This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
(1) Both: it would help readers if we document the "strategy" used to arrange the set of terms, e.g. as text between the heading and the def list. I think the current strategy is to "build up", so each definition can re-use terms already defined, but we do not do so in all cases so it is difficult to be sure. (2) SML: "Reference" probably should be "SML Reference". May be covered by 5388. (3) SML: The def for "model def docs" ends with "...in SML model validation." Later the term "model validation" is defined, but not "SML model validation". Rigorous consistency is most likely to help readers. (4) SML: The def of target complete URI refers to two terms not currently defined, "target" and "SML reference". The former probably needs to be added. The latter should be defined, and may already be covered by 5388. (5) SMLIF: SMLIF introduces some concepts not currently defined in 2.2 in either spec, that we should probably add, e.g. alias, schema document binding. (6) Both: SMLIF section 5.4 is titled "Rule Document Bindings", a term I see in neither spec's 2.2 content. Needs to be reconciled against the other rule/binding terms floating around.
Add: PSVI, mentioned but not defined in 4.1.1 SML Reference. In SML, probably just spell out the acronym Post Schema Validation Infoset, and refer to the Schema Structures spec since that is where the "real" definition lives.
Add: "processor", in 4.2.5 SML Null References Add: "SML processor", in 4.2.6 deref() XPath Extension Function I fully realize these terms, likely having the same meaning, may be rationalized at the editors' discretion and potentially removed by 5406.
reviewed in 01/31 meeting; resolution to mark editorial
Regarding original proposal: 1 - I chose the alphabetical strategy and specified that in both specs. That appears to be the easiest to understand and maintain. All definitions have links to definitions on which they depend. See both terminology sections in SML spec and SML-IF spec. http://dev.w3.org/cvsweb/~checkout~/2007/xml/sml/build/sml.html?rev=1.165&content-type=text/html;%20charset=iso-8859-1#Terminology http://dev.w3.org/cvsweb/~checkout~/2007/xml/sml/build/sml-if.html?rev=1.111&content-type=text/html;%20charset=iso-8859-1#Terminology 2 - already fixed in another bug 3 - I removed SML from all definitions except 'SML reference' and 'SML reference scheme' since 'reference' and 'schemes' are used with different meanings in the specs. All other terms are defined without the 'SML' prefix. 4 - Added definition of target: Target - An element in a model to which an SML reference resolves is called the target of that SML reference. 5 - Added several new definitions to SML-IF. Also referenced SML spec for all other terms. 6 - Removed 'document' from the use of bindings so we now have "rule bindings" and "schema bindings". Rule bindings is defined in SML spec. Schema bindings is defined in SML-IF spec. Regarding comment #1: Fixed as stated. Regarding comment #2: The comment is confusing - e.g., not clear what "add processor" means since that word is already used in section 4.2.5. However, I believe that this comment will be covered in Bug #5406 so nothing is needed here. Also added links to terms throughout both specs. The main changes are in the terminology sections.
+1 for the changes.
+1 for the changes, but the definition of SML Reference Scheme as "A representation of an SML reference" is vague. One needs the read the spec to understand what kind of "representation" is implied here. ("Oh, now I can it, that's what a representation of an SML reference is." I would suggest something more along the lines: SML Reference Scheme: A precisely defined means by which SML references are specified. Also, I'm not sure about the use of the passive voice in the following sentence: "An SML reference may be represented by. . .."
I agree w/ Kirk, overall a vast improvement. SML Ref Scheme is a bit mushy though. It feels like it is mixing the definition (what the scheme consists of) with other statements that are true but not really part of that definition. Proposal: SML Reference A link from one element in an SML model to another element in the same model. An SML reference may be represented by instances of multiple, different SML reference schemes. SML Reference Scheme An SML Reference Scheme RS is a set of rules defining the syntax used to represent an instance of RS in the context of an SML reference, plus a set of rules for resolving an instance of RS to a set of target element nodes. This specification specifies how an SML reference scheme must be defined [4.3 SML Reference Schemes] and also specifies one such SML reference scheme, the SML URI Reference Scheme [4.3.1 SML URI Reference Scheme]. Whenever "reference scheme" occurs in this specification, it should be assumed to mean "SML reference scheme" unless otherwise noted. Despite similar names, the term SML reference scheme is unrelated to XPointer schemes and URI schemes. Target An element in a model to which an SML reference resolves is called the target of that SML reference. Multiple SML references targeting the same element may use multiple, different SML reference schemes. original text + (reasoning for each change above) A representation of an SML reference. (this defines an instance of a ref scheme, which is different than the ref scheme definition) An SML reference may be represented by multiple, different SML reference schemes. (true, but relevant to sml ref not sml ref scheme's def) Also, multiple SML references targeting the same element may use multiple, different SML reference schemes. (ditto, relevant for target not ref scheme def) This specification specifies how an SML reference scheme must be defined [4.3 SML Reference Schemes] and also specifies one such reference scheme, the SML URI Reference Scheme [4.3.1 SML URI Reference Scheme]. (true, first part at least relevant, second part...?) Whenever "reference scheme" occurs in this specification, it should be assumed to mean "SML reference scheme" unless otherwise noted. (relevant) An SML reference scheme is unrelated to XPointer schemes. (relevant, perhaps incomplete if the goal is to distinguish from other easily confused "schemes")
Resolution: remove the second sentence in all three definitions and leave the rest ( looking at comment #7 in bugzilla ) No need for another 'needsReview'.
Fixed per comment #8 with one change - replaced "RS" with "the reference scheme".