Re: How we say Same literals

On 27 Dec 2014, at 11:02, Simon Spero <sesuncedu@gmail.com> wrote:
> If "the reader" is a user of a tool that needs to explain inconsistencies and not a person reviewing or reading a journal article, then explaining things in multiple steps is probably the right approach.
> 
> After explaining that the individual has too many values for the property, what the limit is, and where it came from, you can explain what the values of the property are, and where they came from.
> 
> Finally you can explain that 'the string «"2"»' is not the same as 'the integer «2»'.
> 
> It is probably not helpful to begin the explanation by saying that "leila is a person and not a person", which would be where the clash happens.
> 
This is unlikely. The clash happens with the attempt to merge the data valued nodes and thus is “2”^^string = “2”^^integer when ~( “2”^^string = “2”^^integer). I would be surprised if an implementation populated nodes with the inequalities instead of just checking on merge, i.e.,

def merge_data_nodes(node1, node2):
	if node1 == node2:
		return node1 #why not? there should be nothing to copy from node2
	else:
		throw “There’s a clash! Trying to merge non-equal data values!"

This is assuming you have your values in set up in the right way to use the programming languages equality. You might want to use a special function anyway.

>   Explanations meant for human consumption need to be tested on humans.
> 
> There is a lot of prior work on this topic from the 80s and 90s, which may serve as a guide.

There is a ton of recent work as well i.e., from the 2000s and 2010s. Including on various cognitive aspects and all done in an OWL context. E.g.,

M Horridge, S Bail, B Parsia, U Sattler, Toward cognitive support for OWL justifications
T Nguyen, R Power, P Piwek, S Williams, Measuring the understandability of deduction rules for OWL
TAT Nguyen, R Power, P Piwek, S Williams,  Predicting the Understandability of OWL Inferences
M Schiller, B GlimmTowards, Explicative Inference for OWL.

Plus a lot more on computing justifications. There’s loads more.

Cheers,
Bijan.