- From: Mike Dean <mdean@bbn.com>
- Date: Mon, 12 Feb 2007 22:29:31 -0500
- To: <public-rif-wg@w3.org>
Below is my review of the various pages making up the current draft of RIF Core. Unfortunately, I probably won't be on tomorrow's telecon. In general, I think the current document is usable by rules experts but not by typical software developers. http://www.w3.org/2005/rules/wg/wiki/Core The Dynamically (slowly) generated W3C-styled version indicates a number of formatting errors. I would have preferred to review this version. http://www.w3.org/2005/rules/wg/wiki/Core/Overview This needs text as well as pictures. http://www.w3.org/2005/rules/wg/wiki/Core/Conditions explicated? http://www.w3.org/2005/rules/wg/wiki/Core/Positive_Conditions floats => float [XML Schema] primitive datatypes the possible => possible didactic? Non-terminal names such as CONDIT and DISJ seem unnecessarily terse. Full words would be preferable. Why are some non-terminals all upper case? Is Con short for Constraint? Add link for Section "Multisorted Extensions", like Section Horn Rules It seems premature to give examples and a DTD for RIF Horn Rules in this section. RDF would be a more appropriate camelcase analogue than Java upercase => uppercase The formatting of the box starting with "- Con (constant individual, function or relation)" is not clear. also known as [an] interpretation? will be based => may be based based [on] multivalued logics four mapping[s] comprising I RIF core => RIF Core? (several places) sometimes also go under the name of => are also known as integer, strings => integer, string furthermore => also Add links for "Multisorted Syntax for Primitive Sorts" and "Formalization of Multisorted RIF". dataTime => dateTime Why not use sort rather than type as the attribute name, and avoid introducing another term? xsd:decimal is arbitrary precision. xsd:float is IEEE 32-bit. The types are unrelated. _sateTime => _dateTime xsd:dateTime also allows a form such as "2006-11-20T13:00:00Z", which is used to represent UTC/GMT. mionutes => minutes needs the following [semantic] extensions? character for "element of" (?) in "c ? Con", etc. doesn't render in Internet Explorer 7 (not sure about other browsers) happenstance? Unrendered characters in Ds1 ? ... ? Dsn It seems a bit odd to modify the semantic structures introduced only a few screens back, as if multisorted RIF was an afterthought. I'm concerned that this section tries to cover too much material, with abstract and concrete syntaxes and semantics. RDF and OWL put these in separate documents. http://www.w3.org/2005/rules/wg/wiki/Core/Horn_Rules RIF Validation is a Section rather than an Appendix in the Table Of Contents. Should link to Semantic Structures section, not just Positive Conditions. http://www.w3.org/2005/rules/wg/wiki/Core/RIF-OWL_Compatibility final text missing http://www.w3.org/2005/rules/wg/wiki/Core/RIF-RDF_Compatibility final text missing http://www.w3.org/2005/rules/wg/wiki/A.5_Extension%3A_Ontology_Conditions Integer => integer XML Part 2, Datatypes => XML Schema Part 2: Datatypes http://www.w3.org/2005/rules/wg/wiki/Core/Validation Do we want to promote the use of DTDs? I see they are still part of XML 1.0 (Fourth Edition). DTDs from www.jdrew.org should be copied to and referenced from w3.org space before publication. such as the one of => such as Mike
Received on Tuesday, 13 February 2007 03:29:40 UTC