I actually see a lot of utility for ixml in rdf, especially if you use custom simple types such as "rgba(64,128,64,0.5)"^^Functions:_ColorFunctions as atomic values. This could be mapped to an ixml parser that would generate:
<map>
<number key="red">64</number>
<number key="green">128</number>
<number key="blue">64</number>
<number key="alpha">0.5</number>
<string key="function">rgba</string>
</map>
Which then can be translated with any xslt processor to {"red":64,"green":128,"blue":64,"alpha":0.5:"function":"rgba"} as JSON or [:red 64; :green:128; :blue:54; :alpha 0.5; :function Function:_rgba] as Turtle.
What's the value in this? In a knowledge graph such functional attributes are difficult to parse, especially on ingest. An ixml parser could reduce re this complexity, especially across heterogeneous requirements.
