OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

avdl-comment message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Comment with my YES vote


I have decided to cast BEA's vote for AVDL 1.0 as an OASIS Standard, however I have significant concerns about the specification in its current form. I would not be inclined to approve a future version without substantial improvement in some areas. In general, the document is insufficiently precise about what an implementer must do and must not do to comply with this specification. Experience has shown that tends to lead to practical interoperability problems. Section 2 mixes examples and normative statements in a way that makes it difficult to understand which is which.
I have the following specific suggestions.
1.	Adopt the conventional use of keywords like MUST, SHOULD and MAY as defined in RFC 2119. Add a short section at the beginning stating that you have done so.
2.	Mixing normative text and examples together makes it hard to be precise. Consider an alternative structure, such as a separate section for examples. If you keep the examples intermixed, clearly label each and state at the beginning what is normative.
3.	The schema is not explicitly referenced in the document and it should be. It is reasonable to say that the use of XML Schema is not required, but that AVDL compliant documents must conform to the schema.
4.	Each schema element should be described in a separate subsection. The exact name of the element should be given; along with whether it required or optional; XML attributes if any; relationship or dependencies on other elements; semantics; any syntax constraints not captured by the schema, etc. BTW, I think some of the comments in the schema file would better appear in the specification.
5.	Line 234 says that the complete example is contained in the appendix, but it is not.
6.	Line 180 starts with Traversal Ste. I think you mean Traversal Step. But the paragraph seems to be defining Traversal, Traversal Step and Traversal Step Id. I suggest using three definitions.
I suggest you look at the SAML or XACML specifications as examples of how to construct normative language using the RFC 2119 keywords.


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]