Generating Program Fragments Using Keywords and Context Information

• US 20140282375A1
• Filed: 06/28/2013
• Published: 09/18/2014
• Est. Priority Date: 03/15/2013
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Le, Vu Minh, Su, Zhendong, Gulwani, Sumit, de Halleux, Jonathan Paul, Tillmann, Nikolai
• Granted Patent: US 9,448,772 B2
• Time in Patent Office: Days
• US Class Current: 717/107
• CPC Class Codes: G06F 8/36 Software reuse

Abstract

A program development framework (PDF) is described herein which allows a user to produce a program in piecemeal fashion by successively specifying program fragments. The PDF creates a new program fragment by receiving keyword information from the user that describes a new program fragment, and then identifies context information that pertains to a programmatic context in which the new program fragment appears within the overall program being created. The PDF then generates a set of candidate program fragments that satisfy the keyword information and the context information, and ranks those candidate program fragments based on ranking information. At least part of the ranking information may be based on statistical information that is produced by analyzing a corpus of previous programs produced by one or more users. The PDF then provides the ranked program fragments to the user using various user-friendly presentation strategies.

Claims

1. A method, performed by one or more computing devices, for creating a program, comprising:

receiving one or more keywords from a user who wishes to generate a program fragment as part of a program being created, to thereby provide keyword information;

identifying a context in which the user is generating the program fragment, to provide context information;

generating a set of candidate program fragments based on at least the keyword information and the context information;

ranking the candidate program fragments based on ranking information, to provide a set of ranked program fragments; and

presenting an output presentation to the user based on the set of ranked program fragments,

said receiving, identifying, generating, ranking, and presenting, being performed by said one or more computing devices.

2. The method of claim 1, wherein said receiving comprises receiving said one or more keywords in response to the user invoking a search function.

3. The method of claim 1, wherein said receiving comprises receiving said one or more keywords in response to an editing action performed by the user.

4. The method of claim 1, wherein said identifying comprises identifying at least one global variable which affects the program fragment, said at least one global variable having been previously defined in the program being created.

5. The method of claim 1, wherein said identifying comprises identifying at least one local variable which affects the program fragment, said at least one local variable having been previously defined in the program being created.

6. The method of claim 1, wherein said identifying comprises identifying at least one function which affects the program fragment, said at least one function having been previously defined in the program being created.

7. The method of claim 1, wherein the ranking information is based, at least in part, on statistical information that is produced by analyzing a corpus of programs created by one or more users, each program in the corpus being composed of one or more program elements.

8. The method of claim 7, wherein the statistical information includes:

first statistical information that corresponds to previous programs created by a particular user who is creating a current program; and

second statistical information that corresponds to previous programs created by a plurality of users.

9. The method of claim 7, wherein the statistical information includes frequency information, the frequency information identifying, for at least one program element, a frequency at which said at least one program element appears within the corpus of programs.

10. The method of claim 7, wherein the statistical information includes interrelation information, the interrelation information identifying, for at least two program elements, a prevalence of a relationship among said at least two program elements within the corpus of programs.

11. The method of claim 10, wherein the interrelation information indicates, for a first program element and a second program element, a prevalence at which an output of the first program element is fed, as input, into the second program element.

12. The method of claim 7, wherein the statistical information includes selection information, the selection information identifying, for at least one program element, a prevalence at which users have selected said at least one program element in response to submitting the keyword information.

13. The method of claim 1, wherein said presenting comprises providing, for a particular ranked program fragment, an indication of at least one keyword that was not used to generate the particular ranked program fragment.

14. The method of claim 1, wherein said presenting comprises:

presenting a first output presentation that informs the user about a particular ranked program fragment, for a first level of confidence associated with the particular ranked program fragment; and

presenting a second output presentation that informs the user about the particular ranked program fragment, for a second level of confidence associated with the particular ranked program fragment,

wherein the second output presentation demands less interaction with the user compared to the second output presentation, and wherein the second level of confidence is greater than the first level of confidence.

15. The method of claim 14,

wherein the first output presentation lists the set of ranked program fragments, including the particular ranked program fragment, and

the second output presentation shows code corresponding to the particular ranked program fragment.

16. A computer readable storage medium for storing computer readable instructions, the computer readable instructions providing a program creation system when executed by one or more processing devices, the computer readable instructions comprising:

logic configured to receive one or more keywords from a user who wishes to generate a program fragment as part of a program being created, to thereby provide keyword information;

logic configured to identify a context in which the user is generating the program fragment, to provide context information, the context information specifying one or more of;

at least one global variable which affects the program fragment, said at least one global variable having been previously defined in the program being created;

at least one local variable which affects the program fragment, said at least one local variable having been previously defined in the program being created;

or

at least one function which affects the program fragment, said at least one function having been previously defined in the program being created;

logic configured to generate a set of candidate program fragments based on at least the keyword information and the context information;

logic configured to rank the candidate program fragments based on ranking information, to provide a set of ranked program fragments; and

logic configured to present an output presentation to the user based on the set of ranked program fragments.

17. The computer readable storage medium of claim 16,

wherein the ranking information is based, at least in part, on statistical information that is produced by analyzing a corpus of programs created by one or more users, each program in the corpus being composed of one or more program elements,

and wherein the statistical information includes frequency information and interrelation information,

the frequency information identifying, for at least one program element, a frequency at which said at least one program element appears within the corpus of programs, and

the interrelation information identifying, for at least two program elements, a prevalence of a relationship among said at least two program elements within the corpus of programs.

18. A program creation system, implemented using one or more computing devices, comprising:

an interface module configured to receive one or more keywords from a user who wishes to generate a program fragment as part of a program being created, to thereby provide keyword information;

a context analysis module configured to identify a context in which the user is generating the program fragment, to provide context information;

a candidate program identification module configured to generate a set of candidate program fragments based on at least the keyword information and the context information; and

a ranking module configured to rank the candidate program fragments based on ranking information, to provide a set of ranked program fragments,

wherein the ranking information is based, at least in part, on statistical information that is produced by analyzing a corpus of programs created by one or more users, each program in the corpus being composed of one or more program elements,

and wherein the statistical information includes frequency information and interrelation information,

the frequency information identifying, for at least one program element, a frequency at which said at least one program element appears within the corpus of programs, and

the interrelation information identifying, for at least two program elements, a prevalence of a relationship among said at least two program elements within the corpus of programs,

the interface presentation module being configured to present an output presentation to the user based on the set of ranked program fragments.

19. The program creation system of claim 18, wherein the context information specifies one or more of:

at least one global variable which affects the program fragment, said at least one global variable having been previously defined in the program being created;

at least one local variable which affects the program fragment, said at least one local variable having been previously defined in the program being created;

or

at least one function which affects the program fragment, said at least one function having been previously defined in the program being created.

20. The program creation system of claim 18, wherein the interface module is configured to provide, for a particular ranked program fragment, an indication of at least one keyword that was not used to generate the particular ranked program fragment.