- From: <Simon.Cox@csiro.au>
- Date: Fri, 4 Nov 2011 17:03:21 +0800
- To: <mike@saxonica.com>, <xmlschema-dev@w3.org>
This is RDF/XML. -----Original Message----- From: Michael Kay [mailto:mike@saxonica.com] Sent: Friday, 4 November 2011 8:15 AM To: xmlschema-dev@w3.org Subject: Re: Is this reasonable: classify each element as either an Object or property On 03/11/2011 20:29, Costello, Roger L. wrote: > Hi Folks, > > I am thinking that it is useful to identify each element in a schema as either an Object or a property. > > Objects are described by one or more properties. > > The value of a property may be either simple or an Object. > > Here is an example of an XML document containing two Objects (Book and Person) and three properties (title, author, name): > > <Book> > <title>The Implementation of Functional Programming Languages</title> > <author> > <Person> > <name>Simon L. Peyton Jones</name> > </Person> > </author> > </Book> > > Notice that two of the properties (title and name) have a simple value. One property (author) has an Object value. > > Is this a reasonable desire? > > /Roger > > I think it's a bit 'perfectionist' and can lead to unnecessary and inconvenient extra layers of tagging. It can also lead to unproductive debates about semantic distinctions that have little practical importance. For example, this kind of approach, when taken to extremes, can lead to <Person> <name>....</name> <phone>01189482766</phone> </Person> being replaced by <Person> <name>....</name> <phone> <TelephoneServiceEndpoint> <supplier> <PhoneCompany> <name>BT</name> </PhoneCompany> </supplier> <number> <PhoneNumber> 01189482766 </PhoneNumber> </number> </TelephoneServiceEndpoint> </phone> </Person> Don't laugh: I see exactly this kind of bloat in many standardized XML vocabularies such as FpML. Michael Kay Saxonica
Received on Friday, 4 November 2011 09:04:08 UTC