Support for case-insensitive value checks

[andyward]

This has been discussed before, but it's not clear to me why we have determined that all value comparisons must be case sensitive in IDS. E.g. test cases such as fail-attributes_should_check_strings_case_sensitively_2_2.ids
and fail-specifying_a_value_performs_a_case_sensitive_match_2_2.ids

While I can see there may be some scenarios where it makes sense when dealing with fixed schema values (e.g. #49) it feels like we're making life harder for both IDS and Model authors by not allowing some allowance for case differences - particularly on things like Attribute Values, PropertyValues, Classification values etc. where these are typically free text fields.

Take this simple example from a client's EIR/BEP, where they want to specify floors should have a consistent description:

<ids:specification ifcVersion="IFC2X3" name="04.05 : Building Storeys should have standard Description">
		<ids:applicability minOccurs="1" maxOccurs="unbounded">
			<ids:entity>
				<ids:name>
					<xs:restriction base="xs:string">
						<xs:enumeration value="IFCBUILDINGSTOREY" />
					</xs:restriction>
				</ids:name>
			</ids:entity>
		</ids:applicability>
		<ids:requirements>
			<ids:attribute cardinality="required">
				<ids:name>
					<ids:simpleValue>Description</ids:simpleValue>
				</ids:name>
				<ids:value>
					<xs:restriction base="xs:string">
						<xs:enumeration value="Basement Level 2" />
						<xs:enumeration value="Basement Level 1" />
						<xs:enumeration value="Base level of building (Ground Floor)" />
						<xs:enumeration value="First Floor" />
						<xs:enumeration value="Second Floor" />
						<xs:enumeration value="Third Floor" />
					</xs:restriction>
				</ids:value>
			</ids:attribute>
		</ids:requirements>
</ids:specification>

I can see model authors questioning the value of being asked to change 'Basement level 1' to 'Basement Level 1' when there may be many other high-value issues in the information. At the moment as an IDS author I'd have to express this as a Regex to allow some degree of tolerance.

Should we consider a case sensitivity option on values? Perhaps something like:

<ids:value stringComparison="ignoreCase">
  <xs:restriction base="xs:string">
    <xs:enumeration value="Basement Level 2" />
... etc

Obviously this would only take affect on string values, and I get there may be internationalisation concerns for Unicode etc (but that's a solved problem?)

[atomczak]

We are after consistent and qualitative data and 'Basement level 1' is not equal 'Basement Level 1'. I'm personally against case insensitivity.

[aothms]

Tend to agree, you can use a regex and be really explicit about it if needed.