US 5,754,849 AGrant
Self-Describing Object Providing Dynamic Manipulation of Heterogeneous Data Values and Semantic Identity Between Memory and Transmission Representations
Issue Date:1998-05-19
•13 Claims
•8 Drawing Sheets
Abstract
An object for use in a computer memory and for transmission between computers flexibly stores a large variety of different types of data, in such a manner as to allow easy and efficient transfer and copying of the data between computers. The object includes any number of data elements, each of which is self-describing using type coding. Certain data elements may include other data elements, thereby providing for arbitrarily complex hierarchical arrangements of data in memory. The type codes may be based on the frequency of the types or values of data operated upon, such that the data itself may be used as a type code for at least one type of data, thereby reducing the memory and bandwidth requirements for handling the object. The object is capable of creating a data stream output for copying and transmission that preserves the semantic structure of the object, particularly its hierarchical organization, and is further capable of reconstituting the hierarchical arrangements. Because of the self describing nature of the object and its data elements, marshalling is not needed for inter-computer transfers of the object.
Metadata
Assignee
- Wayfarer Communications, Inc.
Inventors
- Landon M. Dyer
- Robert V. Welland
Application Information
Application Number:US 5939682
Filing Date:1996-01-30
Priority Date:1996-01-30
Art Unit:237
Classifications
IPC:
G06F 1730
Field of Search:
395360365235364612;613;497.01;425;497.0248230492200;900
Patent Drawings (8 sheets)
Description
COPYRIGHT NOTICE
A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by any one of the patent disclosure, as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all copyright rights whatsoever.
BACKGROUND
1. Field of Invention
The present invention relates to data storage arrangements generally, and more particularly, to data storage arrangements useful in application development environments for representing heterogeneous data types in a uniform manner for simplified transfer of data between client and server computers.
2. Background of Invention
Client-server computing has become the predominant model of distributed computing, paralleling the increasing performance of desktop computers and workstations. In a client-server distributed computing environment, multiple computers are connected in a network, and a computer may operate both as client, a consumer of resources and data, and a server, a producer of resources and data for the clients.
In any client-server environment, a client requests operations of a server through a remote procedure call (RPC). In a remote procedure call, a process on a local computer, the client, invokes a process, the server, on a remote computer. Typically, the client transfers data to, or requests and receives data from, the server. In order for both computers to operate on a set of transferred data, there must be an agreed upon set of semantics between the client and the server for defining the arguments passed across the call. This is because both the client and the server may have different internal architectures for representing data, and thus, explicit specification of types of arguments is used to communicate data between the client and server.
For example, if the client requests a procedure from a server that supplies a graphics file from the server, the client may expect that the resulting data be constructed a file name, file type, creation date, data bytes!. The server must have information that defines the data being sent as the same, in the same order so that the client can properly receive and process the data. The determination of the types of data used in a given remote procedure call between a client and a server is determined at the time the client and server applications are written and compiled into executable code.
In order to ensure that both the client and server have the same semantics for a call, the executable code of the sending computer receives the data from the memory, and formats the data being into a standard, network representation. The computer receiving this data then reconstructs this representation to reproduce the original data, which may then be stored in the receiving computer's local memory and acted upon. In most computer systems, the data is stored in memory using implicit typing. That is, the type of the data is not stored directly with the data itself. Instead, the computer program, typically in the procedural module and header files, holds the type information about the nature of each piece of data it receives. This is done in order to save storage space: if every piece of data contained explicit type information, the overall amount of data that could be stored in a given memory would be significantly reduced. Thus, implicit typing is used to maximize local memory capacity.
The conversion and construction of data from its memory representation in a computer memory to its transmission representation on a transmission medium is called marshalling. In a general purpose RPC mechanism explicit typing of the transmission representation is used, since the memory representation used by client and server computers with different memory architectures typically does not allow the memory representation used by one computer to be directly used with the second computer. Thus, the data must be marshalled from its implicit memory representation to an agreed explicit transmission representation. However, marshalling routines are specific to each RPC mechanism, since they designed to construct the data for a particular and machine architecture RPC mechanism. Thus, because of the differences in memory and transmissions representations, computer systems that use different RPC mechanisms, are unable to communicate and freely exchange data in a client-server relationship.
In a conventional system then, once the client and server applications have been compiled, the format of the data being transmitted between for a given procedure call is fixed. The server cannot arbitrarily change the format of the data. Using the previous example, the server could not arbitrarily send back to the client the graphics file as creation date, file permissions, file type, data bytes, file name!, or some other arrangement that is unanticipated by the client. This is because both the client and server have been compiled with specifically linked in marshalling routines. To change the data format requires recompiling, and thereby distributing to the end users or system administrators, new versions of the client and server. Any attempt to arbitrarily change one side of the transaction will cause the other side to at best ignore the data, likely misinterpret it, or worst, crash.
Accordingly, it is desirable to provide a data storage format that preserves the semantic structure of the data between its memory representation in a computer memory and its transmission representation in a transmission medium, in order to eliminate the need to marshall data for communication between clients and servers.
In most application development environments, including object-oriented environments such as Microsoft's Visual Basic or Visual C++ development environments, applications programmers are provided with a rich set of predefined program components, such as objects, modules, and the like, to be used for constructing end user applications, including client-server applications like those described above.
In conventional object oriented systems, the provided objects typically include a number of objects for handling different types of data formats. For example, in Visual Basic, there are provided different objects that separately handle strings, arrays, and so forth, each object having methods particular to the manipulation of one type of data. It would be desirable to provide in an application development environment a single object that flexibly handles a variety of different data types, and allows that object to be easily transported from one computer memory to another without the need for the applications programmer to write the marshalling routines conventionally required, as described.
Finally, in conventional application development environments, data in an object, once instantiated as a particular type, cannot be easily changed to being data of a different type. Typically, there are separate classes for handling significant types of data, such as strings, arrays, maps, and the like. Once a data value is created in an object of one type, it cannot be immediately converted to object of another type. While casting of data types is known, casting does not permanently alter the type definition of a data element: The ability to dynamically manipulate and redefine the types of any data element in a data arrangement is thus desirable to increase the programmer's flexibility in handling the data elements.
SUMMARY OF THE INVENTION
In accordance with one aspect of the present invention, there is provided for an application development environment an object that is capable of storing a large variety of different types of data in a manner that allows simplified transmission of the internal data of the object from a first computer to a second computer over a transmission medium in a communications network. The object internally represents each data element using explicit typing, and arranges the data in a manner that allows it to be read from its memory representation in the first computer into a data stream transmission representation for communication to a second computer, where the object is reconstructed to the equivalent memory representation. The transmission representation in a transmission medium, while semantically identical to the memory representation of the object, is location independent, and thereby may be transported directly between computers, and then reconstituted into another instance of the object. In this manner the object is self describing of the semantic structure of its data, regardless of whether the object is stored in memory or being transmitted.
The present invention allows the applications programmer to transmit complex data arrangements between computers on a network without having to write marshalling code in a first computer that converts the data arrangement of the object in the first computer into a format capable of transmission to a second computer, or the marshalling code in the second computer that reconverts the transmitted data to its original form. This enables programmers to more easily create and maintain client-server applications.
In one embodiment of the present invention, the object, usefully called a "value object," stores in each instance of a data element a type code that defines a type of data, and the data itself. In a preferred embodiment, the types of data supported include scalar types with fixed sizes, blobs, which are variably sized data elements that do not contain other data elements, and aggregations, which are variably size data elements that may include other data elements, either of the same type or heterogeneous types. The value object includes methods to get, set, add, remove, modify, test, and copy particular instances of data or sets of data. As a further aspect of the present invention, any instance of a data element with a particular type may be dynamically redefined as being an instance of another type of data. The value object further includes methods that "flatten" and "unflatten" the data of the value object, that is, convert it from its memory representation in a computer memory to its transmission representation in a transmission medium, and back.
In accordance with another aspect of the present invention, there is provided a method of transmitting data from a first computer to a second computer on a communications network using a data arrangement that is semantically self describing, thereby allowing the data to be transferred without conventional marshalling operations that convert the data from the memory representation in the first computer to an intermediate transmission representation and then to the memory representation in the second computer.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is an illustration the data flow in a system for transmitting a data arrangement between a first and second computer in accordance with the present invention.
FIG. 2 is a schematic representation of a value object in accordance with the present invention.
FIG. 3a illustrates schematic representations of the storage arrangements for scalar type data elements in a value object.
FIG. 3b illustrates schematic representations of the storage arrangements for blob type data elements in a value object.
FIG. 3c illustrates a schematic representation of the storage arrangements for aggregate type data elements in a value object.
FIG. 4 is an illustration of the storage arrangement of an instance of an array in a value object.
FIG. 5 illustrates a schematic representation of the flattened data of various types in a value object, as suitable for transmission between computers.
FIG. 6 is an illustration of data flow in a preferred embodiment of a process for flattening and unflattening a value object.
FIG. 7 is a flowchart for dynamically altering the type of a data element.
DETAILED DESCRIPTION OF THE INVENTION
Referring now to FIG. 1, there is shown a schematic representation of a computer system 100 for transferring data between the memories 114 of two computers across transmission medium in a network 108. In a first computer memory 114a, a first application 102a or programming language environment constructs an instance of a value object 104 from an available class definition (unless otherwise noted, references to an instance of a class, and to the class itself will be made using the same reference number.) The application 102 initializes the value object 104a with some set of data elements 124. While illustrated outside of the value object 104a, the managed data elements 124 are stored "inside" or with the instance of the object.
FIG. 2 illustrates the structure of the data elements 124 held by a value object 104. The data is arranged in an internal memory representation in which each data element 124 of includes a type code 120 defining the type of the data, and a data member 122 that is dependent on the type code 120. In a preferred embodiment, the data member 122 itself may be the actual data for the data element 124, or a reference to another data element, thereby allowing hierarchical arrangements of data in the first computer memory 114a. The type code 120 may be fixed size value, such as a single byte, or greater or lesser number of bits, or it may be
Referring again to FIG. 1, at some point in time, the first application 102a initiates a transfer of the value object 104a to a second application 102b in a second computer memory 114b. This may be done to invoking a remote procedure of the second application 102b, for example. For the example of FIG. 1, there is provided a communication object 113a that handles remote procedures calls and communication between computers. The communication object 113 includes send and receive methods for this purpose. The communication object 113 is shown separately from the application 102 for clarity, but is typically included in it.
When invoked to transmit the value object 104a, the send method of the communication object 113a invokes a flatten method 110 of the value object 104a. In a preferred embodiment, the flatten method 110a is a method of the value object 104 class. Alternatively it may be a separate procedure. The flatten method 110a reads the data elements 124 of the value object 104a from the first computer memory 114a and constructs a byte stream 112 which is passed back to the communication object 113a. The arrangement preserves the semantic structure of the data elements 124 as stored in the memory. In particular, the arrangement in the byte stream 112 corresponds to a depth-first traversal of all hierarchical data elements 124 in the value object 104a, and duplicates the type codes 120 and data members 122 thereof. This transmission representation then is easily transportable to the memory 114b of a second computer on some transmission medium on the network 108. The transmission medium may be any conventional transmission substrate, such as wire, fiber optics, an infrared signal, a wireless signal, or variations of any of these.
A receiving application 102b in the memory 114b of the second computer collects the transmitted data stream 112, for example in a communication object 113b. The communication object 113b invokes an unflatten method 110b to unflatten the data stream 112 and restore the data elements 124 in another instance 104b of the value object. Again, the memory representation of this second instance of the value object 104b is identical in semantic structure to the data stream 112, and to the original value object 104a. The second application 102b is then able to operate on the second value object 104b in an ordinary manner.
Distinct from conventional data transfer methods, in this aspect of the present invention, no conventional marshalling routines are used to marshall the data of the value object 104a into the data stream 112 for transmission to the second computer. This is because, unlike conventional data arrangements as stored in memory which typically use implicit typing, in this aspect of the present invention, the type of each data element 124 is explicitly specified, both in the memory representation of the object, and in its transmission representation in the byte stream 112. In this manner, the applications programmer need only invoke remote procedure mechanism, or other transfer function, without having to write programming code to marshall the data into the transmission representation of the byte stream 112.
Referring now to FIGS. 3a, 3b, and 3c, there are shown schematic representations of the storage arrangements for various types of data elements 124 that a value object 104 can hold. The types of data elements 124 that may be held by a value object 104 may be usefully categorized into three groups:
1) scalar: a data element 124 with a fixed length;
2) blob: a data element 124 with a variable number of units of data of a single type as its data member 122; and,
3) aggregate: a data element 124 with a variable number of any arbitrary types of value objects 104 as its data member 122.
Referring particularly then to FIG. 3a, there are shown the data storage arrangement for scalar data elements 124. In a preferred embodiment, scalar data elements 124 may include Boolean, character, Unicode character, short or integer, long, float, double, and date, illustrated as data elements 124a to 124h, respectively. As generally described with respect to FIG. 2, scalar data elements have a type code 120 that identifies the scalar type, and a data member 122 defined by the type code 120. In FIG. 3a, the type codes 120 for Boolean, character, Unicode character, short, long, float, double, and date are illustrated as the integers 0 to 7, respectively. The other types described herein are illustrated with subsequent integers in FIGS. 3b and 3c. These type code 120 values will be used for discussion purposes, but are not intended to limit the variety of type codes 120 that may be used.
Each scalar data element 124 preferably has a printed representation, useful for outputting the data member, such as to a screen display or printer.
The printed representation Booleans, integers, floats, longs and doubles is conventional.
Character data elements 124b are preferably ASCII characters, represented by the numbers (-128 to 127), though other character sets (e.g. EBCIDIC) may be used. Unicode characters data elements 124c are 16 bit numbers in the range of (-32768 to 32767).
The printed representation of an ASCII character is:
The printed representation of a Unicode character is
where `xxxx` consists of four hexadecimal digits. The notation ".fwdarw." will be used throughout to indicate the printed representation of a value object 104 or other data.
A date data element 124h stores data defining a point in time. Dates are preferably maintained in memory as GMT times, and are then converted to local time by the application 114 or operating system.
The printed representation of a date is an array of the form:
where year is a four-digit year (e.g. 1982), month is a number in the range 1 . . . 12, and day is a number in the range 1 . . . 31, hours are in the range 0-23, and minutes and seconds are in the range of 0-59. Fractions of a second is a 32-bit number in nanoseconds. Alternatively, an is GMT boolean element may be added indicating whether timestamp is GMT or local time, and thereby eliminating the need to always convert to local time. One alternative implementation of a date data element 124h uses an array, and will be described below.
In one embodiment, scalar types do not include other data elements 124 at all. Accordingly, the memory representation of a scalar data element 124 is identical to its transmission arrangement. The flattener 110 need only to read out (or in, during unflattening) the type code 120 and data member 122 directly to (or from) the data stream 112.
Referring now to FIG. 3b, there are shown the data storage arrangements for blob type data elements 124. In a preferred embodiment blob type data elements 124 include string, Unicode string, and binary, illustrated as data elements 124i to 124k, respectively. Each blob data element 124 has a type code 120, and a data member 124 that references a descriptor 130. The descriptor 130 includes a length field 126, and a content field 126. The length field 126 specifies a number of units of data of a specific type.
Thus, a string data element 124i includes in the content field a number of characters. A Unicode string data element 124j includes in the content field 128 a number of Unicode characters.
A binary data element 124k includes in the content field 128 a number of uninterpreted bytes (or other size of data). Binary data elements 124 are useful for storing and transmitting audio data, video data, graphics, and the like. As will be discussed below, because a value object 104 may include, via an aggregate type data element 124, numerous different types of data elements 124 in a single object, binary data element 124, for example of a graphics file, can be directly stored intermixed with string data elements, (e.g. a text file) describing the binary data element 124, and a date data element 124 describing the creation date, modification date, or the like, of the graphics file. One alternative implementation of blob data elements 124k uses an array and will be described below.
As noted above, in the preferred embodiment, a blob data element 124 has as its data member 124 a reference to the descriptor 130. This is used in the memory representation of the value object 104 to access the parts of descriptor 130 in the local memory. Since the local memory reference is not valid when transferred to another computer, the flattener 110, when flattening a blob data element 124, obtains the descriptor 130 and directly builds it into the data stream 112 immediately following the type code 120 for the data element. This preserves the semantic structure of the memory representation and the transmission representation of the data element 124, and hence the value object 104 incorporating a blob data element 124.
Referring now to FIG. 3c, there are shown the data storage arrangements for aggregate type data elements 124. In a preferred embodiment aggregate data elements 124 include array and map, illustrated as data elements 124i to 124k, respectively. Each aggregate data element 124 has a type code 120, and a data member 122 that references a descriptor 130, as with blob data elements 124. The reference is preferably a pointer to the descriptor 130, though other forms of reference may be used, such as indices, handles, or the like.
For an array data element 124l, the descriptor 130d includes a length field 126, and a number of value objects 104, the number specified by the length field 126. Thus, array data element 124l defines a simple vector of values. Array data elements 124l are zero indexed. The value objects 104 contained by an array data element 124l may be of any type, including other array data elements 124l, or map data elements 124m.
The printed representation of an array is:
where each element is printed according to its own printed representation, as described above.
A map data element 124m is a collection of a number of <key 132, value object 104>pairs called map entries 136. Keys are preferably scalar type data elements, and are all of a same specific type. In other embodiments, varied types for keys may be used. The value objects 104 may be of any type, including arrays and maps. Accordingly, the descriptor 130e in a map data element 124m includes a length field 126 specifying the number of pairs of key value 132 and value objects 104. The map entries 136 are not guaranteed to be in any particular order. The printed representation of a map is:
where each key and value object is printed as defined above.
Because the aggregate data elements 124 include other value objects 104 in their data elements 124, they may be used to create hierarchical arrangements of data. The flattening and unflattening process preserves the hierarchical arrangement of data, and thereby the semantic structure of the value object 104 between its memory representation and its transmission representation.
Array data elements 124l and map data elements 124 may optionally include a class field 138. The class field 138 contains a programmer defined value associated with the array data element 124l or map data element 124m. This allows the programmer to define particular operations that manipulate arrays or maps according to their class field 138. When a class field 138 is defined, the provided value is also output in the printed representation of the data element.
FIG. 4 illustrates a schematic representation of an instance of an array data element 124l in a value object 104. Here, the array data element 124l is an array of class 138 "STUDENTS", and has three data elements, each of which is an array data element 124l with two string data elements 124i.
As noted above, in one embodiment, scalar data elements 124 do not include other data elements 124. In an alternative embodiment, scalar types remain fixed in length, but may include other types of data elements 124, or even value objects 104. An example of this would be a date data element 124h defined using a array data element 124l, having a predetermined number of entries, such as seven, one each for year, month, day, hour, minute, second, fraction of second. The number of elements would be stored in the length field 126, as before.
Similarly, blob type data elements may also be alternatively assembled from array data elements 124l. In such an embodiment, a binary data element 124k is an array data element 124l with an optional class field 138 value defining a user selected class value (e.g "movie", "graphic", "code"), and a length field 126 for the length of the binary data, and the data itself.
The type codes 120 illustrated in FIGS. 3a-3d are merely illustrative, in order to demonstrate the structure of the data elements 124 in a value object 104. Table 2 describes the type codes 120 used in a preferred implementation:
In Table 2, the GetType result is used merely to provide a correspondence with similar type codes used in Visual Basic. Note that Visual Basic does not fully recognize the distinctions in type between many of the data elements recognized in the present invention. The second column, Type Code Stored in Memory is the preferred type code 120 for the memory representation of a data element 124. The last column, Type Code for Transmission, is the preferred type code 120 for the transmission representation of a data element in a transmission medium.
In addition to being specified with particular number values, type codes 120 may also be ranges of values, or sets of values. That is, a predetermined set, or range of numbers, may be selected as all identifying a particular type of data element 124. For a simple example, all values between 0 and 9 may be selected as the type codes 120 for short data elements 124d, and values 10 to 19 may be selected as the type codes 120 for float data elements 124f. This is the inverse of the duplication of type codes 120 in the GetType column.
In one preferred embodiment, the type codes 120 for selected ones of the types of data elements 124 are determined according to a frequency distribution of the data values of the data elements. This allows the actual data values themselves to serve as both the type code 120 and the data member 122 for at least one type of data element 124. Having the data values of the data element 124 serve both functions reduces the amount memory required for storing (providing a form of data compression), and the time and bandwidth required for transmitting for these frequently used values and types of data element 124.
The third column of Table 2 illustrates this feature of the invention. For example, it may be determined that a particular set of long data values used in long data elements 124e are frequently stored and transmitted, for example, long data values between -64 and 63. This range of values is then selected as the type codes 120 for long data elements 124e. When a data element 124 is retrieved and the type code 120 is read, if the value of the type code 120 is in the specified range, then the flattener 110, or other code in an application 114, knows that the data element 124 is a long data element 124e, and decodes/encodes the rest accordingly, using that part of the data in the type code 120 space as value of the data member 122. If the value of the type code 120 is outside the range, then data element 124 are treated as described, with specific type codes 120. In this example a range of values is used, but logically, any set of values could be used for this optimization. This optimization of the type code 120 further increase the performance of the system for storing and transmitting data.
The domain over which the frequency determination is made may vary. For example, the frequency distribution may be determined with respect to any variety of applications 114. This type of determination would be useful, for example, when the value object 104 is provided as class in a general programming environment, and may be used by applications programmers to construct any variety of applications, including client-server applications of any type. In this type of embodiment, the selection of the type codes 120 would be made by the provider of the programming environment.
Alternatively, the domain may be limited to particular classes of applications, preferably based on empirical testing of the application behaviors and usage characteristics. Thus, for examples, it may be determined that word processors use string data elements 124i most frequently with the ASCII values. In this case, values between -128 and 127 are used as both the type code 120 and the data member 122 of the data element 122, again, reducing the memory and network requirements for this class of application. The provider of the programming development environment would provide a table that applications programmers can select from when creating their applications that would identify the type of application, and thus the type codes 120 for the data elements 124. In order to facilitate data transfer between applications of different types, each application would include a table with type codes for all other types of applications. When a connection is initially made between applications, they would exchange their application types, and then use that information to decode the value objects between transmitted between them.
Finally, in other alternative embodiments, the type codes 120 may be specified by the applications programmer herself, based on any desirable criteria.
It should be noted that the type codes 120 used in the transmission representation, as for example in Table 2, are then translated to memory representation type codes 120, and vice versa, by the flattening and unflattening methods.
As described above, the value object 104 is able to output its data elements 124 into "flattened" byte stream 112 for transmission over a network. The flattened byte stream 112 has identical semantics to the memory representation with the hierarchical arrangement of data. The flattener 110 is preferably implemented as a flatten method of the value object 104. A preferred implementation of the flatten method operates using recursive descent in order to unnest nested array data elements 124l and map data elements 125m. One pseudo code description of a flatten method is as follows:
The procedure emitNumber(value, destination.sub.13 buffer) outputs the byte of value to the destination buffer, which may be coupled to a network socket or other network connection. In the flatten method the values are the transmission representation type codes 120, and in the unflatten method they are memory representation type codes 120. The procedure emit bytes(number of bytes, source.sub.13 buffer, destination.sub.13 buffer) reads the number of bytes from the source buffer to the destination buffer. The first two cases perform the recursive descent through a hierarchical arrangement of data. The remaining cases structure the output bytes in the transmission representation, that is the actual byte ordering that is transmitted between two computers (or that may be restored locally). The procedure emitClass(class, buffer) outputs a class code (e.g. -50), the length of the class, and bytes from the buffer containing the actual class data.
The flatten method also puts the output in network standard order, as needed. The flatten method thereby creates the transmission representation of a given value object 104, with all hierarchical relationships between the data preserved. One preferred implementation of the flatten method is further described below, using an intermediate data structure, a flat value object, to hold the flattened data stream 112.
FIG. 5 illustrates schematically the flattened transmission representation of the various types of data elements 124 capable of being held by a value object 104. The type code 120 values used in FIG. 5 are the same as those used in FIGS. 3a-3c, and are merely illustrative. The placement of the class field 138 is not shown. If a class field 138 is used, then it preferably prefixes the data stream for the data element 124.
The value object 104 is likewise capable of reconstructing itself from the flattened data stream 112. An "unflatten" method (see the Get method, below, for one implementation) operates in a similar manner to the flatten method to perform this reconstruction. The unflatten method adds in each case a call to allocate a new instance of a data element 124 of the type defined by the type code 120, and then reads the data from a source buffer into the data element.
In addition to being able to hold a large number of different types of data, a value object 104 can dynamically (after instantiation) alter or change the type of any of its data elements. Thus, once a data element 124 in a value object 104 is created as a first type, for example, an short, it may subsequently changed to a different type of data. Examples of methods for these types of manipulations are further described below with respect to the Set methods.
A value object 104 supports a number of methods to get, set, create, test, remove, add, and otherwise manipulate the data elements 124 in the object. As explained above, value objects 104 may include aggregate types of data, such as array and maps, thereby allowing complex hierarchical arrangements of data elements 124. Accordingly, in the methods of the value object, each data element 124 in any aggregate type value object 124 is addressable using a path expression. A path is a sequence of array indices or map entries that specifies a "trip" from an outer value (such as an array or map) to some inner value. More particularly, each element E.sub.i in a path expression is an index into an array or map defined at element E.sub.i-1, where E.sub.0 is a top level array. Paths are most easily explained by examples.
For example, a value object v may be created with a five element array data element 124:
Here, the "5" is the path expression. The third element of the array may be set to some string:
The path in this call to a Set method is a one-element path "2". The final argument to Set, the string "Frogstar", is the value to set the element pointed to by the path. The Print method outputs the printable representation of the value object v. "nil" is the value a data member takes when it is initially created, and not specified.
A nested array data element 124l may be added:
This dynamically changes the third element to a array data element 124l with two elements.
The nested array by adding an index to the path:
The path "3, 1" in two calls to Set above start with the index of the inner array (3) and include an index into the inner array (0 or 1). The last argument to Set is always the value to set.
Paths are also used with map data elements 124m. A value object V including a map data element 125m, that itself containing an array data element 124l and another map data element 124m:
By specifying the key "a" as the first element of a path, further elements may be added to the nested array:
Likewise, by specifying the key "b" as the first path element, elements may be added to the nested map:
Another map may be added inside the already nested map:
The map entry q is printed by calling GetString with the path "b", "z" and "q":
As is evident, depending on the nesting, paths, and the hierarchical arrangement of data storage in value objects 104 may be arbitrarily complex,
As noted, the value object 104 supports a variety of different methods for manipulating the data elements contained therein. Table 3 through Table 12 define the interfaces (in Visual Basic syntax) of the preferred methods of the value object 104. Here path refers to path expressions previously described:
The operation of each of these methods is as follows:
(index)
value (index)
Parenthesis are used to get or set array elements and map entries.
If the first assignment to an uninitialized value object 104 is with parenthesis, the object will automatically turn into an array:
EXAMPLE
value | name
value | name
Visual Basic's "|" syntax can used to get or set map entries by name.
The value must be a map.
A map entry can be created (or set) with assignment:
Likewise, map entries can be retrieved by name:
Debug.Print v|FirstName, v|LastName.fwdarw.Fred Jones
If the first assignment to an uninitialized value object 104 is with "|" syntax, the value object 104 will automatically become a map.
Add
value.Add path,! value
This method adds the value to an aggregate. The path must refer to an aggregate value.
The exact behavior of Add depends on the kind of aggregate that the value is being added to:
EXAMPLE
AddFirst
value.AddFirst path,! value
This method prepends a value to an array. The path must refer to an array, the value can be any value.
This method can also be used to prepend a binary or string to another binary or string. Both values must be of the same type (binary or string).
EXAMPLE
AddLast
value.AddLast path,! value
This method appends a value to an array. The path must refer to an array, the value can be any value.
This method can also be used to append a binary or string to another binary or string. Both values must be of the same type (binary or string).
EXAMPLE
AddUnique
value.AddUnique path,! value
Appends a value to an array (or adds a value to a sorted array or set) if the value does not already appear. The path must refer to an array, the value can be any value.
EXAMPLE
AsGMT
date=value.AsGMT path!
The path must refer to a value that is a date. The date is returned as a GMT time.
EXAMPLE
Clone
value2=value.Clone path!
This method returns a copy of the specified value. The copy is deep; operations on the copy or any of its contents will not affect any elements of the original.
EXAMPLE
Count
count=value.Count path!
Returns the number of elements in an object. What the number means depends on the object's type:
This method is identical to the Length method.
EXAMPLE
Delete
value.Delete path,! index, count
If the value is an array, sorted array or set, the Delete method removes count elements starting at the offset index.
If the value is a string or Unicode string, the Delete method removes count characters starting at the offset index.
If the value is a binary, the Delete method removes count bytes starting at the offset index.
It is not possible to use Delete on any other value types, including symbols and maps.
EXAMPLE
Equals
boolean=value.Equals value2
This method returns true if the value is equal to the argument value, value2. The comparison is deep; the contents of arrays and other aggregates are compared recursively.
Strings are compared without regard to case.
EXAMPLE
ErrorText
string=value.ErrorText (errorNumber)
This method returns the text associated with an error returned by the application programming environment.
EXAMPLE
Exists
boolean=value.Exists path!
This method returns true if the specified value exists. Normally this method is used with a path.
This can be used to test if an entry exists in a map:
If myMap.Exists("fred") Then . . .
Or if an element exists in an array:
If myArray.Exists(42) Then . . .
The IsXxxx set of methods (IsArray, IsSet, IsLong, and the like) provide stronger versions of Exists that also check if the specified item is of a particular type.
EXAMPLE
Extract
result=value.Extract path,! index, count
The Extract method returns a specified piece of an array, sorted array, set, string, Unicode string or binary, depending on the type of the value:
EXAMPLE
Find
index=value.Find path,! expression
The value specified by the path must be an aggregate. The aggregate is searched to see if it contains the value expression. If it does, an index greater than or equal to zero is returned. If the aggregate does not contain the value, the index -1 is returned.
EXAMPLE
Flatten
flattener=value.Flatten
Referring to FIG. 6, the Flatten method makes a copy of the value and returns a flat value object 140, which holds the raw bytes of the original value object 104, and is capable of outputting those bytes as the data stream 112 from which the value object 104 can be reconstructed.
Once a flat value object 140 is constructed by the a call to the flatten method of a value object 104, a call to the Size method of the flat value object 140 returns the number of bytes in the flattened value. A Get method call on the flat value object 140, which takes an offset and a count, returns a byte array containing the appropriate bytes of all of the flattened data elements 124 in the value object 104. For example:
From the bytes, a copy of the original value object 104 can be reconstructed:
Here "v2" is the new value object 104, corresponding logically for example, to value object 104b in FIG. 1. The hierarchical arrangement of the data elements 124 in this second value object 104 is identical to that of the original value object 104, as shown in the final output line.
Notice that no marshalling calls are required in this code example. Thus, the applications programmer is able to easily transmit value objects 104 between computer systems.
Get
result=value.Get path!
Returns the value found by following the specified path.
If the value can be represented directly in Visual Basic, then a Visual Basic value of the appropriate type is returned:
However, if the value specified by the path is an aggregate then Get will return a Value object and you must use Visual Basic's Set syntax if you are performing assignment:
If a value holds a primitive type, such as a string or a number, the Get method without any arguments will retrieve the value:
GetClass
string=value.GetClass path!
Returns the class symbol for an array, map or binary. If the value does not have a class, NULL is returned.
EXAMPLE
GetElem
result=value.GetElem path,! index
Returns the value of the index'th entry in a map.
EXAMPLE
GetKey
result=value.GetKey path,! index
Returns the key of the index'th entry in a map.
EXAMPLE
GetString
string=value.GetString path!
Returns a printable representation of the value at the specified path, as a string. The string returned can be used for several purposes, such as a debugging aid. Alternatively, the string may be fed back into a value object, reconstituting the original value, with the SetString method. This is one of the means by which persistent values may be implemented.
EXAMPLE
GetType
typeNumber=value.GetType path!
Returns an integer indicating the type code 120 of the data element 124 specified by the path. The type code 120 is preferrably the memory representation type 120. However in to provide compatiblity with the Visual Basic programming environment which uses a set of small integers to define the available types, the following type codes are returned:
Note that Table 19 includes a variety of other types of data elements, such as symbol, sorted array, and integer, which as all specifies of the more general types of data elements 124 described above. The OLE object reference type is a data element that holds a reference to an OLE object. Again the type codes here are merely to correspond with Visual Basic types.
EXAMPLE
GetVBString
string=value.GetVBString path!
Returns a string representation of the value specified by the path. The string's characters are encoded in such a manner that it can be written to a file with a print statement, and read back in with an input statement.
EXAMPLE
This example first writes a value to a file using Visual Basic's GetVBString, then re-opens the file, reads from it, and uses SetVBString to reconstitute the value.
Hash
longHash=value.Hash path!
Returns a 32-bit hash of the value specified by the path.
EXAMPLE
Insert
value.Insert path,! index, count
If the value specified by the path is an array, this method inserts count nil values at the index. If the value specified by the path is a string or Unicode string, then count characters whose value is zero are inserted at the index. If the value specified by the path is a binary, then count bytes of zero are inserted at the index. No other value types can be modified with Insert.
EXAMPLE
InsertValue
value.InsertValue path,! index, value2
If the path refers to an array, the value argument (value2) is inserted before the index'th element. If the index is equal to the length of the array, the value is appended to the array.
Example
If the path refers to a string or Unicode string, the value argument (value2) should be another string. That string is inserted before the index'th character.
Example
If the path refers to a binary object, the value argument can be a Visual Basic byte array, a string, or another binary object. The bytes from the value argument are inserted before the specified index.
EXAMPLE
IsAggregate
boolean=value.IsAggregate path!
This method returns true if the value specified by the path is an array, sorted array, set or map.
EXAMPLE
boolean=value.IsXxxxx path!
These methods return true if the value specified by the path exists and is of the appropriate type.
EXAMPLE
IsNumber
boolean=value.IsNumber path!
This method returns true if the value specified by the path is any number (e.g. integer, long, float, or double).
EXAMPLE
Length
length=value.Length path!
This method returns:
The number of elements in an array, sorted array or set.
The number of elements in a map.
The number of characters in a string, Unicode string or symbol.
The number of bytes in a binary.
EXAMPLE
NewArray
arrayValue=factoryValue.NewArray (size , initialValue!!)
This is a factory method that creates a new array. If a size is present, the new array will have the specified number of elements, initialized to nil. If an initial value is present, the array elements are initialized to that value. The original value (used as the factory) is not affected.
EXAMPLE
NewBinary
binaryValue=factoryValue.NewBinary (contents, class)
EXAMPLE
valueobject=factoryValue.NewXxxxx (initialValue )
These factory methods use an existing value object to create another value object of a specific type, with an initial value. The existing value object is not affected.
Example
Example 2
One use of the factory methods is to declare a global, gv, like this:
Global gV As New WValue
Then a factory method is called on gV any time a new value object 104 is desired:
NewMap
mapValue=factoryValue.NewMap
This method uses an existing value object to create a new, empty map value object 104.
EXAMPLE
NewSet
setValue=factoryValue.NewSet
This method uses an existing value object 104 to create a new, empty set.
EXAMPLE
NewSorted
sortedValue=factoryValue.New Sorted
This method uses an existing value object 104 to create a new, empty sorted array.
EXAMPLE
NewValue
value=factoryValue.Newvalue (string)
This factory method uses an existing value object 104 to create a new one from a string. The string is the print-representation of a value (e.g. as produced by GetString).
EXAMPLE
Op "ArrayToMap"
Set value2=value.Op("ArrayToMap", path!)
This Op method requires that the path specify an array. The array is used to construct a new map; each array element is used as the key of a map entry. The value of each map entry is set to nil.
EXAMPLE
Op "GetPat"Op "GetNotPat"
Set mapValue=value.Op("GetPat", , path,! pattern)
Set mapValue=value.Op("GetNotPat", , path,! pattern)
These Op methods require that the path specify map. The pattern is a string. A new map is constructed, whose entries are copies the map entries whose keys start with (or that do not start with) the specified pattern.
EXAMPLE
Op "GetPID"
longvalue=value.Op("GetPID")
This Op method returns the process ID of the Value object in question. By comparing this to the result of GetPID on a Value object known to be created by the current process, this can be used to determine if another Value object was created in process.
EXAMPLE
Op "MapToArray"
Set arrayValue=value.Op("MapToArray", path!)
This Op method requires that the path specify a map. An array is returned that consists of the keys of all of the map's entries.
EXAMPLE
Op "MD5"
string=value.Op("MD5", path!)
This Op method computes the MD5 hash (a one-way secure hash) of the Value object in question. It returns a 32-character string of hexadecimal numbers that is the hash. All elements of the value object are taken into consideration; even a slight change to one element of, for example, a map or an array will result in a completely different hash value.
EXAMPLE
Op "Merge"
value.Op("Merge", path,! mapValue)
This Op method is used to merge maps. The path must refer to a map, which will be the destination of the merge. The mapValue argument must be a map; its map entries are copied into the destination map.
If a map entry exists in both maps, the map entry in the destination is overwritten.
EXAMPLE
Op "RegGet"
value.Op "RegGet", path,! rootHandle, registryPath, registryValueName
This Op method replaces the value's current value by reading the string contained in a registry entry. The registry entry is assumed to contain a printed value. The rootHandle is the longword of an open registry entry, or one of the well-known registry handles such as HKEY.sub.-- LOCAL.sub.-- MACHINE. The registryPath is the path to the registry entry in question. The registryValueName is the name of a value within the registry entry to get.If the registry entry does not exist, the value becomes nil.
EXAMPLE
Op "RegGetString"
value.Op "RegGetString", path,! rootHandle, registryPath, registryValueName
This Op method replaces the value's current value with the string contained in a registry entry. The roothandle is the longword of an open registry entry, or one of the well-known registry handles such as HKEY.sub.-- LOCAL.sub.-- MACHINE. The registryPath is the path to the registry entry in question. The registryValueName is the name of a value within the registry entry to get.
EXAMPLE
Op "RegSet"
value.Op "RegSet", path,! rootHandle, registryPath, registryValueName
This Op method prints (writes the string representation) of the value referred to by the path to the specified registry entry. If the value is nil, the registry entry is removed. The rootHandle is the longword of an open registry entry, or one of the well-known registry handles such as HKEY.sub.-- LOCAL.sub.-- MACHINE. The registryPath is the path to the registry entry in question. The registryValueName is the name of a value within the registry entry to get.
EXAMPLE
Op "RegSetString"
value.Op "RegSetString", path,! rootHandle, registryPath, registry ValueName
The path must refer to a value object 104 that is a string. This Op method writes the string referred to by the path to the specified registry entry. The rootHandle is the longword of an open registry entry, or one of the well-known registry handles such as HKEY.sub.-- LOCAL.sub.-- MACHINE. The registryPath is the path to the registry entry in question. The registryValueName is the name of a value within the registry entry to get.
EXAMPLE
Op "SortBy"
value.Op "SortBy", path,! sortOrder
The value specified by the path should be an array of arrays. These inner arrays are rearranged in the outer array in a specified sorting order. For example, assume a set of arrays describing people, account balances:
The arrays may be sorted by the person's last name by specifying that the second field be sorted on before any others:
The sortOrder parameter controls how the inner arrays are sorted. It is an array whose elements indicate the order in which to consider elements of the sub-arrays. In this case, the sort order:
. . NewValue("1!")
is used, indicating that the second field be the primary sort key. (The factory method NewValue is used to create a temporary sort order array).
Likewise, the account balance may be sorted:
By following a sort key with the string "r", the sort is performed in reverse order:
Op "StripPat"
Op "StripNotPat"
value.Op "StripPat", path,! pattern
value.Op "StripNotPat", path,! pattern
This method expects the value to be a map. The pattern is a string. All map elements whose keys start with (or that do not start with) the specified pattern are removed from the map.
EXAMPLE
Op"ToLower"
value.Op "ToLower", path!
The value referenced by the path must be a string. The contents of the string is converted to lowercase.
Example 1
Example 2
Remove
value.Remove path,! item
If the path specifies an array, the item must be an array index. The item'th array element is removed. If the path specifies a sorted array or a set, the first value that matches the item's value is removed. If the path specifies a map, item is the key of an element to remove.
EXAMPLE
RemoveAll
value.RemoveAll path!
The path must refer to an aggregate. All elements of the aggregate are removed.
EXAMPLE
RemoveFirst
result=value.RemoveFirst path!
The path must refer to an array, sorted array or set. The first element of the array is removed and returned.
EXAMPLE
RemoveLast
result=value.RemoveLast path!
The path must refer to an array, sorted array or set. The last element of the array is removed and returned.
EXAMPLE
RemoveValue
value.RemoveValue path,! valueToRemove
The path must refer to an aggregate. The first element or map entry with the value of valueToRemove is deleted.
EXAMPLE
Set
value.Set path,! newValue
As described above, the data elements of a value object 104 dynamically altered after the value object 104 has been instantiated and give values in its data elements 124. This dynamic alternation is provided by the Set methods. The value specified by the path is set to the new value. If the path refers to an array element or map entry that does not exist, the array is grown as necessary, or a map element is created.
EXAMPLE
FIG. 7 illustrates a flowchart for basic implementation of a set method. The method is called 700 on a particular value, or data element in a value object 104, as specified by a path expression. The method takes a new type code 120, and new value for the data element. The method determines 702 whether the data element 124 is an aggregate type. If so, the data member 120, including the descriptor 130 is deallocated 704. If not, the method tests 706 whether the value is a blob type. Again, the data member and descriptor is deallocated 708 if necessary. In either of these cases, the method then sets 710 the type code 120 to the new type code 120. If the type code is an aggregate (712), then an aggregate descriptor 130 is allocated, and the descriptor of the new value is copied 714 to the body of the new descriptor 130. If the new type is blob type (716), then a blob descriptor 130 is allocated, and the descriptor 130 of the data value is copied 718. In either case, the actual data values are set 720 to the values in the new data value that have been passed to the method.
SetArray
value.SetArray path,! size!
This causes the value referenced by the path to become an array of the specified size. The array's element's initial values are nil. If the path is empty, the Value object receiving this method becomes an empty array.
An array of a particular size can be created by specifying the size:
Likewise, using a path, a sub-array can be created as one of the outer array's elements:
SetBinary
value.SetBinary path,! contents, classSymbol
This method causes the value specified by the path to become a binary. The initial contents may be a Visual Basic string, or a byte array. The class symbol is a string, or a value object that is a symbol or nil.
EXAMPLE
SetClass
value.SetClass path,! symbol
This method allows you to set the class symbol of an array, map or binary. The class symbols of sorted arrays and sets cannot be changed.
EXAMPLE
SetGMT
value.SetGMT path,! date
This method allows the applications programmer to set a time value that has a specific GMT time, rather than a local time. Normally, the Set method converts any date value to GMT internally. This method permits the programmer to bypass this conversion. The corresponding method, AsGMT, allows the programmer to access a time value as GMT directly.
EXAMPLE
SetLength
value.SetLength path,! length
This method is used to set the length of a resizable value. The path must refer to an array, sorted array, set, string, Unicode string or binary. If the new length is smaller than the old length, values at the end of the array, sorted array or set are simply released. If the new length is larger than the old length:
for arrays, nils are added at the end of the array.
for strings, Unicode strings and binaries, zeros are added to the end of the value.
sets and sorted arrays cannot be grown.
EXAMPLE
SetMap
value.SetMap path!
The value specified by the path is turned into an empty map. The previous contents of the value object 104 are lost.
EXAMPLE
SetNil
value.SetNil path!
The value specified by the path is set to nil. The previous contents of the value object 104 are lost.
EXAMPLE
SetSet
value.SetSet path!
The value specified by the path is turned into an empty set. The previous contents of the value object 104 are lost.
EXAMPLE
SetSorted
value.SetSorted path!
The value specified by the path is turned into an empty sorted array. The previous contents of the value object 104 are lost.
EXAMPLE
SetString
value.SetString path,! string
This method sets the value specified by the path to the string's interpreted value. Using the SetString method is an easy way to initialize values.
EXAMPLE
SetSymbol
value.SetSymbol path,! string
The value specified by the path is set to a symbol. The string argument should a string.
EXAMPLE
SetUnicodeChar
value.SetUnicodeChar path,! number
The value specified by the path is set to a Unicode character.
EXAMPLE
SetUnicodeString
value.SetUnicodeString path,! string
The value specified by the path is set to a Unicode string.
EXAMPLE
SetVBString
value.SetVBString path,! string
This method is used in conjunction with the result of the strings returned by the method GetVBString. The SetVBString method accepts a string formatted from a value by GetVBString and reconstitutes the value. The only difference between the methods GetString/SetString and GetVBString/SetVBString is that the latter set of methods escape characters (such as quotation marks) in a manner that is friendly to Visual Basic's Input# statement.
EXAMPLE
Sort
value.Sort path
The path must refer to an array. The elements of the array are sorted in ascending order. If the array contains values of different types (e.g., strings mixed with numbers), then numbers are sorted before strings. If the array contains any sub-arrays, the sub-arrays are ordered in the outer array in increasing order according to the elements that the sub-arrays contain.
EXAMPLE
Version
string=value.Version
This method returns a string indicating a version level of the value object 104, as set by the provider of the application programming environment containing the value object 104 class.
EXAMPLE
In summary, the present invention provides a simple but powerful tool for working with data structures. A value object can contain a rich variety of data types and data elements, including text, tables, numbers, binary objects, arrays, maps, sets and the like. Value objects provide a single, convenient representation for data that works well as a local general-purpose data structure, and that is also easily transported from one computer to another, without the need to use conventional marshalling routines. Value objects support a rich set of data types that allows information to be expressed simply and naturally.
When value objects are transmitted or stored, their representation is relatively compact. The semantic structure of the value object is preserved between its memory representation in a computer memory and its transmission representation in a transmission medium. Value objects are self-describing, making it easy to design message formats that are flexible and extensible. There is no inherent need for message format "version numbers," as in conventional messaging systems. Value objects can easily be made persistent by storing them in database tables or text files. This makes store-and-forward services particularly easy to design and implement.
Claims
We claim:
1. A computer readable memory having a plurality of contiguous storage locations and storing an object including: at least one data element each data element comprising: a) in at least one first storage location a type code defining a type of a data for the data element, the type being one of a group comprising: a first type of data element having a fixed length, the first type of data associated with first type code; a second type of data element having a variable length but not including other data elements, the second type of data associated with a second type code; and, a third type of data element having a variable length and capable of including other data elements, the third type of data associated with a third type code; and, b) in at least one second storage location contiguous with the at least one first storage location, a data member selected from a group comprising: at least one data value of the data element; or, at least one reference to another data element.
2. The computer readable memory of claim 1, wherein the type code for at least one type of data element is determined according to a frequency distribution of the types of selected data elements in a sample of data.
3. The computer readable memory of claim 1, wherein, the type code for at least one type of data element is the data value of the data element when the data value is within a specified range of data values.
4. The computer readable memory of claim 1, wherein: the data member is determined according to the selected type code, and being one of a group comprising: a fixed number of storage locations where the type code is the first type code; a length value specifying a number of units in the data member, and the number of units, each unit having a fixed number of storage locations, where the type code is the second type code; and, a length value specifying a number of units in the data member, and the number of units, each unit being a data element of any of the types of the data elements, where the type code is the third type code.
5. The computer readable memory of claim 1, wherein the data member is selected from a group further comprising: a reference to a descriptor including a length field, and a number of contiguous storage locations holding data values, the number being specified in the length field.
6. The computer readable memory of claim 1, wherein the data member is selected from a group further comprising: a reference to a descriptor including a length field, and a number of contiguous sets of storage locations, each set of storage locations including a key value, and a data element, the number of sets being specified in the length field, such that any data element may be retrieved according to its key value.
7. A computer readable memory having a plurality of contiguous storage locations and storing an object including: at least one data element, each data element comprising: a) in at least one first storage location a type code defining a type of a data for the data element, the type being one of a group comprising: a first type of data element having a fixed length, the first type of data associated with a first type code; a second type of data element having a variable length but not including other data elements, the second type of data associated with a second type code; and, third type of data element having a variable length and capable of including other data elements, the third type of data associated with a third type code; and, b) in at least one second storage location contiguous with the at least one first storage location, a data member selected from a group comprising: at least one data value of the data element; or, at least one reference to another data element; and a plurality of instructions for controlling a computer to convert the object from a memory representation using memory dependent locations to a transmission representation using memory independent data bytes.
8. The computer readable memory of claim 7, wherein the plurality of instructions reads the object from the memory, and recursively descends through each data element including a reference to another data element, and converts each such reference from a data element to a contiguous sequence of data bytes for the type code and data member of the data element, the contiguous sequence of data bytes adapted for transmission to a second computer readable memory.
9. A computer readable memory having a plurality of contiguous storage locations and storing an object including: at least one data element, each data element comprising: a) in at least one first storage location a type code defining a type of a data for the data element; and, b) in at least one second storage location contiguous with the at least one first storage location, a data member selected from a group comprising: at least one data value of the data element; or at least one reference to another data element; and a first procedure to retrieve any selected data element within the object; a second procedure to remove any selected data element within the object; a third procedure to add a new data element at a designated location within the object; at least one fourth procedure to determine whether a selected data element within the object is of a selected type; at least one fifth procedure to change the type code for any selected data element within the object to a selected type code; and at least one sixth procedure to replace the data member for any selected data element within the object with an input data member.
10. A computer readable arrangement of data in a transmission medium, the arrangement of data having a plurality of contiguous memory independent data bytes that form an object including: at least one data element, each data element comprising: a) in at least one first data byte a type code defining a type of data for the data element, the type code selected from a group comprising: a first type code for a first type of data element having a fixed length; a second type code for a second type of data element having a variable length but not including other data elements; and, a third type code for a third type of data element having a variable length and capable of including other data elements; b) in a plurality of second data bytes contiguous with the at least one first data byte, a data member determined according to the selected type code, and being one of a group comprising: a fixed number of data bytes where the type code is the first type code; a length value specifying a number of units in the data member, and the number of units, each unit having a fixed number of data bytes, where the type code is the second type code; and, a length value specifying a number of units in the data member, and the number of units, each unit being a data element of any of the types of the data elements, where the type code is the third type code.
11. A computer readable memory coupled to a transmission medium to receive a computer readable arrangement of data, the computer readable memory including a plurality of instructions to read the data and create in memory an object by storing in contiguous storage locations: data bytes of a first type code; data bytes of data members for data elements of a first type corresponding to the first type code; data bytes of a second type code and a reference to data members for data elements of a second type corresponding to the second type code; and data bytes of a third type code; and a reference to data members for data elements of a third type corresponding to the third type code.
12. A computer implemented method of transferring data between a first and second computer, comprising: storing in a memory of the first computer an object including: at least one data element, each data element comprising: a) in at least one first storage location a type code defining a type of a data for the data element; and, b) in at least one second storage location contiguous with the at least one first storage location, a data member selected from a group comprising: at least one data value of the data element; or at least one reference to another data element; flattening the object to produce a memory location independent byte stream representation of the object; transferring the byte stream representation from the first computer to a memory of the second computer; and, reconstructing in the memory of the second computer the object from the byte stream representation.
13. A computer system, comprising: a network providing a transmission medium; a first computer having a first computer memory having a plurality of contiguous storage locations and storing a first object having a first structure comprising: at least one data element, each data element comprising: a) in at least one first storage location a type code defining a type of a data for the data element, the type selected from a group comprising: a first type of data element having a fixed length, the first type of data associated with a first type code; a second type of data element having a variable length but not including other data elements, the second type of data associated with a second type code; and, a third type of data element having a variable length and capable of including other data elements, the third type of data associated with a third type code; and b) in at least one second storage location contiguous with the at least one first storage location, a data member selected from a group comprising: at least one data value of the data element; or, at least one reference to another data element; wherein the first computer is communicatively coupled to the network, and converts the object from the first structure to a computer readable arrangement of data comprising: at least one data element, each data element comprising: a) in at least one first data byte a type code defining a type of a data for the data element, the type code selected from a group comprising: a first type code for a first type of data element having a fixed length; a second type code for a second type of data element having a variable length but not including other data elements; and, a third type code for a third type of data element having a variable length and capable of including other data elements; b) in a plurality of second data bytes contiguous with the at least one first data byte, a data member determined according to the selected type code, and being one of a group comprising: a fixed number of data bytes where the type code is the first type code; a length value specifying a number of units in the data member, and the number of units, each unit having a fixed number of data bytes, where the type code is the second type code; and, a length value specifying a number of units in the data member, and the number of units, each unit being a data element of any of the types of the data elements, where the type code is the third type code; and a second computer having a second computer memory, and communicatively coupled to the network to receive the computer readable data arrangement, and to convert it to a second object semantically identical to the first object, and having the first structure.
Patent Citations (19)
| Patent | Date | Inventor | Cited By |
|---|---|---|---|
| US4575827 | 1986-03-01 | Kulakowski | |
| US4893232 | 1990-01-01 | Shimaoka et al. | |
| US5034914 | 1991-07-01 | Osterlund | |
| US5129076 | 1992-07-01 | Freeman et al. | |
| US5129084 | 1992-07-01 | Kelly, Jr. et al. | |
| US5165023 | 1992-11-01 | Gifford | |
| US5212786 | 1993-05-01 | Sathi | |
| US5247646 | 1993-09-01 | Osterlund et al. | |
| US5261080 | 1993-11-01 | Khoyi et al. | |
| US5338923 | 1994-08-01 | Grieu | |
| US5421012 | 1995-05-01 | Khoyi et al. | |
| US5426747 | 1995-06-01 | Weinred et al. | |
| US5432928 | 1995-07-01 | Sherman | |
| US5442771 | 1995-08-01 | Fillep et al. | |
| US5479656 | 1995-12-01 | Rawlings, III | |
| US5488717 | 1996-01-01 | Gibson et al. | |
| US5551020 | 1996-08-01 | Flax et al. | |
| US5561785 | 1996-10-01 | Blandy et al. | |
| US5592342 | 1997-01-01 | Hall et al. |