Objectives for the Web Service Hackathon

Background

As the databases in life science is keep growing and daily updated, integration of these databases by mirroring in one organization is not feasible. Instead, it is required to develop the way to utilize variety of databases directly over the distributed environments.

This situation brought on SOAP/WSDL web services to be launched at the major database centers around the world including NCBI and EBI. In the same way, DDBJ, KEGG, PDBj, and CBRC has been released relevant web services in Japan.

Several projects like BioMOBY and myGrid/Taverna projects have been started to utilize these services in an unified way, and it is strongly believed that this kind of integration should be accomplished by the SOAP/WSDL based web service technology in this field.

However, inconsistent specifications of the data structures and the naming conventions among these services prevent the interoperability. Besides, the insufficient documentation of each services is one of the bottle necks to popularize these services. Additionally, the web service is originally designed to be utilized by every programming languages which have support for SOAP/WSDL, but there are several services which can't be used with some languages. It is also hard to handle the cases like temporal service down or whatever error occurred while executing a job for end users.

To improve this circumstance, it is a key to standardize the data structures to be passed among existing web services and to increase the number of tools and databases accessible through the web service for creating practical bioinformatics workflows.

Efforts in DBCLS

Thus, we at DBCLS are planning to construct a proxy-like server, which aim to provide unified and consistent usages of the existing services. For this purpose, we would firstly consider:

• provides sufficient documentation for the usage of each services
• ensures every operations to be accessible by the widely used languages like Perl, Ruby, Python and Java.
• translates data structure exchanged among servers to create seamless workflows
• concatenates pipelines over several steps to be done on the server side in typical cases for the effecitivity
• handles various errors returned by external services at the server side as much as possible

By conducting our proposal, not just providing user friendly environments to many researchers, the server can act as an infrastructure for constructing practical workflows effectively utilizing existing services. At the same time, we will also provide new web services for tools and databases to be developed at DBCLS.

Objectives of this Hackathon

To discuss about a standardisation of:

• data structures exchanged among services
• naming convention of the methods (operations)
• technique of the job management for the time consuming query

which at present differ within each providers.

Therefore, we are going to held developer's meeting in February of 2008. There, key members related to each web service providers around the world and core developers from BioMOBY and Open Bio* (BioPerl?, BioRuby, BioPython?, and BioJava?), would be offered to attend this meeting, staying for about a week in Japan.

Historically, the class design, which represents bioinformatics data types, has been defined by each Open Bio* project, independently. In this opportunity, to define the standard specification for the biological objects through the web service will also benefit interoperability among these libraries. Seamless integration of the remote (web services) and local (installed tools) environment is an another challenge for those projects.