ws-wsif-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aleksander Slominski <as...@cs.indiana.edu>
Subject Re: [ANN] updated "Super" Dynamic Invoker (SDI)
Date Thu, 02 Dec 2004 17:37:20 GMT
Ananth Krishna wrote:

>I had some questions/suggestions regarding the Super Dynamic Invoker
>(..and subsequently wsif and axis). I thought I d share them on the
>wsif-dev list as it seemed most appropriate to post it here (apologies
>if it wasn't the right place for it)
>
>A) Is the SDI part of WSIF ? 
>
no

>If not will be included in future releases
>soon ?
>  
>
possibly SDI or some form of it - concept and capabilities are more 
important but all ultimately depends on user feedback - do you like it 
and how to improve it :)

>B) On the developer list of Axis, I see there will be support for XML
>Infoset soon. I am assuming we are seeing this as being a good thing for
>Dynamic invocation using complex xsd types in some ways ?If so, will
>WSIF move towards using this feature in Axis when it is available ?
>  
>
absolutely - i am main supporter of having XML Infoset API and it is my 
main focus in work on Axis2

>Th reason I ask all this is that currently we have a custom web service
>invocation framework that we developed a few years back when there was
>little or no support for dynamic invocation using complex xsd types and
>the ones available then were a bag of spanners and very hard to work
>with. Besides, we also had to plug-in a WS-Security implementation to
>our framework and WSS4J wasn't available then.
>
>However, our implementation suffers the same limitations as the Dynamic
>Invoker (in the initial/current? release of WSIF) in that we generate
>java types using castor bindings to invoke services. In our upcoming
>projects a client framework for dynamically invoking services without
>having to explicitly generate java types from xsd's or use generated
>stubs is very important in our software architecture. 
>
>Initially WSIF seemed like a good option but we realised soon that it
>suffered from :
>
>A) having to generate java types for complex xsd types.
>
>B) I believe sun licences for JMS and J2EE are too restrictive for us.
>
>C) The official, stable release of WSIF (2.0 I think) uses a vintage
>version of axis which doesn't include some bug fixes that we require
>(including a patch for Attachments that we submitted and is in Axis
>1_2RC2)
>  
>
this is fixed in CVS that uses the latest Axis 1.2RC2  - i should make 
an interim wsif release soon to gather more feedback

>Our approach to using WSIF currently was to try and have a version that
>is perhaps WSIF Lite (by ripping off support for JMS,J2EE bindings and
>having only a light weight web service client infrastructure). As we
>were thinking about it, we came upon the Super Dynamic Invoker which
>seems in some ways like WSIF Lite. If our thinking regarding the SDI is
>right, it would be good to collaborate efforts on it to achieve more
>than we hope to if we went on our own, say..
>  
>
Ananth, i am ver glad to hear it. 

It would be great if we could discuss in detail use cases and list of requirements for WSIF
dynamic invoker - for example how exactly to deal with complex types?


>It would be good to know what the core developers on wsif think about
>this..
>  
>
as you are interested in shaping future of WSIF you may soon become one 
of them. currently most of initial developer are no longer actively 
working on WSIF so we definitely need "fresh blood" :)

thanks,

alek

-- 
The best way to predict the future is to invent it - Alan Kay


Mime
View raw message