FtanML

Daniel Sullivan asked, in a post which I've only just seen,

 >could you make a comment on the relationship/differences/utility of 
FtanML vs. uxml?

FtanML was a student project to design a simple markup language. The aim 
was to have the simplicity of the JSON model, but with the addition of 
mixed content; and to do the mixed content without the verbosity of XML. 
We decided that compatibility wasn't going to be a constraint. In the 
end I think it's a very pleasing integration of JSON and XML, and it has 
the nice feature that the data model is pretty much JSON (the mixed 
content ends up just being lexical sugar, represented beneath the 
surface by JSON arrays of elements and strings). It improves on JSON in 
some respects and it also improves on XML in many respects. The fact 
that it's not compatible with the rest of the XML eco-system gives it 
little chance of survival in the wild - except perhaps to people who 
like JSON but recognize its limitations.

MicroXML I think has different objectives, most notably the aim to be a 
subset of XML and therefore to work within the XML ecosystem (XPath, 
XQuery, schema languages, etc). I worry that MicroXML isn't radical 
enough to make it worth adopting: it's relatively low pain, low gain, 
but the low gain doesn't justify the cost of adoption. FtanML by 
contrast is too radical: it's high gain, high pain.

Michael Kay
Saxonica

Received on Thursday, 30 August 2012 16:31:28 UTC