Talk:Web Services Resource Framework
|WikiProject Computing||(Rated Start-class, Low-importance)|
the content here naiively belived the WSRF specification and unduly criticised all other approaches to implement state in web services.
I've made the core of the document less opinionated, and moved opinions (with citations) to the Issues section. As with all WS pages, you need a politics section to understand the past behind the specification, and its future.
SteveLoughran 12:47, 5 October 2006 (UTC)
Self Published Citations
Stuartyeates has tagged this as containing self published citations. What is being referred to here? The technical references are to OASIS standards, which, as a recognised standards body, is capable of issuing normative standards. WS-RF does refers to both the draft (non-normative) WS-Addressing and the final version is one example of the flaws of WS-RF; a criticism isn't raised in the article -though anyone is free to do so.
The only other (current) citation is the "Web Service Grids: An Evolutionary Approach" article, which was issued by the [National eScience Centre. That is part of the UK academic system and as the paper was co-authored by staff as many of the main universities in the UK, it is hard to say "self published". That paper was controversial, but as Tony Hey was one of the authors, you cannot say the people who wrote it were not qualified -or the organisation lacks the authority to self publish. It defined the UK grid organisation's policy regarding WS-RF. I am one of the people to be acknowledged in the paper, but as Jim Gray is also in that list, that is not something about which I wish to hide: we considered ourselves to be the peer-review.
Web Services Resource Framework
"A web service by itself is nominally stateless, i.e., it retains no data between invocations." This is erroneous and ambiguous. What do you mean by a "Web Service". Something built using SOAP+WSDL? SOAP is just a protocol. One can easily build stateful interaction on top of it, or not. HTTP is stateless. Dido. Rather it should be ~ "No standard existed to define stateful interactions in the Web Services framework / family of specification. Where required Web Services implemented stateful behaviour in custom ways.