* david_marston@us.ibm.com wrote: >Alex and Bjoern seem to be converging on an understanding that I think >is intended by SpecGL: >A spec specifies the behavior of one or more classes of product. That >is, the class is the conformance subject. Each implementation is one >instance of the class, and it can be tested for conformance. Hmm. Various W3C Technical Reports define and/or use the term "implementation" as opposed to specifications, documents, etc. >Example: the XSLT spec has one class of product, an XSLT processor. It >does not define the notion of conformance for a stylesheet editor, a >validator, or any other product. Why are stylesheets not a class of product here? Is this considered good practise? I must admit that I find it odd that the XSLT 1.0 Rec does not mention stylesheets in its section on conformance or not to consider XSLT stylesheets a class of product of the XSLT 1.0 Recommendation. I would appreciate if the Specification Guidelines require that all requirements of a specification are directly bound to a class of product (which would consider the conformance section of XSLT 1.0 bad practise).Received on Thursday, 4 September 2003 19:30:16 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 10 September 2007 00:41:11 GMT