WSRP TCK / Pluto Provider Implementation

["Rudnev, Alex (OCTO)" <Al...@dc.gov>]

I was going through Pluto's discussion and found some statement I 100% agree
with (sorry for taking it out of the context):

> Date: Wed, 22 Oct 2003 21:38:32 +0200
> From: Stefan Hepper sthepper@hursley.ibm.com
...
> Pluto is also the reference implementation of JSR 168 and therefore 
> passes all compliance test cases from Sun (around 400) that test all the 
> MUST statements that are automatically testable.
>
> The code quality differs between the container and the sample portal 
> driver. The container was designed to be extensible and useable in 
> products, whereas the portal driver is just some sample code that is 
> neither well structured nor extensively tested.
>
> Stefan

and which helps me to formulate the following two questions:

a) Is there some kind of test plan / test cases, which would allow us to
define unit tests (currently the only unit test I see is
oasis.names.tc.wsrp.v1.wsdl.WSRPServiceTestCase, generated by axis). Have
been any documents of this nature produced during WSRP TC discussions? Are
there any plans in IBM or other vendors to contribute some documents (which
most probably exist internally) to WSRP4J, or it's better to start
developing it from scratch?

b) Currently, WSRP4J Pluto Provider is based on this sample portal driver.
Does it make sense to kick a dead horse (I'm overstating here of course) or
it's better to start some additional effort in background (current
implementation is working somehow, so we can take our time) to re-implement
this wsrp4j-provider-api-to-pluto-container-api connector from ground up
(without dependencies on portalImpl)? Or is this pluto-provider just another
sample which is not supposed to be used in production environments (so, why
bother)?

Regards,
Ru

Re: WSRP TCK / Pluto Provider Implementation

[Richard Jacob <ja...@apache.org>]

Rudnev, Alex (OCTO) wrote:

>a) Is there some kind of test plan / test cases, which would allow us to
>define unit tests (currently the only unit test I see is
>oasis.names.tc.wsrp.v1.wsdl.WSRPServiceTestCase, generated by axis). Have
>been any documents of this nature produced during WSRP TC discussions?
>
Yes kind of.
There were/are two subcommittees concerned with this issues.
1. Interop
This sc deals with interoperability among various vendors implementing wsrp.
Here loosly coupled tests are performed as far as the implementations go.
2. Conformance
This SC defined conformance statements and levels.
Also a WSRP conformance test suite is being implemented.
http://www.alphaworks.ibm.com/tech/wsrptk

>Are
>there any plans in IBM or other vendors to contribute some documents (which
>most probably exist internally) to WSRP4J, or it's better to start
>developing it from scratch?
>
The WSRP TC generated some sheets, not sure if they are public yet.

>b) Currently, WSRP4J Pluto Provider is based on this sample portal driver.
>Does it make sense to kick a dead horse (I'm overstating here of course) or
>it's better to start some additional effort in background (current
>implementation is working somehow, so we can take our time) to re-implement
>this wsrp4j-provider-api-to-pluto-container-api connector from ground up
>(without dependencies on portalImpl)?
>
I agree here. Since the portal impl will not be seriously continued we 
should be only dependant on the container.
However, the puropse here was to reuse as much as possible from pluto.
I think this would be a larger refactoring needing investigation what 
this would mean, what needs to be done etc.


Richard