01 or the number of similar usage information in each case provided in the respective second memory location 16-02 the triple. Compared to the variant shown in Figure 3a, however, several piece counter must be managed separately.
  FIG. 3 c illustrates a memory format using the example of a list-like representation of memory states (FIGS. 3 c-1 to 3 c-5) in the course of a historical storage of dissimilar or identical user information in a successive use according to a type of service. The aforementioned triad of memory locations is referred to below as a list.
  The number of lists results from the number of usage information. The latter are in the exemplary embodiment, for the sake of simplicity, the postage value, the weight level and the date. In fact, lists can be stored for a much larger number of usages become. Each list is processed by the microprocessor at each franking in the memory mode (steps 417 to 430) to be processed within the franking mode 400.
  At the beginning, each list is empty. A parameter comparison in step 417 results, for example, in that a letter with a postage value = 1.10 DM is to be franked, that the weight of the letter is classified in a first weight level GW1 and that the calendar module 8 indicates the date 31.08.98. In the memory mode (steps 417 to 430), the microprocessor generates the state shown in FIG. 3 c-1 in the usage memory 16.
  Subsequently, a similar use of the franking machine in the franking of the next letter. In this case, the microprocessor merely increments the piece counter in the three lists and thus generates the state in the use memory 16 shown in FIG. 3c-2 in the storage mode (steps 417 to 430).
  Now another eight letters of the same weight level GW1, with an equal postage value of 1.10 DM are franked with the same date 31.08.98. This results in the state shown in Figure 3c-3 in the user memory 16. It was only the piece counter increased to Z = 10. The first stack thus comprised ten similar letters.
  Subsequently, in case of unequal use, two letters of the second weight level GW2 are franked, with a second postage value of 3.00 DM, but with the same date 31.08.98. This results in the state shown in Figure 3c-4 in the user memory 16. In the date list, only the piece counter has been increased to Z = 12. In the two other lists for postage value and weight level, a new line was entered. From the list it can be seen: The franked first stack thus comprised ten similar letters and the historically subsequent franked second stack thus comprised two similar letters.
  On the following day on 01.09.98 a letter of the first weight level GW1, with a first postage value of 1.10 DM, will be franked. This results in the state shown in Figure 3c-5 in the user memory 16. Since the piece counter in the first line of the date list is Z = 12, the date refers to the entries in the first two lines of the lists for postage value and Weight level, but not on the entry in the last line. On the other hand, it is clear that another letter of the first batch was franked on the following day, because in the date list, a new line was entered, which is the last line entered in the two lists for postage value and the weight level assigned.
  Such a data storage is space-saving and still allows later after query the data from the data center any evaluation with appropriate allocation of data to each other.
  FIG. 3 d illustrates another advantageous memory format using the example of a list-like representation of the memory states (FIGS. 3d-1 to 3d-5). The difference from the format according to the first variant according to FIG. 3a is the avoidance of logging of identical events in a new line. Rather, only the qualitatively altered or disparate events are logged in the order of their occurrence. This takes place in the form of an event list 500 with assigned reference list 610. In the event list 500, code is entered whose reference to qualitatively changed or dissimilar events results from the associated reference list 610. From the event list 500, the history of a historical storage of qualitatively changed or disparate events (usage information) can be seen. The reference list 610 stores reference entries for the code for the different usage information.
  The reference entries require less memory space than the entries of the events because the same reference entry is not repeatedly entered if the same usage behavior is repeated and the same services are used.
  It is provided in a preferred embodiment that the nonvolatile memory 5a is a parameter memory in which the property type and value of the current usage information are detected as data by the microprocessor 6 of the postage meter when acquiring current usage data. The usage memory 16 of the device has areas for storing an event list 500 and a reference list 610. A program memory 11 is connected to the microprocessor 6 and contains a program for the memory mode 417b - 430b, whereby the microprocessor 6 is programmed,
• that code is stored in the reference list 610 and the event list 500, with each code storing data for describing the type and value of a property associated with the reference list 610,
• that one of the code for comparing the current usage data with the cumulatively stored usage data is taken from the event list 500,
• in that the associated data concerning the species are found and compared with the respective current data on property types of current usage data taken from the parameter memory 5a, wherein if the retrieved data is unequal to the species, a further code is taken from the event list 500 until all those codes of the event list 500 have been queried, which relate to qualitative properties and are set in relation to respective current data on property types of current usage data, wherein the microprocessor 6 generates a new code in the case of inequality of data to Art and in the Reference list 610 and the event list 500 writes, wherein also the data on the type and the value of the new property of the code zugeodnet stored in the reference list 610, and wherein the same nature of the microprocessor 6 nor the value based on the in the reference list (610) stored data with the value of the aktuel len property refers to equality and in case of inequality to the value of the new property a new code which stores data on the type and value of the new property associated with the new code in the reference list 610 and places the new code and associated one-piece code in the event list 500, which is terminated with a list end. Code is completed,
• that the microprocessor 6, however, only increments a piece counter Z when equal and overwrites a corresponding current piece code in the event list 500 with the current piece code.

Although the event list also stores the quantitative events accordingly for the same usage behavior, but not historically. The quantitative events are counted and also stored in the form of a variable code. In a sequential same use of the terminal, the codes are changed according to a number of pieces. For each quantitative event of a service type, a unique code exists in the program memory 11. The microprocessor accesses the predetermined code stored in the program memory 11 and writes it in the list. It changes at least one code in the event list 500 as part of the execution of each service. For the sake of simplicity, the service in the illustrated example is preferably limited to, but not limited to, franking mailpieces. A variety of different codes can also be reserved for another service or dynamically generated in the process. Each code is on a separate line in the list and comparatively requires only a small memory space in the user memory 16. Advantageously, the space required per line is reduced to a memory addressable by the microprocessor with a number of memory cells corresponding to the number of digits in a suitable number system translated codes. In practice, of course, for the machine language in binary code converted hexadecimal code used. In the simplest case, one byte, ie 8. 8 bits, is sufficient per line.

• Code 0 für das Frankieren an einem ersten Datum, ggf. mit Stunden,
• Code 1 für das Frankieren mit einem 1.Standardwert 1,10 DM,
• Code 2 für das Frankieren mit einem 2.Standardwert 3,00 DM, usw.
• Code 3 für eine Gewichtsstufe bis 20g,
• Code 4 bis 239 sind reserviert für weitere Eigenschaften,
• Code 240 für eine Frankierung von einem einzigem Poststück,
• Code 241 für ein Frankieren von zwei Poststücken,
• Code 242 für ein Frankieren von drei Poststücken, usw. bis
• Code 250 für ein Frankieren von elf Poststücken,
• Code 251 bis 255 reserviert für weitere Ereignisse und zur Steuerung.

For the sake of better understanding, the following clarification of the memory states takes place on the basis of codes converted into the decimal number system for one byte lines. The events illustrated with reference to FIG. 3d are defined, for example, by:

• Code 0 for franking on a first date, possibly with hours,
• Code 1 for franking with a 1.standard value 1.10 DM,
• Code 2 for franking with a 2. standard value 3.00 DM, etc.
• Code 3 for a weight level up to 20g,
• Codes 4 to 239 are reserved for additional properties,
• Code 240 for a franking of a single mailpiece,
• Code 241 for franking two mail pieces,
• Code 242 for a franking of three mail items, etc. to
• Code 250 for franking eleven mail pieces,
• Code 251 to 255 reserved for further events and for control.

As illustrated by the state shown in Figure 3d-1 in the usage memory 16, initially the reference list 610 is still empty and the event list 500 contains in a first line (in the first memory location) only one code 255 for the list end.
After billing for the first franking with the postage value of 1.10 DM, a first date code, the date and possibly the hour or more accurate time values are entered into the reference list 610 in a first storage location 611, for example. There are now "0 = date, time", ie reference code and the corresponding description entered. The microprocessor generates the first reference code. In this case, the memory location number of the first memory location 611 can be used to form a first reference code 0, for example by subtracting a constant code from the address code of the memory location. Using the memory location number of the second memory location 612, the default value code for the postage value may be generated. This and an associated description for the type and value, for example, "1 = postage value 110" are written in the reference list 610.

Thus, the state shown in FIG. 3d-2 results in the usage memory 16. The event list 500 contains four lines, namely a first line (first memory location 501) with a first date code 0 and a second line (second memory location 502) default value Code 1 for the postage value 1.10 DM, a third line (third memory location 503) with a one-piece code 240 and a fourth line (fourth memory location 504) with a code 255 for the end of the list.
The figure 3d-3 shows the state in the user memory 16 after the settlement for the tenth franking with the postage value of 1.10 DM. In the reference list 610 are still only two lines: "0 = date, time" and 1 = postage value The event list still contains only four lines with a first date code, a default value of 1 for the postage value of 1.10 DM, but with a ten-piece code 249 and again a code 255 for the end of the list.

The figure 3d-4 shows the state in the user memory 16 after billing for another two frankings, but with a second standard postage value of 3.00 DM. In the reference list is after the line "postage 110" still a second line "postage value 300 "registered. Starting from the first date code 0, the microprocessor may increment a counter to form a first default value code 1 and subsequently later a second default value code 2 which is automatically assigned in the second line to the description "postage value 300". The event list now contains six lines with a first date code 0 for the date, a first default value 1 for the postage value 1.10 DM, with a ten-piece code 249, with a second default value 1 for the postage value DM 3.00, with a two-piece code 241 and again a code 255 for the end of the list.

Figure 3d-5 shows the state in the usage memory 16 after billing for another franking, but again with a first standard postage value of 1.10 DM. In addition, this time also a scale 22 is connected, which provides a weight value 20g. In the reference list, the third line "2 = postage value 300" is entered after the first line "0 = date" and the second line "1 = postage value 110". In the reference list, after the third line "2 = postage value 300", a fourth line "3 = weight value 20" is still entered. The event list now contains nine lines with a first date code 0, a first default code 1 for the postage value of 1.10 DM, with a ten-part code 249, with a second default value code 2 for the postage value 3.00 DM, with a two-piece code 241, with a first standard value code 1 for the postage value 1.10 DM, with a first weight value code 3, with a one-piece code 240 and again with a code 255 for the list end.

Only a certain number of events, which are reflected, for example, with one byte per memory location (memory location), must be defined in the program memory 11 before a first startup. User behavior in mixed mail processing is characterized by a reduced collection of similar mail. A twelve-piece code 251 for franking twelve pieces of mail constitutes a limit found by experience, which is rarely exceeded by the user. When this limit is exceeded, a control code 252 and immediately thereafter a value code for the number of pieces from this aforementioned limit is written. Another control code 253 with immediately following value code allows the further pushing out of the limit for a piece number acquisition. According to a suitable number system selected limits can in principle also other memory space-saving counting ways.

The undefined events are dynamically defined during the machine runtime. The assignment of code to events should be during the term of the franking machine take place automatically. The events are included in the reference list only in the context of the storage mode before the franking process. The reference list can contain a detailed description of the event.

The method provides that a change of a parameter (property) is detected as an event in the event list 500. If an event occurs, which occurred earlier, then only an entry in the event list needs to be made. The entered code is also in the reference list 610 at a position pointed to by the pointer drawn in Fig. 3d-5.
For each entry in the event list 500, a constant number of bytes or at least 1 byte is advantageously reserved. The number of different events is preferably limited to 240. If this number is insufficient, however, then two bytes are consecutively written in an event list 500 *, the first byte having the code 254 and referring to a second reference list 620 (shown in FIG. 8). If two bytes with the code 254 follow one another, then reference is made to a third reference list 630, which is not shown for reasons of space. According to this principle, the detectable number of different events can be arbitrarily increased.

With reference to FIG. 2, a flow chart for the storage of usage data in the storage mode (sections 417 to 430) has been explained, which is executed within the franking mode 400. However, the method according to the invention is not limited to this specific embodiment. In principle, a storage of usage data can also take place after the completion of the franking print routine or immediately after the step 405 for determining a print request. In conjunction with a varied memory organization, the special sequence in the memory mode is also substituted, which is explained in detail with reference to FIGS. 4 and FIGS. 7a, 7b and 7c for two variants.

FIG. 4 shows a flowchart for the storage of usage data according to the memory organization explained in FIG. 3c. It is assumed that a list-like storage of features. The features are predetermined properties of the machine in connection with franking, in particular franking parameters and settings or errors. The processing takes place after determining the current feature or its change, for example via a comparison of the already stored parameters, settings and errors with the currently stored parameters, settings and errors that have occurred for each feature in the same way. In a first step 420a, it is determined that the "store feature" JOB is to be executed, and it branches to the first step 421a to call up a last feature M _m stored in the list. In a subsequent step 422a, the comparison is made with the current feature. If the characteristic is identical to the listed feature M _m , only the counter Z _{m of} the listed feature is incremented to M _m : = M _m + 1 (step 425 a). Otherwise, in step 423a, a new feature M _{n is added to} the list end and the associated counter is set to a seed Z _n : = 1 (step 424a).

In FIG. 7b, a flowchart for the storage of usage data is explained in more detail, which relates to the storage format explained with reference to FIG. It is assumed that a storage of events in an event list 500 as listed specific code with associated reference list 610. An event should now be added to the history as a result of a state change of qualitative properties. The list-like storage of events in the event list relates to both qualitative and quantitative properties, the latter each having a hard-coded code associated with it. The reference list 610 associated with the event list 500 contains only events related to qualitative properties and an associated code generated in a dynamic programmable dymamic process. One to each Event exists at least a pair of data or code in the event list 500. The one of the qualitative properties associated programmable code is entered in a first memory location 501. The fixed programmable code associated with the quantitative properties is entered in a second memory location 502. The storage in the event list 500 ends with the list end code stored in the third storage location 503.
To store data in the event list 500 or reference list 610, a number of steps are executed by the microprocessor. If, within the franking mode (for example, after the billing according to step 416 in FIG. 2), a phase has been reached in the sequence that the "save event" JOB is to be executed, then in a step 417 b a call is made to the parameter list, in which a multiplicity to current properties are included. For example, a property E1 of the first kind relates to the set postage value, where the index u of the event E1u illustrates the selection of one of 1, ... u, .., w different discrete values. A property E2 of the second kind relates, for example, to the set weight level. According to a postage tariff table, although a weight level is assigned a discrete postage value, further shipping parameters are included in the postage value calculation, so that the postage value can not always be deduced from the weight level. The weight range is therefore divided so finely in weight levels that the concrete weight in the range of a change is detected from one to the next level. A later evaluation in the data center or at the mail carrier can be used, for example, according to a statistics on user behavior to specify the grading in the postage tariff table differently. For example, a third type property E3 refers to the automatically set date for a day, which may be similarly subdivided into discrete time increments as previously described. There are further properties of i-th to k-th type, which are suitable to make services at least statistically detectable. The different properties may be the same values (For example, 110 g and 10 10 10 ^-2 DM) assigned. They are then distinguished by their type (corresponding to a unit of measure in a physical quantity). The order of the species is arbitrary. The parameter list is realized, for example, in the nonvolatile memory 5a. It contains a preprogrammed number of memory locations in the program memory for the different parameters (eg postage value, weight value, time of day / date, etc.).
FIG. 5 shows the basic configuration of a memory location of the parameter memory. A first byte is provided for the identification "i-th ART" of a parameter type (postage, weight, date, etc., etc.) and another four bytes "u-th VALUE" for the concrete value of the parameter. For each type, the discrete parameter value may assume only one of the plurality of different possible values. Each discrete value should be considered a qualitatively different property.

FIG. 6 shows the basic configuration of a memory location in the user memory 16 for the reference list 610. A first byte is for an associated code "CODE", another byte "ART" is for the identification of a parameter type (postage, weight, date, .. etc.) and another four bytes "VALUE" are for the specific value the parameter provided.

According to FIG. 7b, a call-up of the parameter list takes place in step 417b. A flag P named MERKER_NEW_EIGENSCHAFT is set to P: = 0 in step 418b, and the first run variable i representing the number of the parameter to be processed is set to the first parameter, that is, the running variable i : = 1.

A first local variable E receives the value of the property to be processed ( E _iu ) in step 419b. Since the run variable i : = 1, the local variable E = E _{1u becomes} the first in the parameter list queried listed property, eg postage value, to which the specific value 1 is assigned in this example. In step 419b, a second run variable j : = 1 is set and a byte counter is set to the initial value n: = N bytes, which corresponds to the currently reached number of lines in the event list.
Now branch to point a1 of a search routine 600, which searches the event list 500 for certain free-defined code. A simplified search routine 600 will be explained in more detail with reference to FIG. An interrogation step 600a for N = 0 is anticipated, which enables an apportionment to the point c1 in FIGS. 7b or 7c, which is effective only immediately after a re-initialization. In the latter case, the previous value N = 0 is increased by one (step 704). Otherwise, if N ≠ 0, then in the next first sub-step 601 of the search routine 600, a memory location is selected by now defining the n-th memory location as n: = N-j. Thus, a preparation is made in order to be able to compare in the following step 420b the successively called properties of the already detected property types of the reference list with the property type called from the parameter memory. The second run variable is set to j : = 1 and is subtracted from byte counter value N. This results in the last line below the list end code, which is addressed and read out. The evaluation of the code read out takes place in sub-step 601 by comparison with the codes which are stored in the program memory 11 in a defined manner.
The microprocessor now determines in sub-step 602 whether, in the event list, there has been stored on the nth memory location (line for the nth byte) a hard-coded code because of a predefined meaning. If so, then point a3 is reached and branched to sub-step 603 where the one second variable j is incremented by one before then branched back to sub-step 601 to decrease the byte counter value n, thus causing the penultimate line below the end-of-list code gives 255. If it is determined in sub-step 602 that a code programmed in advance (due to non-predefined meaning) has been stored, the point a2 is reached. It this code denotes a qualitative property, which can be evaluated in the subsequent step 420b.

The respective second byte for the identification of a parameter type in the memory location of the reference list shown in FIG. 6 contains-each associated with a code-a characteristic value for a qualitative property "A second local variable EL is set to this characteristic value which corresponds to the qualitative property of the event code listed in the event list 500 in the nth memory location. The characteristic value shall correspond to one of the types, for example the date, weight, postage, etc., etc. List 610 contains, for example, the assigned characteristic value for this type for the (freely-defined) code listed in the event list 500. The second local variable EL, which was thus set to a characteristic value corresponding to the second byte of each memory location ( 6), now serves as an actual value The characteristic value of the ith type serves as a setpoint for a comparison with the actual value of the second one local variable EL in step 420b.

In step 421b it is now checked whether the same species was found. If so, then the process branches to step 422b. In step 422b, the search in the reference list 610 continues to the value for which 4 bytes are reserved in memory location in third place. A value of a current property from the parameter memory to which the first local variable E has been set in step 419b may now be compared in step 422b with the aforementioned value detected in the reference list 610.

However, if it is determined in step 421b that the same type has not been found, a branch is made to step 423b in order to check whether all characteristics of properties already acquired earlier have already been called. This is only the case with a byte counter value of n = 0. At n = 0, a branch is made to the point c1 at the beginning of a subroutine 700. Are still not all characteristics of already previously detected properties have been called before, ie n ≠ 0, then the search routine 600 is branched back to the point a3. The search routine 600 thus serves to prepare for the step 420b. In step 420b, prior to the comparison, the second local variable EL is set, the latter being set to the corresponding characteristic from the reference list 610, the characteristic being assigned to the code which is in the nth memory location in the event list 500 is read. In this way, a check is made one by one, and then, in step 421b, whether one of the previously acquired characteristics matches the called current property in kind. If this is the case, then the values can still differ for the same type and it is again branched to step 422b. In step 422b, in the reference list 610, the stored concrete value corresponding to the value of the variable E is searched for. Of course, if the check in the following query step 424b shows that there is a match also in terms of the value, then no new qualitative property needs to be detected in the event list 500 and a branch is made to step 425b to increment the first run variable i by one. The run variable i is thus set up in the next property type. In the following query step 426b it is determined whether all k types have already been checked for which current parameter values can be detected in the parameter memory. If this is not the case, then branch back to step 419b. This backward branching provides a loop to compare all current properties with the previously stored properties.

Now, by means of the search routine 600, the event list 500 is again searched from top to bottom for a code, the retrieved code corresponding to a type checked in step 420b for compliance with the currently relevant property type "ith ART" in step 422b, if the match is determined in step 421b, the current property value is retrieved from the parameter memory to compare with the stored property value from the reference list 610. However, if inequality is detected in step 424b, a branch is made to point c2 for a second entry into subroutine 700. Now, the microprocessor starts the subroutine 700, which will be explained in more detail with reference to FIG. 7c.

Figure 7c shows the flowchart subroutine 700 of Figure 4b which is at least suitable for storing the code for the new property in the event list 500 or for giving at least one warning if the list is full. In sub-step 701 at point c1 at the beginning of subroutine 700, the memory location number reached is first queried in accordance with the last stored freely defined code and compared with a desired number. If the target number has not yet been reached, then the reference list 610 is not yet full. Now branch to sub-step 702 to complete the reference list. Otherwise, when the target number is reached, a "list full" warning is generated in sub-step 706 and branched to the point g (Figure 2). Starting from point c2 for a second entry into the subroutine 700 and from the sub-step 702, the query step 703 is reached, in which the achieved byte count value N is compared with a corresponding setpoint value to decide whether the event list 500 is full or not yet full. If the event list 500 is full, the program branches again via the sub-step 706 to the point g (FIG. 2). Otherwise, if the event list 500 is not yet full, the process branches to sub-step 704. In sub-step 704, a new event is added to the list of events as a code at the list end, and then the byte counter value is incremented N: = N + 1, before branching to sub-step 705. In sub-step 705, the flag_new_property p: = 1 is set, and then the point c3 is reached at the output of the subroutine.

If none of the stored types was equal to a type currently detected in the parameter memory, then branching is made to the point c1 at the beginning of the subroutine 700. But was the value listed in the reference list to a previously stored type not equal to the value currently detected in the parameter memory, then branching to point c2 at the beginning of subroutine 700 is to store the code for the new property in event list 500, the flag new_property p being at the value 1 or otherwise to give at least one warning if the list is full.

From step 426b of the flow chart shown in FIG. 7b, a branch is then made to step 427b when the loop has been processed and thus i> k, ie all current properties have been compared with the previously stored properties. In step 427b, the flag_new_property p is evaluated, which can only have the values 0 and 1. If p = 1, a branch is made to step 428b and a one-piece code 240 is entered at the end of the event list. The byte counter value is incremented by one. N: = N + 1. In a final step 430b, the end-of-list code 255 is set at the end of the event list 500.
Otherwise, at p ≠ 1, a branch is made from step 427b to step 429b to increment the piece code at the end of the event list by one. From steps 429b or 430b, the end of the process is reached.

It is envisaged that a line-like storage of codes in an event list, which are related to usage information, wherein a usage information is the number of pieces.

Usage information related to a new event is assigned a freely definable code by the microprocessor, the association being stored in a reference list. The reference list can be stored in a second memory area of the user memory 16 or on reserved memory locations 16-0n + 1 in the first memory area.

It is further contemplated that the codes related to the number of items are fixed, with each new event an assigned piece number counter is set to a value of one and in the event list a corresponding one-piece code is entered. The firmly defined code, which are not freely definable and not the reference list can be removed, are stored in the program memory 11 before.

The invention is not limited to any of the present embodiments. The storage of usage data may be at any point in the usage memory and in any suitable form, such as variable length data sets. Although cumulative storage significantly reduces storage space requirements, it is not necessary with a very large amount of free space available. On this basis, a number of other variants within the scope of the claims conceivable. Thus, obviously other other embodiments of the invention can be developed or used, starting from the same basic idea of the invention, which are encompassed by the appended claims.

Claims (15)

1. An arrangement for the storage of data about the utilization of a terminal device, comprising at least a memory and a microprocessor, wherein a non-volatile utilization memory (16) for the storage of previous utilization data is connected with the microprocessor (6), characterized in that the microprocessor is programmed for

   - entering a storage mode for storing utilization data according to a preceding utilization,

   - entering a communication mode, wherein, in said communication mode, the microprocessor (6) is programmed for transmitting data from the utilization memory (16) to a remote memory (31), the data transmission being made for carrying out a statistical analysis of the utilization data remotely from the terminal device, and

   - reinitializing the utilization memory (16) for restoring the pre-defined storage capacity in the utilization memory (16) after completing the data transmission.
2. An arrangement according to Claim 1, characterized in that, in the storage mode, the microprocessor (6) is programmed for

   - comparing current utilization data with the utilization data already stored,

   - storing a new data record with utilization data in the utilization memory in historical order according to utilization when said utilization data are changed or not equal compared with the data contained in the data record previously stored,

   - forming a counter for utilization data, as well as

   - incrementing the counter and replacing respective data of a data record already stored by new data reflecting the incremented counter reading when the current data are unchanged or equal compared with the data record already stored.
3. An arrangement according to Claim 1, characterized in that the utilization memory (16) is arranged separately from a non-volatile memory (5a, 41) for the post registers and that utilization data are stored independently from accounting data.
4. An arrangement according to Claims 1 and 2, characterized in that the microprocessor is programmed for organizing the allocation of storage locations to the data of the data record in the current process according to the events occurring, wherein the data of the data record include at least one parameter value and one piece number.
5. An arrangement according to Claims 1 to 3, characterized in that the terminal device is a franking machine with a microprocessor (6) that is programmed for entering an input/display routine (209) and a franking mode (400) within a system routine (200), wherein the input routine (209) for collecting current utilization data by the microprocessor of the franking machine in the non-volatile memory (5a) includes instruction steps for the input of a current utilization information comprising at least the franking value, the weight and the dispatch, and wherein the franking mode (400) comprises the storage mode (417 - 430, 417a - 430a or 417 - 430b) for a cumulative storage of the current utilization information together with the preceding utilization information.
6. An arrangement according to Claims 1 to 5, characterized in that the non-volatile memory (5a) is a parameter memory in which the microprocessor (6) of the franking machine, when collecting current utilization data, collects the property type and value of the current utilization information as data; that the utilization memory (16) includes areas for the storage of an event list (500) and a reference list (610);
   that a program memory (11) is connected with the microprocessor (6) and contains a program for the storage mode (417 - 430) by means of which the microprocessor (6) is programmed;
   that codes are stored in the reference list (610) and the event list (500), wherein, assigned to every code, data for describing type and value of a property are stored in the reference list (610);
   that one of the codes is taken from the event list (500) for comparing the current utilization data with the cumulatively stored utilization data;
   that the assigned data are found according to their type and compared with the respective current data of the types of property of current utilization data taken from the parameter memory (5a), wherein, if the called data of the type are unequal, a respective further code of the event list (500) is taken until all of those codes of the event list (500) that refer to qualitative properties have been called and compared with respective current data of types of property of current utilization data, wherein, in case of inequality of the type data, microprocessor (6) generates a new code and writes it into the reference list (610) and the event list (500), wherein also the data of the type and value of the new property are stored in the reference list (610) assigned to the code, and wherein, in case of equality of the type, microprocessor (6) checks the value on the basis of the data stored in the reference list (610) and the data referring to the value of the current property as to equality and, in case of inequality with the value of the new property, generates a new code, stores the data of type and value of the new property in the reference list (610) assigned to the new code and includes the new code as well as a related one-piece-code into the event list (500) that is concluded with a list-end code;
   that the microprocessor (6), however, in case of equality, only increments a piece counter (Z) and overwrites a respective current piece-code in the event list (500) with the current piece-code.
7. A method for the storage of data about the utilization of a terminal device, characterized by

   - provision of a utilization memory (16) with pre-defined available storage capacity;

   - utilization of a terminal device with a collection of current utilization data that are assigned to at least one utilization information, wherein the current property of a utilization or of an alteration is determined by comparison of respective current utilization data with utilization data already stored; and

   - storage of a new data record with utilization data in the utilization memory (16) if the data are changed or unequal compared with the data record stored before;

   - incrementing of a counter and replacement of a data record already stored before by new data that reflect the incremented counter reading if the current utilization data are unchanged or equal compared with the data record stored before;

   - data transmission of the utilization data from the utilization memory (16) to a remote memory; and

   - reinitialization of the utilization memory (16) for restoring the available pre-defined storage capacity in the utilization memory (16).
8. A method according to Claim 7, characterized in that the remote memory is a memory (31) of a remote data central and that, a short time before or upon storage overflow, the utilization data stored in the utilization memory (16) of the terminal device are loaded into the memory (31) of the data central.
9. A method according to Claims 7 to 8, characterized in that, after the transmission of utilization data and their storage in the remote memory, the inquiry step (211) of a system routine (200) determines that the data have been transmitted; and that, in step (213) within the scope reinitialization of the utilization memory (16), a list-end code is put on a memory location at the start of an event list and subsequently, in step (215), a display text is generated, before branching back to point s of the system routine (200).
10. A method according to Claim 7, characterized in that, at the data central, there is generated statistics of any kind on a utilization of a terminal after loading those utilization data from the utilization memory (16) into the remote memory (31) that had been stored in a non-volatile matter in the utilization memory during the utilization of the terminal device.
11. A method according to Claim 7, characterized in that a space-saving historical storage method is used for the storage in the utilization memory, wherein data of a data record are stored or replaced in a line-by-line manner during the utilization of the terminal device with a collection of current data that are assigned to at least one utilization information, and that the data are compressed by means of a modem during their transmission to the data central.
12. A method according to Claim 9, characterized in that, in step (213) within the scope of the reinitialization of the utilization memory (16), a list-end code is placed at the start of an event list and the previously stored data on the other memory locations are deleted.
13. A method according to Claim 9, characterized in that a data record with utilization data in the utilization memory comprises at least one code; and that codes referring to utilization information are stored line by line.
14. A method according to Claim 13, characterized in that a utilization information is a quantitative property and/or a piece number.
15. A method according to Claim 14, characterized in that a freely definable code is assigned to a utilization information related to a new event, the assignment being stored in a reference list; that the codes referring to the piece number are clearly defined, wherein, for every new event, an assigned piece number counter is set on a value one and a respective on-piece code is entered on the event list.

Images