Subject: [LC response] To Daniel Barclay
Thank you for your comment
on the OWL 2 Web Ontology Language last call drafts.
Thank you for your careful proof reading. Most of the problems you identified had either already been fixed or have now been fixed -- please see the relevant diff for details .
Regarding "side-effect", the hyphenated form is in the OED , so we didn't change it.
Regarding xsd:hexBinary and xsd:base64Binary, Section 2.2.3 of the XSD 1.1 specification states that "For purposes of this specification, the value spaces of primitive datatypes are disjoint, even in cases where the abstractions they represent might be thought of as having values in common." Thus the value spaces of xsd:hexBinary and xsd:base64Binary are disjoint, as both are primitive datatypes.
Regarding an annotation property providing an annotation for an IRI, this is correct. This is because punning can be used to associate the same IRI with more than one entity.
Regarding individual equality axioms, we do intend equality and not identity: stating that two individuals are equal means that they denote the same element, not that the individuals themselves identical (which they clearly are not).
Please acknowledge receipt of this email to <mailto:firstname.lastname@example.org> (replying to this email should suffice). In your acknowledgment please let us know whether or not you are satisfied with the working group's response to your comment.
on behalf of the W3C OWL Working Group
CUT AND PASTE THE BODY OF THE MESSAGE (I.E. FROM "Dear" TO "Group") INTO THE BODY OF AN EMAIL MESSAGE. SET THE To:, CC:, AND Subject: LINES ACCORDINGLY.
PLEASE TRY TO REPLY IN A WAY THAT WILL ALLOW THREADING TO WORK APPROPRIATELY, I.E., SO THAT YOUR REPLY CONTINUES THE THREAD STARTED BY THE ORIGINAL COMMENT EMAIL
* Section 3.1 says: The version IRI MAY, but need not be equal to the ontology IRI. That is missing the closing comma: The version IRI MAY, but need not be, equal to the ontology IRI. * Similarly, section 5.8 says: Each IRI I used in an OWL 2 ontology O can, and sometimes even needs to be declared in O ... but also needs a comma: Each IRI I used in an OWL 2 ontology O can, and sometimes even needs to be, declared in O ... * Section 3.1 says: 3.1 Ontology IRI and Version IRI ... The following list provides conventions for choosing ontology IRI and version IRI in OWL 2 ontologies. That should say: 3.1 Ontology IRIs and Version IRIs ... The following list provides conventions for choosing ontology IRIs and version IRIs in OWL 2 ontologies. (Per usage in RFC 3987, starting with its title ("Internationalized Resource Identifiers (IRIs)"), and per normal English patterns for acronyms, the plural of "IRI" is "IRIs".) * Section 3.1 says "side-effect." That seemingly should be simply "side effect." * Section 3.2 says: Each ontology document can be accessed from an IRI ... That's imprecise and possibly a bit confusing. A document can be accessed _using_ an IRI or _via_ and IRI, but it can't actually be accessed _from_ an IRI--an IRI is only a string of characters. Nothing except those characters can be accessed _from_ the IRI. I wonder if there's some other wording to use that's more correct (e.g., "... can be accessed _from_ a location denoted by an IRI") but not too wordy ("via an IRI"?). * Section 3.7 says: Either Table 2 or the prefix declarations of the ontology document being parsed MUST contain a declaration for pn: associating it with a prefix IRI PI. Something seems backwards about that wording. (The "must" sounds like a condition that applies to Table 2, so that if the condition isn't satified, Table 2 could be at fault.) It might be clearer to apply the "must" to the abbreviated IRI's prefix name (e.g., "the prefix name must be one for which there is a declaration contained in ..." (or a less wordy version of that)). * Section 3.7 says: Each part of the ontology document matching the prefixDeclaration declares ... That apparently should be: Each part of the ontology document matching prefixDeclaration declares ... or: Each part of the ontology document matching the prefixDeclaration production declares ... (or possibly: Each part of the ontology document matching the prefix declaration declares ... depending on other wording). * Section 3.7 says: The abbreviated IRI is split into a prefix name pn: — the part up to ... and The IRI <http://www.example.com/ontology1#> is associated with the prefix name : (this prefix is often called "empty" or "default"). Strings like that, especially those containing symbols frequently used for punctuation, going double for isolated symbols, really should be quoted. * Section 4 says: OWL 2 ontologies can refer to well-known data values such as strings or integers. Each kind of such values is called a datatype, and the set of all supported datatypes is called a datatype map. That and other wording that refers to "_a_ datatype map" seems to imply that OWL 2 supports multiple datatype maps, although other text seems to say that OWL 2 defines exactly one datatype map. * Section 4 says: The value space is a set determining the set of values of the datatype. Isn't the value space the set of values itself? If so, that wording should be simplified. (If it really is some other set that somehow determines the set of values, then that, especialy the determination, should be clarified.) * Section 4.2 says: Although floating-point numbers are numbers, they are not contained in the value space of owl:real. Should that say "Although floating-point values are numbers ..." (or "... represent numbers"), or maybe "Although values of type xsd:float and xsd:double represent numbers ..."? * Section 4.3 says: ... only basic language ranges [BCP 47] are normative in the rdf:langRange constraining facet. That sounds ambiguous. Usually, "normative" applies to text in a specification (saying which part "counts" and which doesn't), but the usage here is different and less clear. Does the above text mean: 1) that only basic language ranges (but not extended language ranges) are _required_ to be supported, 2) that only basic language ranges are _allowed_, or 3) that somehow only basic language ranges must be interpreted normatively (per the spec.) but extended language ranges don't have to be?) (The mentions of "normative constraining facets" might also be ambiguous.) * Section 4.5 says: ...: * xsd:hexBinary * xsd:base64Binary As specified in XML Schema [XML Schema Datatypes], the value spaces of these two datatypes are disjoint. Is that actually correct? The XML schema specification's section 220.127.116.11, Value Space (for hexBinary) says: The ·value space· of hexBinary is the set of finite-length sequences of zero or more binary octets. The length of a value is the number of octets. Its section 18.104.22.168, Value Space (for base64Binary) says, identically: The ·value space· of base64Binary is the set of finite-length sequences of zero or more binary octets. The length of a value is the number of octets. According that that wording, both the value space of hexBinary and the value space of base64Binary are the same set--the set of finite-length sequences of zero or more binary octets. Is the XML schema specification buggy (intending to make the two value sets disjoint but not wording those sections right (perhaps by defining each value set to have values _corresponding_ to (but not being) those sequences, so that the value sets weren't the same set))? * Section 4.7 says: A time instant may not contain a time zone offset ... That wording seems to be ambiguous, meaning either "a time instant _must_ not have a time zone offset" (it must not have) or "a time instant _might_ not have a time zone offset" (it might be that it does not have). (The latter meaning seems to be what was intended in this case.) * In section 5, in Figure 2, the association line between Entity and IRI has a navigability arrow at one end. That does not appear to be correct: a) presumbly, the OWL spec shouldn't be that concrete (talking about navigability in the first place), and b) the association is navigable in the other direction too--otherwise you wouldn't be able to give an IRI and have an OWL interpreter get to the entity or entities associated with that URI. (Also, can the other end of the association also give a cardinality to make clear whether one IRI can denote different entities (re "punning")?) * Section 5.2 says: An IRI used to identify a datatype in an OWL 2 DL ontology MUST * identify a datatype in the OWL 2 datatype map (see Section 4), or * have the xsd: prefix, or * be rdfs:Literal, or * not be in the reserved vocabulary of OWL 2 (see Section 2.4). IRIs from the reserved vocabulary MUST NOT be used to identify datatypes in an OWL 2 DL ontology, with the exception of rdfs:Literal, the IRIs of the datatypes in the datatype map, and the IRIs with the xsd: prefix . The second paragraph is confusing: (It seems that) it doesn't specify anything that the previous paragraph with its bullets doesn't already specify, but it doesn't indicate that it is intentionally restating implications of the first paragraph. If it is simply intentionally restating things, it should probably begin with "That is, " or something similar. * Section 5.5 says: Annotation properties can be used to provide an annotation for an ontology, axiom, or an IRI. Regarding "IRI": Is it really for an IRI, or for the entity denoted by an IRI? * Section 5.6.2 says: If an individual is not expected to be used outside an ontology ... That should be something like If an individual is not expected to be used outside a particular ontology ... (The current wording is somewhat ambiguous, possibly meaning "... outside any ontology...") * Section 7 says: Datatypes, such as strings or integers, can ... That wording is confusing because it abbreviates something too much. (Strings are not datatypes--clearly, a string is not a datatype.) That should probably say something like: Datatypes, such as the set of strings or the set of integers, can ... or maybe something referring to (some system's) names of datatypes, as in: Datatypes, such as the xsd:string or xsd:integer, can ... * Section 9.6.1 says: An individual equality axiom SameIndividual( a1 ... an ) states that all of the individuals ai, 1 ≤ i ≤ n, are equal to each other. Is that really "equal" or should that say "identical"? * Section 10.2.3 says: An annotation property domain axiom AnnotationPropertyDomain( AP U ) states that the domain of the annotation property AP is the IRI U. That wording and the object and association role names in Figure 23 seem to be confusing. Upon a first reading of that section 10.2.3 text, it appears that it inaccurately says the domain of the property is the IRI instead of saying the the domain is the entity that the IRI denotes. Then one can see that that occurrence of "domain" probably refers to the association role name "domain" of the structural object. However, that is confusing, since elsewhere the word "domain" refers to the domain of a properties (represented by a structural objects, specified by syntax). Additionally, there might not be any text saying whether the annotation is associated with the IRI or associated with the entity (or entities) denoted by the IRI. The terminology of section 10 should be reviewed and possibly revised.