RE: Adoption and support for WS-* in general and WS-Notification in particular

[Pim van der Eijk <pvde@sonnenglanz.net>]

Hello,

Most of the WS-* specifications were first implemented in products from major vendors around the time the specifications were developed, quite a few years ago. Encouraged by vendors, analysts and consultants etc. many end users adopted them at the time, and continue to do so. It is true that OASIS/W3C committees stopped developing newer versions of the specifications or even closed, but this does not mean users immediately stop using these specs for existing applications, in fact some continue to adopt them also for new applications. As long as there are customers and the impact of maintaining support for these specs in newer versions of vendor products is limited, vendors seem willing to continue to support them.ÂÂÂ

In healthcare, WS-* including WS Notification are used in the IHE Technical Framework, providing for so-called document subscriptions, see https://wiki.ihe.net/index.php/Document_Metadata_Subscription. IHE is very widely used, including the part that uses WSN. So there is at least one active large user community.  IHE implementations rely on underlying toolkits (Java application servers, .NET, Biztalk etc.) to provide WS-* functionality, so this provides a business case for the providers of those toolkits too.

I reviewed WS Notification for a customer project in the past two years, where we implemented an event brokering service.ÂÂ My personal view on WSN is that:

• From a functional point of view, WS Notification covers the basics of notification data exchange well. In particular, it support push and pull data access.
  
• It would have been useful if the WSN specifications would have separated the core notifications functionality from the WS-* binding. This would make it easier to provider alternative transport protocol bindings, including bindings for protocols that did not exist at the time the specifications were written.  My project required support for several protocols, includings ones not based on SOAP.
  
• WS Notification does not provide a complete, fully dynamic model, e.g. there are no defined interactions to dynamically create or delete topics. My project required an ability for users to create (and later delete) their own topics to selectively and securely share information, from applications using messaging, with the same ease of a mailing list or social media group.
  

So in summary, I think WSN could be an option for communities that have existing actively maintained WS-* based exchanges that are organized relatively statically.  In other cases, there are many other alternative newer interoperable technologies (including from OASIS) to be considered.

Kind Regards,

Pim

Â