CNet is reporting that IBM has donated a Java implementation of the Web Services Inspection Language (WSIL) to the Apache Project. This is great news that I hope leads to more discovery of its potential uses as a better means for the publishing and discovery of Web services today.
I've been keenly interested in the WSIL specification for sometime because it represents a more practical approach for discovering Web services today then UDDI. It has been well documented that UDDI has some significant issues from attempting to be too grandiose and thereby maligning its true purpose. Should we be focusing our attention on WSIL rather then UDDI? I think so. Let me explain.
UDDI is a specification for a registry server where service providers can publish their identity, classify their offering and provide technical information for service consumers to bind to a service. There is a collection of XML formats for storing the different record types in UDDI that can be accessed with a SOAP API.
WSIL is a relatively simple XML file format that is published with a standard URI on a provider's web site and read by consumers to discover and use the provider's listed web services.
Without noting their details, the difference between UDDI and WSIL are subtle. In comparison to WSIL, UDDI is significantly more complex (multiple XML formats and an API) and requires more overhead (server software) to be implemented. WSIL is a lightweight and decentralized mechanism for discovery of Web services.
WSIL assumes the service consumer has found and at least knows the provider -- perhaps they have has established a business relationship through traditional means. In reality, most business partners today do not find one another in a registry, and it is unlikely that they will replace negotiating agreements with automated ad-hoc means. (For a more in depth comparison of UDDI and WSIL read "WSIL: Do we need another Web Services Specification? Explaining the difference between UDDI" by Tarak Modi.)
In a lot of ways, WSIL is quite REST-ful[1] in nature -- its like RSS for Web services. RSS is a file format with pointers to published content that can be syndicated and aggregated. WSIL is a file format with pointers to published Web services that can be discovered and bound.
Its also far less resource intensive which is important to lowering the barrier to adoption and broadening the potential uses of this information.With Web services capable of being published and discovered so simply, new uses of WSIL and novel services would likely develop that could further diminish the need for UDDI in its current form.
This is an exciting possibility to me that I think deserves more attention. WSIL offers a more practical and lightweight approach to an area that needs a better answer if Web services are to proliferate. IBM's donation to the Apache Project is helpful in furthering WSIL's potential. More support is needed in other languages and tools and more importantly recognition by the standards bodies.
---
[1] I'm not a devotee of REST, but I think it makes some good observations that should be noted by the Web services community.
Leave a comment