Ways and scope of interop tests discussion

[Juan Carlos Cruellas <cruellas@ac.upc.edu>]

Dear all,

The co-chairs of the TC would like to advance in the decission making 
process on the interoperability tests.

This message identifies a number of issues that we propose to discuss at 
the TC level and that we think may help in clarifying the scope and the 
way in which the interop tests may be conducted. Please feel free to add 
any additional issue that you consider relevant to be discussed for 
reaching this goal.

Among other relevant issues there are two that we would like to comment 
here:

1. The way in which the interoperability tests will be conducted.

2. The profiles on which the TC will conduct interop tests.


1. WAY OF CONDUCTING INTEROP TESTS

On the first one, we initially envisage three potential situations which 
would make it possible different ways:

a. For a certain profile there is more than one tool that implements the 
same relevant feature. In this case, the way for conducting the tests 
would be the expected one in interoperability tests on tools: once 
specified the test cases, each tool would generate requests and the 
others would generate the correspondent responses.

b. For a certain profile there is only one tool that implements certain 
features. In this case we could conduct a let us say kind of "tool-human 
being" interop tests. This could consist in that one party (tool or 
human being) would generate requests and the other (human being or tool) 
would generate corresponding responses....well, in the case of a human 
being, the responses would obviously be qualitative: would indicate what 
that human being has understood from the request and a detailed 
description of what a server should, in his/her opinion do, generate and 
give back to the requester....

c. For a certain profile there is not any tool implementing features. 
Then the only thing that we could do here is a kind of interop between 
human beings for checking if they have actually understood the same 
things after having read the protocol.

Would you see any other way to conduct the interop tests? if so, could 
you please describe them?


2. PROFILES TO BE CHECKED.

As for the profiles to be checked, well, ideally if we could incorporate 
all of them that would be great...but this greatly depends on agreeing 
the way of conducting the tests and the willingness of the different TC 
members to participante in them.

We would like to suggest to start making a poll for knowing the exact 
situation of each TC member.

Could you please let us know:

a. Tools that implement some profile of the DSS-X? if so, what profiles 
exactly and whether it is a tool that supports all the features of the 
profile or only some of them?

b. Are you willing to participate in the interop tests? and if so, in 
the tests of which profile(s) would you like to be involved?


Could you please react to this message by email so that at the next call 
we have been able to exchange some views and we may have made some 
progress on the issue?

As mentioned above, please feel free to raise any issue different to the 
ones that we have mentioned and that you feel that are relevant for the 
interop tests.

Best regards

Stefan and Juan Carlos.