Method and system for generating program source code of a computer application from an information model

• US 7,032,210 B2
• Filed: 05/02/2002
• Issued: 04/18/2006
• Est. Priority Date: 11/11/2001
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Foucher, Marie, Alloing, Francois, Macquet, Xavier
• Primary Examiner(s): Dam, Tuan
• Assistant Examiner(s): CHOW, CHIH CHING
• Application Number:
  US10/138,892
• Publication Number:
  US 20030110472A1
• Time in Patent Office: 1,447 Days
• Field of Search: 717/106, 717/108, 717/147, 717/105, 717/109
• US Class Current: 717/106
• CPC Class Codes: G06F 8/71 Version control security arrangements therefor G06F21/57; Configuration management

Abstract

A method for generating program source code of a computer application from an information model representing a high level specification of the computer application, comprises a step of inserting reconciliation annotations in the program source code, the reconciliation annotations representing a modification rule and a modification state for each language element contained in the generated source code. Then, upon receiving a current version of the program source code, the current version resulting from the modification by a user of an old generated version of the program source code, and a new generated version of the program source code, the new version reflecting changes in the application information model, a step of reconciling the current version with the new version of the program source code according to the reconciliation annotations inserted in each of the current and new program source code versions, and a step of generating a reconciled version of the program source code are performed.

Claims

1. A method in a computer system for generating program source code of a computer application from an information model representing a high level specification of the computer application, said method comprising the computer-implemented steps of:

inserting reconciliation annotations in said program source code, said reconciliation annotations representing at least a modification rule and a modification state for each language element contained in said generated source code;

upon receiving a current version of said program source code, said current version resulting from the modification by a user of an old generated version of said program source code, and a new generated version of said program source code, said new version reflecting changes in the application information model, reconciling the current version with the new version of the program source code according to the reconciliation annotations inserted in each of the current and new program source code versions; and

generating a reconciled version of the program source code, whereby user modifications in an old version of the program source code and subsequent changes in the application information model are automatically consolidated in a reconciled version of the program source code.

2. A method according to claim 1, wherein each of the program source code versions is represented by a program tree in a memory of the computer system, the program tree being a tree data stnicture having nodes, each node representing a language element of the program source code.

3. A method according to claim 2, wherein the step of inserting reconciliation annotations in a generated program source code comprises the step of creating in the program source code, for each node of the program tree representing the program source code, a program construct having a plurality of data fields, said plurality of data fields including:

a node identification field containing a node identifier for identifying the semantic type and the name of the source code language element corresponding to the free node;

a modification rule identification field for containing an identifier of the modification rule applicable to the corresponding language element; and

a modification state identification field for containing an identifier indicative of the modification state of the corresponding language element.

4. A method according to claim 3, wherein said plurality of data fields further includes:

a model entity identification field for containing an identifier of an entity of the information model, on which depends the corresponding language element, said entity identifier being invariable all through the code generation process; and

a model entity name identification field for containing the name of the information model entity on which depends the corresponding language element, said model entity and identification fields being used in conjunction to a predefined renaming rule for implementing a renaming procedure.

5. A method according to claim 4, wherein the modification rule identified by the value of said modification rule identification field, has one of the following meanings:

the language element to which it applies is not allowed to be modified (Not Updateable—

NU);

the language element to which it applies is not allowed to be deleted (Not Deleteable—

ND);

the language element to which it applies is not allowed to be modified nor deleted (Not Updateable nor Deleteable—

NUD);

or

the language element to which it applies can be modified or deleted (UD).

6. A method according to claim 5, wherein the modification state identified by the value of said modification state identification field, has one of the following meanings:

the language element to which it applies has not been modified (Unmodified);

the language element to which it applies has been modified (Modified);

or

the language element to which it applies has been deleted (Deleted).

7. A method according to claim 6, wherein said reconciliation step between said current version and said new version of the program source code, comprises the sub-steps of:

for each node (current node) of the current version free, reading the reconciliation annotations of the current node and those of the corresponding node (new node), if it exists, of the new version tree; and

determining a reconciled node according to the modification state of the current node and either to the modification rule of the corresponding new node, if it exists, or otherwise, to the modification rule of the current node.

8. A method according to claim 7, wherein, in order to generate the reconciled version of the program source code, each reconciled node is detemilned by applying a number of predefined reconciliation rules which include at least the following rules:

if a current node is unmodified and a corresponding new node exists, then the new node is to be used in the reconciled version;

if a current node is unmodified and a corresponding new node does not exist, then the current node is to be omitted in the reconciled version;

if a current node has been modified and a corresponding new node exists, and the modification rule is “

Not-updateable-nor-Deleteable”

or “

Not-Updateable”

, then the new node is to be used in the reconciled version;

if a current node has been modified and a corresponding new node exists, and the modification rule is “

Not-Deleteable”

or “

Updateable-and-Deleteable”

, then the current node is to be used in the reconciled version;

if a current node has been modified and a corresponding new node does not exist, and the modification rule of the current node is “

Not-Updateable-nor-Deleteable”

or “

Not-Deteteable”

, then the current node is to be omitted in the reconciled version;

if a current node has been modified and a corresponding new node does not exist, and the modification rule of the current node is “

Not-Updateable”

or “

Updateable-and-Deleteable”

, then the current node is to be used in the reconciled version;

if a current node has been deleted and a corresponding new node exists, and the modification rule is “

Not-Updateable-nor-Deleteable”

or “

Not-Deleteable”

, then the new node is to be used in the reconciled version;

if a current node has been deleted and a corresponding new node exists, and the modification rule is “

Updateable-and-Deleteable”

or “

Not-Updateable”

, then the new node is to be omitted in the reconciled version;

if a current node has been deleted and a corresponding new node does not exist, then none of them is to be used in the reconciled version;

if a current node has been added and a corresponding new node exists, and the modification rule is “

Not-Updateable-nor-Deleteable”

or “

Not-Updateable”

, then the new node is to be used in the reconciled version;

if a current node has been added and a corresponding new node exists, and the modification rule is “

Updateable-and-Deleteable”

or “

Not-Deleteable”

, then the current node is to be used in the reconciled version; and

if a current node has been added and a corresponding, new node does not exist, then the current node is to be used in the reconciled version.

9. A method according to claim 8, wherein the reconciliation step includes a renaming procedure which relics on the information contained in said reconciliation annotations, within said model entity identification and name fields, and on a predefined renaming rule, said renaming procedure comprising the sub-steps of:

extracting from the program source code current version tree (“

current tree”

), for each current node, the information contained in said model entity identification and name fields;

parsing the program source code new version tree (“

new tree”

); and

for each current node, if it is determined that the corresponding information model entity has been renamed, according to the content of the model entity name field associated with the corresponding new node, then renaming the current node, according to the renaming rule.

10. A method according to claim 9, wherein the reconciliation step is preceded by a step of validating, in said current version of said program source code, the modifications made by the user to said old generated version of the program source code, said validation step comprising, for each modification found into said current version, the sub-steps of:

reading the modification rule annotation in the old version, which corresponds to the same tree node;

canceling any modification in the current version that violates the corresponding modification rule in the old version, or otherwise; and

allowing the modification and updating accordingly the reconciliation annotations contained in the current version, in particular the modification state annotation.

11. A method according to claim 10, wherein a program source code is expressed in Java programming language and the corresponding program tree is structured according to the JDOM (Java Document Object Model) standardized format.

12. A method according to claim 11, wherein said program construct having a plurality of fields containing reconciliation information, is implemented as a specific Java method.

13. A method according to claim 10, wherein a program source code is expressed in XML or HTML programming language and the corresponding program tree is structured according to the DOM (Document Object Model) standardized format.

14. A system for generating program source code of a computer application by carrying out a method as claimed in any one of the preceding claims, said computer system comprising:

insertion means for inserting reconciliation annotations in said program source code during the generation process, said reconciliation annotations representing at least a modification rule and a modification state for eaeh language element contained in said generated source code; and

a reconciliation component including;

means for receiving a current version of said program source code, said current version resulting from the modification by a user of an old generated version of said program source code, and a new generated version of said program source code, said new version reflecting changes in the application information model; and

means for reconciling said current version with said new version of the program source code according to the reconciliation annotations inserted in each of the current and new program source code versions; and

means for generating a reconciled version of the program source code;

whereby user modifications in an old version of the program source code and subsequent changes in the application information model are automatically consolidated in a reconciled version of the program source code.

15. A system for generating program source code of a computer application comprising:

insertion means for inserting reconciliation annotations in said program source code during a generation process, said reconciliation annotations representing at least a modification rule and a modification state for each language element contained in said generated source code; and

a reconciliation component including;

means for receiving a current version of said program source code, said current version resulting from the modification by a user of an old generated version of said program source code, and a new generated version of said program source code, said new version reflecting changes in the application information model;

means for reconciling said current version with said new version of the program source code according to the reconciliation annotations inserted in each of the current and new program source code versions;

means for generating a reconciled version of the program source code, whereby user modifications in an old version of the program source code and subsequent changes in the application information model are automatically consolidated in a reconciled version of the program source code;

a core engine containing information required for calling program functions which implement the reconciliation procedure, a renaming procedure, and a validation procedure, the core engine being independent of the programming language used for the program source code and independent of a corresponding dedicated tree structure format;

a first interface containing at least information for declaring the methods which make it possible to actually run the reconciliation, the renaming or the validation procedure on language-specific Document Object Model (DOM) trees;

a second interface containing information for declaring methods which enable the consulting of the information items associated to a given tree node, information for declaring a method which enables the detection of a user modification during the course of the validation procedure, and information for declaring a method which enables the processing of a node during the renaming procedure; and

an extension module containing a number of extension units, each extension unit being dedicated to one particular programming language supported by said program source code generation system and its corresponding tree structure format, each extension unit containing two programming language-specific classes which implement respectively said first and second interfaces.

16. A computer program in a computer readable medium, said program comprising program instructions adapted to implement a method comprising the steps of:

inserting reconciliation annotations in said program source code, said reconciliation annotations representing at least a modification nile and a modification state for each language element contained in said generated source code;

upon receiving a current version of said program source code, said current version resulting from the modification by a user of an old generated version of said program source code, and a new generated version of said program source code, said new version reflecting changes in the application information model, reconciling the current version with the new version of the program source code according to the reconciliation annotations inserted in each of the current and new program source code versions; and

generating a reconciled version of the program source code, whereby user modifications in an old version of the program source code and subsequent changes in the application information model are automatically consolidated in a reconciled version of the program source code.

17. A productaccording to claim 16, wherein each of the program source code versions is represented by a program tree in a memory of the computer system, the program tree being a tree data structure having nodes, each node representing a language element of the program source code.

18. A product according to claim 17, wherein the step of inserting reconciliation annotations in a generated program source code comprises the step of creating in the program source code, for each node of the program tree representing the program source code, a program construct having a plurality of data fields, said plurality of data fields including:

a node identification field containing a node identifier for identifying the semantic type and the name of the source code language element corresponding to the tree node;

a modification rule identification field for containing an identifier of the modification rule applicable to the corresponding language element; and

a modification state identification field for containing an identifier indicative of the modification state of the corresponding language element.

19. A product according to claim 18, wherein said plurality of data fields further includes:

a model entity identification field for containing an identifier of an entity of the information model, on which depends the corresponding language element, said entity identifier being invariable all through the code generation process; and

a model entity name identification field for containing the name of the information model entity on which depends the corresponding language element, said model entity and identification fields being used in conjunction to a predefined renaming rule for implementing a renaming procedure.

20. A product according to claim 19, wherein the modification rule identified by the, value of said modification rule identification field, has one of the following meanings:

the language element to which it applies is not allowed to be modified (Not Updateable—

NU);

the language element to which it applies is not allowed to be deleted (Not Deleteable—

ND);

the language element to which it applies is not allowed to be modified nor deleted (Not Updateable nor Deleteable—

NUD);

or

the language element to which it applies can be modified or deleted (UD).

21. A product according to claim 20, wherein the modification state identified by the value of said modification state identification field, has one of the following meanings:

the language element to which it applies has not been modified (Unmodified);

the language element to which it applies has been modified (Modified);

or

the language element to which it applies has been deleted (Deleted).

22. A product according to claim 21, wherein said reconciliation step between said current version and said new version of the program source code, comprises the sub-steps of:

for each node (current node) of the current version tree, reading the reconciliation annotations of the current node and those of the corresponding node (new node), if it exists, of the new version tree; and

determining a reconciled node according to the modification state of the current node and either to the modification rule of the corresponding new node, if it exists, or otherwise, to the modification rule of the current node.

23. A product according to claim 22, wherein, in order to generate the reconciled version of the program source code, each reconciled node is determined by applying a number of predefined reconciliation rules which include at least the following rules:

if a current node is unmodified and a corresponding new node exists, then the new node is to be used in the reconciled version;

if a current node is unmodified and a corresponding new node does not exist, then the current node is to be omitted in the reconciled version;

if a current node has been modified and a corresponding new node exists, and the modification rule is “

Not-Updateable-nor-Deleteable”

or “

Not-Updateable”

, then the new node is to be used in the reconciled version;

if a current node has been naodilied and a corresponding new node exists, and the modification rule is “

Not-Deleteable”

or “

Updateable-and-Deleteable”

, then the current node is to be used in the reconciled version;

if a current node has been modified and a corresponding new node does not exist, and the modification rule of the current node is “

Not-Updateable-nor-Deleteable”

or “

Not-Deleteable”

, then the current node is to be omitted in the reconciled version;

if a current node has been modified and a corresponding new node does not exist, and the modification rule of the current node is “

Not-Updateable”

or “

Up dateable-and-Deleteable”

, then the current node is to be used in the reconciled version;

if a current node has been deleted and a corresponding new node exists, and the modification rule is “

Not-Updateable-nor-Deleteable”

or “

Not-Deleteable”

, then the new node is to be used in the reconciled version;

if a current node has been deleted and a corresponding new node exists, and the modification rule is “

Updateable-and-Deleteable”

or “

Not-Updateable”

, then the new node is to be omitted in the reconciled version;

if a current node has been deleted and a corresponding new node does not exist, then none of them is to be used in the reconciled version;

if a current node has been added and a corresponding new node exists, and the modification rule is “

Not-Updateable-nor-Deleteable”

or “

Not-Updateable”

, then the new node is to be used in the reconciled version;

if a current node has been added and a corresponding new node exists, and the modification rule is “

Updateable-and-Deleteable”

or “

Not-Deleteable”

, then the current node is to be used in the reconciled version; and

if a current node has been added and a corresponding new node does not exist, then the current node is to be used in the reconciled version.

24. A product according to claim 23, wherein the reconciliation step includes a renaming procedure which relies on the information contained in said reconciliation annotations, within said model entity identification and name fields, and on a predefined renaming rule, said renaming procedure comprising the sub-steps of:

extracting from the program source code current version tree (“

current free”

), for each current node, the information contained in said model entity identification and name fields;

parsing the program source code new version tree (“

new tree”

); and

for each current node, if it is determined that the corresponding information model entity has been renamed, according to the content of the model entity name field associated with the corresponding new node, then renaming the current node, according to the renaming rule.

25. A product according to claim 24, wherein the reconciliation step is preceded by a step of validating, in said current version of said program source code, the modifications made by the user to said old generated version of the program source code, said validation step comprising, for each modification found into said current version, the sub-steps of:

reading the modification nile annotation in the old version, which corresponds to the same tree node;

cancelling any modification in the current version that violates the corresponding modification rule in the old version, or otherwise; and

allowing the modification and updating accordingly the reconciliation annotations contained in the current version, in particular the modification state annotation.

26. A product according to claim 25, wherein a program source code is expressed in Java programming language and the corresponding program tree is structured according to the IDOM (Java Document Object Model) standardized format.

27. A product according to claim 26, wherein said program construct having a plurality of fields containing reconciliation information, is implemented as a specific Java method.

28. A product according to claim 25, wherein a program source code is expressed in XML or HTML programming language and the corresponding program tree is structured according to the DOM (Document Object Model) standardized format.