- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Wed, 08 Oct 2008 14:38:50 -0400 (EDT)
- To: public-owl-wg@w3.org
Here are my responses to the current set of review comments. My changes are http://www.w3.org/2007/OWL/wiki/index.php?title=ManchesterSyntax&diff=13815&oldid=13780 I've removed comments that have been dealt with. I've left in comments that may need further work or concurrence from the reviewers. peter ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|What OWL 2 tools may or may not do is not part of status.}} OWL 2 tools may implement extensions to the Manchester Syntax, and may only implement part of it, for example only the descriptions. Removed the sentence ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:04, 8 October 2008 (UTC)|Direct hyperlinks to P4 and Topbraid should be at the bottom of the page, in the references section}} I'm not completely convinced of this. These are not references to document, but instead are references to projects. Does anyone else have anything to say on this? ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:04, 8 October 2008 (UTC)|Propose a slight rewrite to improve readability (e.g. four sentences start with "The Manchester OWL syntax...") }} Edited the paragraph to make it more readable. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|The use of labels to to replace URIs is central to productive use of Manchester syntax when URIs are not meaningful, as is common in many ontologies. In addition it is often recommended that, in general, URIs not have meaninful information encoded in their strings (see e.g. http://www.w3.org/DesignIssues/Axioms.html). Given this and the goal of making the Manchester syntax readable and user friendly, this specification should say precisely how to use labels in this way, while providing, as a backup, the ability to use URIs when they are not available.}} I still view this extension as outside the scope of this document. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|Backwards reference to OWL 1 not necessary. Syntax "represents", rather than "stores". Perhaps replace starting "The Manchester OWL syntax stores" with: "Despite the fact that it adopts a frame-like syntax, parsing the Manchester OWL syntax into the OWL 2 structural specification is quite easy, as it is easy to identify the axioms inside each frame. }} ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:04, 8 October 2008 (UTC)|That the Manchester syntax is readable for non-logicians is not its most distinguishing feature wrt. the other syntaxes. I don't suppose logicians would find RDF/XML more readable than the Manchester syntax (and, logicians seems a bit arbitrary, why not computer scientists?). I understood from earlier comments by [[User:AlanRuttenberg|Alan Ruttenberg]] that the syntax is meant to overcome accessibility issues as a whole: e.g. it is suitable for screen readers. Perhaps this accessibility issue could be stressed more.}} Rewrote the section to talk about the frame-like aspect of the Manchester syntax and how it is different from the OWL 2 syntaxes. Kept the reference to the OWL 1 abstract syntax, for its frame-like nature. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:28, 8 October 2008 (UTC)|BNF grammar should be explained by e.g. a pointer to http://www.w3.org/2007/OWL/wiki/Syntax#BNF_Notation}} Added a bit about the BNF used here, stolen from Syntax. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:28, 8 October 2008 (UTC)|Add reference to unicode and UTF-8 specs}} Added a reference, but what it the desired document? ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|Better to indicate these clearly at point of use. It's not clear what "keyword use dominates" means.}} Better to indicate this at the beginning of the grammar section. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|The Curie syntax is in last call. We should plan on citing it as normative, perhaps leaving an editor's note to that effect.}} Added reference to CURIE document. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|Why {only} non-URIs for these. ([[User:RinkeHoekstra|RinkeHoekstra]]: refers to definition of datatype) What about: number, real, real+, double, decimal, long, int, short, byte, datetime, text, etc..}} The short forms are for the "common" datatypes. It might be useful to add more, but it is not necessary. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|Do we want the trailing f for 1e10f, and 1e10 to be disallowed?. What about doubles? NaN, -0, +/- INF}} Allowing no "f" could be done, but it is not necessary. The unusual values are all allowed in the long syntax. ** {{Review|[[User:MikeSmith|MikeSmith]] 20:25, 7 October 2008 (UTC)|The grammar above could be simplified by referencing the Syntax document for terms that are defined identically, such as URI. Since such references are present for terminals with other dependencies, such as full-IRI, referencing Syntax seems to make sense.}} This document tries to be self-contained. ** {{Review|[[User:MikeSmith|MikeSmith]] 20:33, 7 October 2008 (UTC)|The grammar for typedLiteral and stringLiteral could be aligned with the grammar in Syntax. This might reduce the likelihood of confused readers.}} ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:14, 8 October 2008 (UTC)|Agree with Mike's comments on alignment with other syntaxes.}} Done. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:14, 8 October 2008 (UTC)|Wrt. datatypes (integer, decimal, float and string): these are probably mapped onto standard XML Schema datatypes, but this is left implicit.}} Added sentence at beginning of section. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:28, 8 October 2008 (UTC)|Inconsistent use of upper/lower case: full-IRI (previous section) vs. Full-IRI here}} Fixed (full-IRI). ** {{Review|[[User:MikeSmith|MikeSmith]] 20:40, 7 October 2008 (UTC)|Is rdf:langPattern intentionally missing or is pattern overloaded?}} Unintentionally missing. Added. pattern can't be overloaded as it is used in internationalized string. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|Will we not follow the lead of functional syntax and have facets be URIs?}} Good question. No, as <=, etc., need translation anyway. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|Missing pipe(or) separating "that not" from "and not"}} If you mean the first alternative from conjunction, then this is not an alternative. ** {{Review|[[User:AlanRuttenberg|AlanRuttenberg]] 05:10, 8 October 2008 (UTC)|The OWLED paper on Manchester syntax defines an "onlysome", "xor", and "ValuePartition" "macros" not listed here.}} Currently not in this version. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:28, 8 October 2008 (UTC)|Find the 'miscellaneous' in the title a bit imprecise, but can't come up with a better name myself.}} Any suggestions? ** {{Review|[[User:MikeSmith|MikeSmith]] 21:03, 7 October 2008 (UTC)|KeyFor changed to HasKey in the other syntaxes. Can it be changed here?}} Not only "can", but "it produces a nicer syntax". Done ** {{Review|[[User:MikeSmith|MikeSmith]] 21:00, 7 October 2008 (UTC)|Could this paragraph be simplifed by saying that the Manchester syntax is only valid for ontologies satisfying the typing constraints defined in Syntax?}} Simplified, or at least changed. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:32, 8 October 2008 (UTC)|Example ontology should import Dublin Core definitions. The dc:creator data property should probably be replaced with some made-up property.}} I don't think that DC is importable into the Manchester syntax. This example tries to show how DC can be cleaned up. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:37, 8 October 2008 (UTC)|Typo: betwen->between. Consider rephrasing 'translating frames axioms'.}} Done. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:37, 8 October 2008 (UTC)|What does 'almost directly' mean? What would the mismatch be?}} Rewritten. The formal bit shows how to fix up the remaining stuff. ** {{Review|[[User:RinkeHoekstra|RinkeHoekstra]] 13:42, 8 October 2008 (UTC)|Strings with language tags should be mapped onto literals of type rdf:text, see [[Syntax#Strings]]}} They do - as x@y is a literal of type rdf:text
Received on Wednesday, 8 October 2008 18:39:42 UTC