[bt-spec] FW: Draft response to 'Conformance Requirements forSpecificat

bt-spec message

Subject: [bt-spec] FW: Draft response to 'Conformance Requirements forSpecifications' v0.4 14 Jan 02

Here is a draft response to 'Conformance Requirements for Specifications' v0.4 14 Jan 02 that you sent around awhile ago for you to review. My assumption is that you would like to receive responses and will produce a response from the BTP team based on the responses received.

The OASIS Business Transaction Protocol (BTP) TC team have received your document 'Conformance Requirements for Specifications' v0.4 14 Jan 02 and have attempted to apply it to our BTP specification.

We have found it a helpful reminder and guide as to what we should include in our Conformance section. Given that your document is currently only at version 0.4 we have not attempted to provide detailed comments at this stage. However, we would like to make the following comments aimed at helping you to help us more.

1) We appreciate that OASIS TCs produce many different kinds of specification concerning quite different entities which makes providing specific guidance not easy. However several OASIS specifications concern protocols (as does ours) and several concern 'processors'. We suggest that for these more common cases you could provide some more specific guidance.

2) While we would not want you to be unnecessarily restrictive or proscriptive, it would be helpful if you were to provide a 'model' conformance clause perhaps as an informative annex. You might need more than one for the different cases - e.g. one model for protocol specifications and another for processor specifications, and perhaps a third very generic one. Taking the protocol one it would show a recommended order of sub-headings / tackling the various conformance issues and concerns you address in the main part of your document. It should also indicate the level of detail one is recommended to go to.

3) The sample conformance claim in Appendix A is a good idea but seems to rather miss the target at present. A proper proforma for Conformance claim proforma would be helpful. The current text of this appendix seems on one hand to be very 'testing' orientated and on the other does not address the many issues called out in the main part of your document (such as for which levels or profiles conformance is claimed, options implemented / not implemented, and so forth). It has to be born in mind that standardised test cases / testing suites, etc are unlikely to exist for a new OASIS specification, and it is not generally worth producing such until the specification has proved itself in the market place (although we accept the desirability of standardised testing to help promote compatibility / interoperation).

4) With respect to sub-section 8.1.2 Specifying Conformance Claims we suggest that a note is added pointing out that any requirement for strict conformance should be very carefully considered and is not usually a good thing as it can prevent sensible interworking of implementations of later versions with this version.