[regext] null vs empty string: was Re: rdap mobile app avail

["Marc Blanchet" <[e-mail]>]

On 29 Aug 2019, at 4:11, Marc Groeneweg wrote:

> So, we have our implementation live! But how can we make sure (if not 
> ICANN but at least the community) what compliancy means... (e.g. we 
> return a null when we don't have a value and "" when we have an empty 
> string value. The mobile app wants "" in all empty situations... Just 
> an interpretation?).
>

There are many cases. Let’s try to cover them.
a) the server has no value to send. sends empty string as value
b) the server has no value to send. sends null as value
c) the server has no value to send. do not send the member and its value
c) the server has value to send but because of policy (such as GDRP or 
else), the value has to be redacted. In this context, it may want to 
tell the user the data was redacted

I’ve seen all a),b),c) and d).  I’ve only seen b) on your server, 
but that is not the point. let’s try to find the right answer.

As Andy was saying, to me, the server should do c) whenever possible.
However, some members are required by the standard or by policy (icann 
profile). Then either we do a) or b).

I guess RFC7483 and RFC7095 are almost silent on what to do. Except for 
ADR fields:
For ADR member in jCard, RFC7095 says: « Its equivalent in jCard is a 
structured
    property value, which is an array containing one element for each
    text component, with empty/missing text components represented by
    zero-length strings. »

However, some members are redacted and need to be signaled to the user. 
Right now, it is pretty bad, as people are sending all kind of strings 
in caps letter, in one language (en). Given that it is not normalized, I 
can’t find it was redacted and as a client show it nicely. I think we 
need to define some normalized attribute to signal it was redacted.

Marc.