axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sanjiva Weerawarana <sanj...@opensource.lk>
Subject RE: [Fwd:[Axis2] WSRF Implementation]
Date Thu, 02 Mar 2006 15:22:00 GMT
Wow interesting indeed .. Axis2 appears to be penetrating deeper in
places we didn't know about ;-).

I hope that the Muse folks realize they can't just suck in a bunch of
code from IBM .. it needs to go thru the incubator.

Sanjiva.

On Thu, 2006-03-02 at 09:27 -0500, Hawkins, Joel wrote:
> Update on the WSRF saga - from the Muse mailing list. 
> 
> <snip>
> We are currently looking at merging the three Apache projects (WSRF, 
> Pubscribe and Muse) into a single entity under the Muse project.  While 
> I have done some work on this there have been other things happening 
> which have made me reconsider the project as a whole.
> 
> We have been talking with a team at IBM, whom we work with through the 
> spec committees, to do a joint next-generation implementation of WSDM, 
> and its sub-specifications.  The implementation would not only position 
> the WSDM effort in a more future-proof way (i.e. move towards WSDM 2.0), 
> but continue to move the project forward in a positive direction.  IBM 
> has code which they would like to contribute as the foundation for this 
> effort.  The code will initially be run in Axis2 which would please 
> those who are seeking an Axis2 implementation of WSDM (WSRF,WSRN, etc).  
> 
> IBM is willing to contribute base code, which will be reviewed by us, 
> and developers to the effort.  IBM would like to be part of moving the 
> Apache WSDM effort forward, and is excited about the possibilities of 
> being part of the project.  They are even planning on contributing 
> tooling around this effort to the Eclipse Project.
> 
> I believe this to be the right thing to do for the direction of the Muse 
> project.  Bringing IBM to the project will entice more users, generate 
> more interest in the project and bring more active committers moving the 
> effort forward.
> 
> Here is my +1 for this effort.  I hope others will join me in support of 
> this.
> 
> I apologize that I have not been forthcoming thus far, however I had to 
> consider the legal side of things before making any statements regarding 
> this.
> </snip>
> 
> Part of this effort will be carried out in conjunction with the Eclipse TPTP
> and Corona projects, which will have the additional benefits of bringing
> Apache and Eclipse closer together, and injecting the Axis2 stack into the
> OSGi platform. Pretty exciting stuff, at least to me!
> 
> Cheers,
> Joel Hawkins
> 
> 
> 
> -----Original Message-----
> From: Hawkins, Joel 
> Sent: Monday, February 27, 2006 9:06 AM
> To: 'axis-dev@ws.apache.org'
> Subject: RE: [Fwd:[Axis2] WSRF Implementation]
> 
> 
> Steve,
> 
> Funny you should mention those features - they're exactly what we're doing
> for integrating WSRF with our Corona project! I'll take a look at smartfrog
> - thanks.
> 
> Sanjiva, I'll check in with Muse and WSRF to see if anyone over there's
> contemplating an implementation.
> 
> Cheers,
> Joel 
> 
> -----Original Message-----
> From: Steve Loughran [mailto:stevel@apache.org]
> Sent: Monday, February 27, 2006 6:27 AM
> To: axis-dev@ws.apache.org
> Subject: Re: [Fwd:[Axis2] WSRF Implementation]
> 
> 
> Sanjiva Weerawarana wrote:
> > AFAIK no one has so far signed up to doing a WSRF impl over Axis2. If
> > you're willing to do it that's very cool! 
> > 
> > When you're ready we can create a new maven module for that code or if
> > its more appropriate you can talk to Muse folks and ask them what their
> > plans are. To be honest, I haven't looked at Muse at all so have no idea
> > how easy/hard it'll be to port to Axis2 vs. writing one from scratch.
> > 
> 
> I have hints of a WSRF app on the smartfrog repo, but its all LGPL so 
> you can't use it.
> 
> What you could do is look at the architecture which has the following 
> features
>   -uses URIs to index things, not the fancy (and unstable, less 
> interoperable) WS-A reference props/params
>   -doesnt have a simple resource -> object binding. Instead each 
> endpoint gets to provide its own mapping code, which is useful for a 
> high availability solution where state can be stored in a db, 
> tuple-space or whatever.
>   -does not use xmlbeans, but works at the Axiom level. Keeps it 
> lightweight.
> 
> The thing to remember is that although you could export object instances 
> as WSRF endpoints, you shouldn't. distributed object applications have 
> known failures, which we don't want to reimplement.
> 
> Not looked at muse; don't know its current state.
> 
> 
> =steve
> 
> 
> 
> The contents of this e-mail are intended for the named addressee only. It
> contains information that may be confidential. Unless you are the named
> addressee or an authorized designee, you may not copy or use it, or disclose
> it to anyone else. If you received it in error please notify us immediately
> and then destroy it. 
> 
> 
> 
> The contents of this e-mail are intended for the named addressee only. It
> contains information that may be confidential. Unless you are the named
> addressee or an authorized designee, you may not copy or use it, or disclose
> it to anyone else. If you received it in error please notify us immediately
> and then destroy it. 
> 


Mime
View raw message