[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: xml:href, xml:rel and xml:type

• From: Michael Kay <mike@saxonica.com>
• To: xml-dev@lists.xml.org
• Date: Mon, 16 Apr 2012 18:03:45 +0100

>> If that's the case, then I'm sceptical about
>> whether this information belongs in the XML.
> Why are you sceptical?
>
Remember third normal form?

My postcode is RG4 7BS. That's a relationship, between me and where I 
live. If I want to record my postcode in an XML data file, that is the 
way I should record it. If someone wants to use that fact to get 
information about that postcode from a web service provide by the UK 
Post Office, or the Ordnance Survey, or Google, that's a matter for the 
user of the data and/or the application. The address of the web service 
and the domain name of its provider doesn't need to be stored 
redundantly with every postcode. To do so is simply bad normalization; 
and it violates the objectives of normalization, which is to maximize 
potential for change. If the application decides to get its information 
about postcodes from Google rather than from the PostOffice, why should 
that mean changing the address data for every customer?

Michael Kay
Saxonica

• References:
  • xml:href, xml:rel and xml:type
    • From: "Rushforth, Peter" <Peter.Rushforth@NRCan-RNCan.gc.ca>
  • Re: xml:href, xml:rel and xml:type
    • From: Michael Kay <mike@saxonica.com>
  • RE: xml:href, xml:rel and xml:type
    • From: "Rushforth, Peter" <Peter.Rushforth@NRCan-RNCan.gc.ca>
  • Re: xml:href, xml:rel and xml:type
    • From: Michael Kay <mike@saxonica.com>
  • RE: xml:href, xml:rel and xml:type
    • From: "Rushforth, Peter" <Peter.Rushforth@NRCan-RNCan.gc.ca>