From owner-incits-v1@incits-v1.org Tue Jun 19 07:39:55 2007 Date: Tue, 19 Jun 2007 07:39:06 -0700 (PDT) From: jon.bosak@sun.com Subject: INCITS-V1: SyntacticConformance-Durusau-6 Sender: owner-incits-v1@incits-v1.org To: incits-v1@incits-v1.org Reply-to: incits-v1@incits-v1.org X-PMX-Version: 5.2.0.264296 X-Authentication-warning: jurassic-x4600.sfbay.sun.com: bosak set sender to jon.bosak@sun.com using -f In his comment Durusau-6, Patrick says: DIS 29500 only constrains syntactic conformance to the proposal. As a practical matter, this will result in applications that claim conformance to the proposal but which have in fact re-defined the semantics as defined in the proposal. This will lead to a lack of interoperability due to non-disclosed modification of the semantics as defined in DIS 29500. The statements in DIS 29500 that "document conformance is purely syntactic" and "application conformance is purely syntactic" strike me as bizarre. The point of XML is to associate a set of agreed-upon syntactical constructs with agreed-upon semantics. That's why we do this. Interoperability lies precisely in an agreement about semantics. For something that purports to define an XML vocabulary to say that it doesn't require applications or documents to recognize semantics is to say that it fails in the primary purpose of an XML vocabulary. The DIS here seems to want to have its cake and eat it, too. It wants to define semantics but leave applications free to do what they please. So in the normative part of Sections 2.3-2.5, it says that applications and instances can ignore semantics, and then in the informative "Interoperability Guidelines" of Section 2.6, it tries to make sense of this by appealing to applications and instances to conform to semantics after all. For example: A presentation editor that interprets the preset shape geometry "rect" as an ellipse does not observe the first guideline because it implements "rect" but with incorrect semantics. The overall impression here is one of confusion and an imprecision of goals. So, for example, the statement in 2.5 that "application conformance is purely syntactic" directly contradicts the requirement in 11.3.1 that "A WordprocessingML consumer shall treat the contents of such legacy text files as if they were formatted using equivalent WordprocessingML, and if that consumer is also a WordprocessingML producer, it shall emit the legacy text in WordprocessingML format," which clearly both assumes a known semantic framework and requires a conformant application to respect that framework by translating the constructs of the legacy format into "equivalent" DIS 29500 constructs, that is, constructs having the same semantics. As Patrick notes, The net result of these clauses is that conformance to the semantics as defined by DIS 29500 can be changed for a document instance or by an application that claims conformance to DIS 29500 from those defined therein. In other words, users and application designers will be subject to arbitrary changes in the semantics allegedly defined by DIS 29500, resulting in a lack of interoperability and portability, which are goals for any standard issued under the JTC 1 Directives. And in practice, the effect of this is to allow behavioral details to be determined on an ad hoc basis by the dominant software. The proposed fix is fine as far as it goes: That Part 1, Section 2.3 be amended to include conformance to the semantics as defined by DIS 29500 to claim conformance to DIS 29500. Without the days at my disposal that it would take to investigate this properly, however, I'm left wondering whether the confusion evident here might be more pervasive. Jon