Microsoft Technology Licensing, LLC

Erik Meijer

Matthew J Warren

Anders Hejlsberg

Wolfram Schulte

Dinesh Chandrakant Kulkarni

Christopher Lovett

Barend Hermanus Venter

Chia-Hsun Chen

Luca Bolognese

John Wesley Dyer

Jeffrey Van Gogh

Mads Torgersen

Dragos A Manolescu

Moe Khosravy

Christopher Wellington Brumme

Christian Liensberger

Avner Y Aharoni

Gary Shon Katzenberger

Srikanth Mandadi

David N Schach

Andrew Eyring Kimball

Peter A Hallam

Derek Denny-Brown

David Lawrence Remy

Ralf Lammel

James H Hogg

Lucas James Hoban

Craig T Sinclair

Peter Drayton

Peter H Golde

Rene Jacco Bouw

Mark W Fussell

Evgueni Zabokritski

Ion Vasilian

Dario Russi

Dare Obasanjo

Dinesh Chandrakant Kulkami

Ilia G Ioffe

Matthew J Wararen

Serge Lidin

An operation (such as a relational query) may be processed on a processing engine (such as a relational database server) on behalf of a client. A conventional processing involves the delivery of the operation to the processing engine, which executes the entire operation to completion and returns a result data set. It may be more efficient to allocate part of the operation to be performed on the client, but a developer may be unable or unavailable to rewrite the operation in a distributed manner. Instead, the operation may be automatically partitioned into a pre-engine client portion, a processing engine portion, and a client portion, and the instructions of each portion may be automatically allocated respectively to the client, the server, and the client. The partitioning may be adjusted to conserve computing resources, such as bandwidth and storage, and the instructions may be reordered to improve the processing of the operation.

Automated client/server operation partitioning

Various aspects for providing a runtime for dataset transformation are disclosed. In one aspect, data consumption requests are received. The requested data is retrieved from a web service in an arbitrary form, and subsequently transformed into a standardized result from the arbitrary form. In another aspect, a user interface is displayed to a content provider, and an input identifying a transform function selected by the content provider is received via the user interface. A transform assembly is then generated based on the input, which facilitates retrieving data from a web service in an arbitrary form, and transforming the data into a standardized result by calling the selected transform function. In yet another aspect, third party code is associated with a transform assembly. An execution of the third party code is then managed such that at least one aspect of the transform assembly is operated according to the execution.

Web services runtime for dataset transformation

A method of binding elements of a structured document to an observer structure includes obtaining the logical structure of a document. Nodes representing information contained in the document are mapped to an observer structure which can include both a user interface or a programming object. The user interface may be a graphical user interface including a display form of a grid-like structure to contain the structured document information. The data binding which maps the structured document information to the observer form may be directionally controllable such that any change made to the information at the observer may be reflected in the source structured document.

Method and system for mapping between structured subjects and observers

A fuzz testing system is described herein that represents event sources, channels, processors, and consumers as first-class entities in an application. Abstracting event-related entities allows fuzzing through injecting, dropping, reordering, and delaying events from within the application. This translates into the ability to localize the areas under test and perform fuzzing in a systematic manner. In some embodiments, the fuzz testing system tests concurrent asynchronous and event-based code, and can generate event streams based on a given statistical distribution. Representing events, event sources, processors, and sinks as first-class objects provides easy access to the event handlers and facilitates implementing fuzzing by introducing event processors between the source and the sink. Thus, the fuzz testing system improves the testability of applications and APIs with asynchronous behavior and provides a uniform framework for introducing fuzz testing into such applications.

Fuzz testing of asynchronous program code

Data source queries may be specified in various languages, such as SQL, XQuery, and object-oriented languages, and each language may have various linguistic features that enable different types of queries. It may be advantageous to formulate mixed language queries having portions specified in a first language and portions specified in a second language, and to provide mixed language query evaluation components that can parse portions of queries of different languages. It may also be advantageous to permit the translation of single language queries into multiple languages to be parsed by such mixed language query evaluation components. Such techniques may permit queries to utilize linguistic features and processing techniques of one language that are less efficient, less direct, or simply unavailable in another language.

Data query translating into mixed language data queries

Queries targeting various data sources are processed in a query processing pipeline that parses the query into a set of operations (e.g., an expression tree or a translated SQL query) using a set of query operators, each handling a particular type of operation. The query operators are often designed in an unspecialized manner, such that each query operator handles one query operation in an atomic, generic manner (e.g., sorting generic data items for an ORDER BY clause.) More efficient queries may be devised by including specialized queries that operate in common but special cases, such as a sorting of a particular data type (e.g., a floating-point number sort) or a sequence of two or more operations that are often performed together (e.g., a WHERE test of an attribute followed by a SELECT of the same attribute.) The use of specialized operators may result in the formulation of more efficient queries.

Query processing with specialized query operators

An arrangement by which pointers may be represented in a restricted software execution environment that provides access to only reference types but not pointers is realized by modeling both pointers and value type boxing using a set of boxes (i.e., containers) which describe how to retrieve the value associated with the pointer, and emitting code for accessing the set of boxes. A decompiling and recompiling process is utilized in which code arranged to run in an execution environment that supports pointers is scanned to reveal pointer and boxing usage. In one illustrative example, code is responsively emitted to create one or more specific boxes. At runtime in the restricted execution environment, an instance of a box class is created to thus emulate a pointer to the value.

Representing pointers and boxing in environments using only reference types

A sensor network may be created by collecting information from a plurality of mobile devices, such as cellular telephones. The mobile devices use sensors, such as microphones, cameras, accelerometers, biometric readers, etc. to detect the sensory information. Sensory output related to detected sensory information may be transmitted from a mobile device to a receiver that receives the sensory outputs. The received sensory outputs from a plurality of mobile devices may be aggregated to generate aggregated data, which may be transmitted to one or more receivers that use the aggregated data to perform a function and/or present the aggregated data to a user. Additionally, the aggregated data may be reviewed by a reviewer component that can create (revised) detection instructions regarding how one or more mobile devices are to detect sensory information and/or what sensory information to detect (e.g., to make resulting aggregated data more relevant).

Mobile sensor network

Queries against data sources (such as language-integrated queries to be applied against relational databases) may be prepared for processing by a query processing pipeline. This pipeline performs services including transforming the query into a form more easily applied by the data source, e.g., into SQL, by matching query terms with query operators configured to handle various query operations. Many query processing pipelines include single-item query operators that perform individual operations (e.g., an updating query operator that specifies an updating of individual records of a table.) An integrated query pipeline may be devised that includes both single-item query operators and multiple-item query operators that specify operations on sets of data items. An integrated query processing pipeline may analyze the query to determine whether a single-item or multiple-item operation is specified in order to select an appropriate query operator, which may improve the efficiency and performance of the data source querying.

Query processing pipelines with single-item and multiple-item query operators

The utilization of data services (such as web services) may involve the invocation of services, which may reduce performance and overburden the server. In some scenarios, the invocation may be specified in a language that does not support batch invocation, but it may not be feasible to re-code the invocations in a language that supports batching. Instead, the service invocations may be automatically translated into a batch logic specified in a batch-capable language and sent to the service as a batch. For example, a local proxy may receive a language-integrated query having two or more service invocations, compile the query into an expression tree including the two or more invocations, and send the expression tree to the service host to achieve the batch processing of the service invocations.

Automated translation of service invocations for batch processing

Object instance annotation technology is disclosed. This technology is capable of annotating any number of instances for any number of objects. Object instance annotations can be of various types, and annotations can serve as data extensions. Objects are enabled for annotation either by developers, or, alternatively, automatically by a programming language or framework. Furthermore, an interface is provided for annotations. This interface has a variety of operations, some of which allow for adding, retrieving, or removing annotations. Annotations can be reference types or even value types. They may also be private class types. Once object instances are enabled for annotating, and the interface is set up, this interface can annotate a multitude of different kinds of objects. For example, it can annotate generic objects, XML trees, CodeDom trees, and so on.

Instance annotation in object-oriented programming

A type system employing structural subtyping is disclosed herein. A core type system supports several structural types, such as stream, choice, intersection and sequence. Also part of the core type system is a new invariant type, which denotes values whose dynamic type is the same as its static type, and type restrictions for limiting a range of a base type. Furthermore, a streamlined structural version of delegates, called structural delegates and a validation method thereof are introduce into the type system. To further facilitate type safety, strict statically checked interface casts are introduced.

Core object-oriented type system for semi-structured data

Growing community of inventors

Mercer Island, WA, United States of America

Erik Meijer