International Business Machines Corporation

Yen Lu

Jason Ashley Sholl

Jeffrey Yu Kwan Liu

Christopher Lawrence Brealey

Jeffrey Y Liu

A programming model allows application programmers to extract Web services from various types of resources. The programming model defines relationships between various resource types and methods for each resource type to extract a Web service or another linked resource. The programming model also provides caching capability to store previously extract resources for future requests.

Extracting web services from resources using a web services resources programming model

Extracting web services from resources using a web services resource programming model

A query for data in a UDDI registry is developed. A URI query having parameters that identify the data is provided. A UDDI query having a UDDI specified format based on the parameters in the URI query is generated. The UDDI query is sent to the UDDI registry.

UDDI metadata query development

A method and system for storing metadata related to Web service entities is disclosed. The method according to a preferred embodiment comprises browsing a Web service and storing metadata relating to an entity of the Web service as a binding description in a format for assisting in an inspection of a site for available services. Preferably, the Web Services Inspection Language, or other format for assisting in the inspection of a site for available services, may be used as a format for storing metadata related to Web service-related entities. Because WSIL is XML-based, the documents that store the metadata are human readable, platform independent and conform to an open standard.

Persistent data storage for metadata related to web service entities

Embodiments of the invention, given a conventional makefile which builds targets in serial mode, will process (e.g., compile, link, pre-process, execute, etc.) commands in parallel on those targets which, as a result of a lack of interdependency, can be built in any order. Embodiments may first identify one or more targets (i.e., object files) which are candidates for parallel compilation ('parallelization'). From the targets identified instructions are generated, which may be stored in a file, which provides the identity of targets for parallelization and the order in which these targets should be built. The instructions may then be executed (resulting in some instructions being executed in parallel) thus providing reduced compiling time and, thus, improving productivity.

Growing community of inventors

North York, Canada

Yen Lu