What are Web Services? Do you need to know about them?
I would argue basically you just need to know they are a new technology for building applications. If you don't understand the concepts and terminology around application programming interfaces (APIs) and object- or component-based development and design, I think it's a bit hard to grasp their purpose.
The idea is basically Service Providers provide modules that do "something" e.g. your organization might create an Author Profile service that when given an author's name, returned a biographical profile and/or photo and/or birthdate. Another organization could use this to extend the functionality they offer on their website, calling you service to display author photos beside book entries.
As some areas of science become more and more data-driven, it becomes useful to have various services to manipulate, process, or otherwise provide services around this data. Biology and genomics are two areas where some of these types of services have been developed.
Conference reporting from ASIST 2005 on presentation "Workflows and Service Composition in the domain of biomedical research".
Workflows and Service Composition in the domain of biomedical research
Tim is looking at the combination of existing services (database systems, analysis tools, WSDL) to form higher level services - mapping disimilar services to form cohesive whole.
Web services are a good fit for collaboration in the biomedical domain because independent researchers create heterogeneous datasets.Web service standards provide the access modality to support collaboration.
The Taverna System based on web services provides graphical representations of workflow (services) in XML (SCUFL).Measures of coherence in service composition:
semantic meaning and syntactic meaning must match for in/output to work
www.biomoby.org presents a set of standards for sharing datasets.
from ASIS&T Annual Meeting Blog » Collaboration in Digital Libraries
The earlier poster said that he didn’t get Tim Patrick’s talk, so I thought I’d give my own interpretation of it.
There are two different meanings of collaboration in this session, one is collaboration between different groups of people, internal or external, covered by the other speakers. Collaboration in Tim’s sense is also called workflow.
Tim’s own project does this using Taverna, which is an open source workflow engine and also includes tools for building workflows. The things that workflows are built of are called services, and services are accessible over the net. Taverna doesn’t provide services, but enables the user to build complex workflows made up of Web Services (such as retrieving genetic sequence or analyzing that sequence.) Web Services can be made available by different groups, inside the same institution or across different institutions.
Being able to combine these services, and compose them into larger services, allows people to solve ‘big’ problems. What’s required to make this collaboration possible was the main focus of the talk.
To make this possible, many things are needed, one of the most important is a data-type ontology. In Tim’s field, the most important data-type ontology is called BioMoby. Taverna uses an XML format called SCUFL to save its workflows, and it uses BioMoby. People in different institutions can reuse these files to access services, provided they have permission to do so. Reusing these files is handy because most of the people who use Taverna have a lot of domain-specific (in this case Biology) knowledge, but not a lot of knowledge about linking these services together, and by reusing SCUFL files they don’t need to develop it.
For data-type ontologies to work, they need two sorts of agreement in what the data types are — these are called syntactic and semantic integrity. Syntactic integrity means that data types are represented (written) the same way, and semantic integrity means that the data type has the same meaning. One language that has both of these features is OWL-S, previously known a DAML.
from ASIS&T Annual Meeting Blog » Collaboration in Digital Libraries (Tim Patrick)
Comments