Home » railML newsgroups » railml.common » [all versions] Using sequences in XSD schemas
[all versions] Using sequences in XSD schemas [message #2714] Thu, 06 May 2021 22:37
Thomas Nygreen is currently offline  Thomas Nygreen
Messages: 66
Registered: March 2008
Member
Dear community!

A topic that arises from time to time is the use of xs:sequence vs. xs:all in the railML XSDs. Together with xs:choice, these are three options for how to attach child elements to a parent. I'll try to sum up their differences briefly:

sequence:
* the multiplicity of the children is specified individually
* children must appear in the same order as in the XSD
* supports xs:any

all:
* the multiplicity of the children can only be 0..1 or 1..1, and is specified individually
* children can appear in any order
* does not support xs:any

choice:
* designed for selecting one child from a set of possible children

The most common background for these questions is that with xs:sequence, the order of the child elements has to be the same as in the schema file, while with xs:all the order is arbitrary. This may seem simpler, but it leads to ambiguities more easily than with the stricter xs:sequence. That is why xs:all neither supports xs:any or children with multiplicity greater than 1. With the restricted multiplicity, xs:all is mostly not suitable for our purposes, apart from in the root element and the "views", where no children have multiplicity greater than 1.

Since xs:all has this limited functionality, and since mixing xs:all and xs:sequence depending on the multiplicity of the children will most likely confuse developers using railML as to when the order of the children matters, the coordinators recommend that we keep using xs:sequence for railML.

As a side note, there is one element in railML 3 element with xs:all, and that is the <railML> root element, allowing an arbitrary order of the top-level elements for each schema. So you can have <timetable> before <infrastructure> if you want. I don't know why, but it is there already, so we should probably not change it.

Best regards,
Thomas Nygreen


Thomas Nygreen – Common Schema Coordinator
railML.org (Registry of Associations: VR 5750)
Altplauen 19h; 01187 Dresden; Germany www.railML.org
 
Read Message
Previous Topic: Suggested refined definitions and extension to organizationalUnits
Next Topic: Passing arbitrary attributes / use of "any" fields
Goto Forum:
  


Current Time: Thu Mar 28 13:22:09 CET 2024