Summary
In which Andy suffers an extreme bout of cynical disbelief in the merits of Resource Oriented Architecture and "solving" the discovery problem by giving up and just having a description as part of a "Resource".

In Replacing Service Oriented Architecture with Resource Oriented Architecture Alex Bunardzic slanders SOA by saying it implies Any consumer-to-be must keep an inventory of the available services of interest that are on the web.

Funny - I thought it was the role of Registries to keep track of things.

His ROA solution, however, glosses over the problem of finding services with the broad assumption that they are out there on the Web and so all they need to do is encapsulate a description. Presumably Google Solves All and uses that description to make the service findable.

That simplistic, singular approach seems to ignore the possibility that there might be multiple reasons to use a given service and therefore multiple ways to find it.

Once we start thinking about composing real business solutions, the (legal) need for accountability complicates a blithe approach to service selection and services are only likely to be used because they are available through a corporate registry.

Imagine a CEO signing off on a Sarbanes-Oxley report that figures were calculated using some service casually discovered because it ROAred loudest!

Talk Back!

Have an opinion? Be the first to post a comment about this weblog entry.

RSS Feed

If you'd like to be notified whenever Andy Dent adds a new entry to his weblog, subscribe to his RSS feed.

Digg | del.icio.us | Reddit

About the Blogger

Andy is a free-lance developer in C++, REALbasic, Python, AJAX and other XML technologies. He works out of Perth, Western Australia for a local and international clients on cross-platform projects with a focus on usability for naive and infrequent users. Included in his range of interests are generative solutions, software usability and small-team software processes. He still bleeds six colors, even though Apple stopped, and uses migration projects from legacy Mac OS to justify the hardware collection.