- From: Pat Hayes <phayes@ihmc.us>
- Date: Fri, 15 Jan 2010 11:43:24 -0600
- To: Jeremy Carroll <jeremy@topquadrant.com>
- Cc: Michael Schneider <schneid@fzi.de>, Jiří Procházka <ojirio@gmail.com>, Semantic Web <semantic-web@w3.org>
+1. We have to be very careful about what exactly "optional" is supposed to mean in a standard whose intention is to support interoperability. If every processor is obliged to be able to process it, its not optional. If they aren't, then interopreability is compromised. Pat On Jan 15, 2010, at 11:23 AM, Jeremy Carroll wrote: > Michael Schneider wrote: >> >> r own specs as /optional/ features of RDF. >> This reminds me to OWL 2, where the concept of an n-ary datarange >> is now >> also an optional feature, and there is a first type of n-ary >> dataranges, >> namely linear equations, available in its own document [1], maybe >> others >> will follow over time. This strategy settled some longish and heated >> discussion within the OWL Working Group about whether n-ary >> dataranges >> should go into the core language or not, and if yes, which and to >> what >> degree, etc. >> >> [1] <http://www.w3.org/TR/2009/NOTE-owl2-dr-linear-20091027/> >> >> > > I think optional features are, in general, a bad idea. > They are technically flawed solutions to political problems. > If there is not consensus that a feature should be in, then it is > non-standard, and should not be included. > > See > http://www.w3.org/TR/qaframe-spec/#option > "The greatest way to undo the utility of a specification is with too > many optional features." > > Jeremy > > > ------------------------------------------------------------ IHMC (850)434 8903 or (650)494 3973 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32502 (850)291 0667 mobile phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes
Received on Friday, 15 January 2010 17:43:54 UTC