On 10 December 2017 at 10:41, Gerard Meijssen <gerard.meijssen(a)gmail.com>
This tool has its attractions but it does entrench
cultural dominance even
further as it stamps "quality" on items. The items with the most statements
are the ones that are most likely to get good grades. Items from Indonesia
let alone countries like Cameroon or the Gambia are easily substandard.
As quick reply I would say don't shoot the messenger :-). If Recoin finds
that some of the most important Indonesian national heroes are less well
documented than some minor league Baseball players then....that is firstly
food for thought.
I agree though that labels like "basic information" and red symbols can
create an unintended negative perception, but I have no idea how to do it
differently (albeit from lowering the thresholds). It could indeed be that
many of the red items are already very useful as they are, yet, who could
decide this and create necessary annotations? Thus the idea of Recoin is to
show a *relative* completeness/informativeness based on how much is known
about other, similar entities.
The question is: do you dwell on what is
"good" or is it a tool that
helps to improve things?
An expandable list of things to improve should be loaded directly below the
page title (unless it is pushed down by other plugins/scripts)?
On 10 December 2017 at 10:24, Fariz Darari
I was wondering: is there any correlation between the number of triples
entity with the entity's Recoin completeness indicator?
Thanks for the interesting analysis! Yes, there should be some positive
correlation, yet the way we compute the indicator (look at avg frequency of
top 5 missing properties) it is much more important that the frequent
properties are there, than that many are there in general. This is not
scientifically grounded though (I am not aware of any scientific work
discussing how one should compute such indicators), so what we compute is a
simple heuristic, and we are glad for any suggestions for different