JPH04317160A - Data processing system - Google Patents

Abstract

PURPOSE:To share data in a program operating on a computer whose architecture is different. CONSTITUTION:The procedures of access and architecture conversion are previously registered for respective classes of data. Furthermore, a flag showing generated architecture is set on data on a file. When there is the read request of data from an application, the flag showing architecture on data is compared with the architecture of hardware to be executed. When they do not coincide, the procedure of the architecture conversion of the class is called, is converted into a usable state, and is returned to the application so as to be refferred to. Thus, data can be shared on the program operating on the computer different in architecture without using a converter.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing system for sharing data between programs running on computers with different architectures.

0002

2. Description of the Related Art When the same program is ported and executed on computers with different architectures, the arrangement and structure of data may differ due to operations specific to each architecture. In this case, in order to share files, conventionally a data conversion program called a converter was created separately, and the data was passed through it to convert the format before use. "Operating Manual"; Published by Fuji Xerox Co., Ltd., June 1988).

0003

[Problem to be Solved by the Invention] Conventionally, when programs running on computers with different architectures share data, it is impossible to access it unless the data is actually converted using a conversion program. Therefore, if machines are connected via a network and each server contains data of different types, each machine cannot share files containing data of different types at the same time. Therefore, when a network was set up, there was a problem that the file could not be used unless a converter was installed and the file was converted to a separate file. Additionally, in order to create such a converter, it was necessary to investigate everything about how files are exported, and there was also the problem that the converter could not be created until the file format had been completely determined.

[0004] The present invention is aimed at solving the above-mentioned problems, and involves comparing the data format and the format of the data request source at the time of a data reference request, and if they do not match, converting the data to the format of the data request source and returning the data. The purpose of the present invention is to provide a data processing system that allows data to be received in a desired format and that allows programs running on computers with different architectures to share data without the need for a converter.

[0005]

[Means for Solving the Problems] As shown in FIG. 1, the present invention provides a data processing system in which data is shared by programs running on computers with different architectures. a management means 2 that manages data for each type of data and also manages mutual conversion procedures and data access procedures between different data formats; and a registration means 3 that registers mutual conversion procedures and data access procedures between data formats. The management means compares the data format with the format of the data request source using the identifier at the time of a data reference request, and if the formats do not match, calls the conversion procedure to convert to the format of the data request source and return the data. It is characterized by

[0006] Furthermore, the management means 2 is characterized in that it changes the identifier to an identifier indicating the format of the data conversion result.

[0007]

[Operation] In the present invention, the access and architecture conversion procedures are registered in advance for each data class by the registration means, and furthermore, the management means sets a flag indicating the architecture created for the data on the file, and the application When there is a request to read data from , the management means compares the flag indicating the architecture on the data with the architecture of the executing hardware, and if they do not match, calls the architecture conversion procedure for that class and uses it. After converting the data into a possible state, it is returned to the application so that it can be referenced, making it possible to share data among programs running on computers with different architectures without using a converter.

[0008]

[Example] Fig. 1 is a diagram showing the functional blocks of the present invention, Fig. 2
3 is a diagram illustrating the device configuration, FIG. 3 is a diagram illustrating data and a data manager, FIG. 4 is a diagram illustrating data processing according to the present invention, and FIG. 5 is a diagram illustrating a processing flow. In the figure, 1 is a data storage means, 2 is a data management means,
3 is a conversion/access procedure registration means, 11 is a CPU, 12
is the main memory, 13 is the display, 14 is the input section, 15
16 is an external storage device, 16 is a bus, 20 is a node manager, 21 is node subclass data, 30 is a body manager, 31 is body subclass data, and 40 is an application program.

The data storage means 1 shown in FIG. 1 stores hypertext type data. Hypertext is text that has a structure; for example, in a book, chapters, sections, etc. are the nodes, and a single sentence, that is, the basic unit of text, is the body. Note that the area above the node is a text file.

As shown in FIG. 3(a), this node includes a linear 21a, a spatial 21b, and a box 21c.
, browser 21d, search 21e, and so on. For example, in a book, the linear 21a has meaning in the order of chapter 1, chapter 2, etc., and is a node for creating a sentence. The spatial 21b is related to positional information, indicating where it is, and is a node used when drawing a diagram, and the box 21c is a node for storing some text like a holder. These linear, spatial, and box nodes are called data type nodes, while browser and search are called functional type nodes. The browser 21d is a node for visually showing the relationship between nodes on the screen and the structure of the nodes, and the search 21e is a search node for tracing nodes that have a certain relationship.

[0011] Furthermore, as shown in FIG. 3(b), the body includes a link 31a, a bitmap 31b, and a text 31.
There are types such as c, graphics 31d, and so on. For example, the link body 31a is used to relate bodies, the bitmap body 31b is a cutout of a part of an image, and the text body 31c represents a string of characters that has a meaning. , the graphics body 31d similarly shows a group of images with meaning.

These data are stored in data storage means 1 and managed by data management means 2, respectively. The data management means 2 is the node manager 20 shown in FIG. 3(a) or the body manager 30 shown in FIG. 3(b), and each data access procedure and data conversion procedure are is registered. The data management means 2 and the conversion/access procedure registration means 3 are a data management section 111 and a conversion/access procedure registration section 112 of the CPU 11 shown in the device configuration of FIG.
Corresponding to this, a flag 122 indicating data and data type, and a conversion access procedure 123 are stored in the main memory 12, and the CPU
11, the main memory 12 is connected to the display 1 through the bus 16.
3. It is connected to the input section 14 and the external storage device 15.

[0013] By the way, when an application issues a request to access data called linear, the node manager 20 calls the data access procedure of the linear 21a, and detects that the request source and the data managed by the node manager are of different types. If this is the case, the manager determines the type, converts the type using the data conversion procedure, and returns it to the application. Depending on the system, for example, in the case of 32-bit integer data, the upper 16 bits and lower 16 bits may be reversed, or the real number type of the data may be different. It is converted through a procedure.

To explain this situation with reference to FIG. 4, when the application 40 makes a data access request, the node manager 20 looks at the judgment flag 22a from the node data on the file, the management information that is the node manager's data 22, and makes a request on the application side. It is determined whether the format of the data matches the format of the data, and if they do not match, the type is changed using a conversion procedure and the data is passed to the application. In this case, when data is converted, the determination flag 22a is changed to a flag indicating the converted format. This is because there was a request for data in a different format from the application side, and there is a possibility that there will be similar requests in the future, so we converted the data format in advance and made it clear by changing the flag that it was converted. This is to help with processing.

The present invention was implemented using a system called "Viewcars" developed by the applicant. In this system, data is composed of upper classes called nodes and bodies, and subclasses defined for each data type. , node, and body data operations were handled by the respective node managers and body managers. For this reason, we implemented a function to register data access procedures and architecture conversion procedures in the node manager and body manager, and implemented an architecture judgment flag for node and body data, and each manager When a reference request is made, the registered data access procedure is called to access the data and determine the data architecture, and if it does not match, the registered conversion procedure is called to convert and return the data. That's what I do.

Next, the processing flow of the present invention will be explained with reference to FIG. When there is a data access request, the node data at that time is read and the determination flag is checked (steps 41 to 43). Judging from this flag, if conversion is necessary, it calls a conversion procedure suitable for the data type, changes the judgment flag, writes the node data, and returns the data to the side that made the access request (step 4).
5-48). If no data conversion is required, the data will be returned as is. In the linear card conversion example, the conversion procedure is called using the upper 1 of the 32-bit integer in the data
The 6 bits and lower 16 bits are exchanged and the format of real number data is converted (51, 52).

[0017]

[Effects of the Invention] As described above, according to the present invention, in the past, the data access part and the architecture conversion part were separated, so converter creation could not be started until the file format was determined, and the program When there is a change in the file format, there has been a problem that the converter forgets that the corresponding conversion module is in the converter, or that it is easy to make mistakes because it is a separate program.However, according to the present invention, the data class Because the parts that are accessed and the parts that perform conversion can be centrally managed, mistakes are less likely to occur, and the data manager can create conversion functions even if the file format is not determined.

[Brief explanation of drawings]

FIG. 1 is a diagram showing functional blocks of the present invention.

FIG. 2 is a diagram showing the device configuration of the present invention.

FIG. 3 is a diagram for explaining data and a data manager of the present invention.

FIG. 4 is a diagram for explaining data processing according to the present invention.

FIG. 5 is a diagram showing a processing flow of the present invention.

[Explanation of symbols]

1...Data storage means, 2...Data management means, 3...Conversion/
Access procedure registration means, 11...CPU, 12...Main memory, 13...Display, 14...Input section, 15...External storage device, 16...Bus, 20...Node manager, 21...Node subclass data, 30...Body manager ,
31...Body subclass data, 40...Application program.

Claims (2)

[Claims] Claim 1: In a data processing system in which data is shared by programs running on computers with different architectures, data is managed for each type of data to which an identifier for identifying the data format is assigned, and the data is managed in different data formats. management means for managing mutual conversion procedures and data access procedures between data formats; and registration means for registering mutual conversion procedures and data access procedures for data formats; A data processing system characterized in that the data is compared with a format of a data request source, and if the formats do not match, the conversion procedure is called to convert the data to the format of the data request source and return the data. 2. The data processing system according to claim 1, wherein the management means changes the identifier to an identifier indicating a format of the data conversion result.