Bug: Activities with @humanitarian="true" are not labelled as humanitarian

[andylolz]

This activity has @humanitarian="1", and is flagged as humanitarian.

This activity has @humanitarian="true", but is not flagged as humanitarian. (Note that @humanitarian="true" is also valid XML. In fact, it’s the canonical representation for a boolean!)

It looks like this needs fixing in activities.css:
https://github.com/devinit/D-Portal/blob/45561c0f05c1fe26de670d7a78864063c596a51b/ctrack/art/high/activities.css#L41-L48

[notshi]

Many thanks, @andylolz. We're working on SAVi 2.0 right now and will add that to the list.

[notshi]

I feel like it would make more sense to keep this as an explicit codelist; ie. 0 or 1, not 4 possible values.
And maybe stick to that approach consistently throughout the standard.

At the current state, it catches you off guard with it's native xml eccentricities.

[andylolz]

I know you know all this, but just to state it here:

It wouldn’t be enforceable at schema level, since true/false are still valid booleans. IATI could recommend using 0/1, but some publishers would still use true/false so you’d still need to handle the edgecase anyway. Plus, just the thought of a long discourse thread proposing this as guidance makes my eyes water. <Duncan Bannatyne voice> For that reason, I’m out!

[notshi]

Ack - I had to search who that was!

Maybe we could get some progress with the new validator people and have it implemented as a suggestion rather than a rule :}

[stevieflow]

Hi @andylolz @notshi - thanks for this

We've some evidence here in terms of the use of 1/0 & true/false (and also True/False it seems) in IATI booleans: http://dashboard.iatistandard.org/booleans.html

Note - DFID have just started to include humanitarian="true" in their data, so it'd be quite useful to see that flagged in d-portal

I'm with @andylolz - so long as you supply a valid boolean, we should be able to use your data

NB: @davidmegginson might be able to point to the is_truthy() function that deals with a range of booleans...

[andylolz]

Thanks, @stevieflow.

We've some evidence here in terms of the use of 1/0 & true/false (and also True/False it seems) in IATI booleans: http://dashboard.iatistandard.org/booleans.html

^^ Worth noting that this dashboard page lists publishers using False and True (i.e. title case). These are invalid values – false and true must be lower case according to the XML spec. I’ve made a ticket for that here: IATI/IATI-Dashboard#543

[stevieflow]

Denmark use humanitarian="true", I have discovered today: http://d-portal.org/q.html?aid=DK-1-128253-30530

[notshi]

Many thanks, both - this is now fixed.