0-1 or 0-n string bodies

[azaroth42]

From #79, there was no consensus around how many strings could be used as the body of an Annotation. The current WD has 0-1, but there was unresolved discussion as to whether there should be an array of strings.

My proposal is close wontfix (e.g. use the current 0-1) -- the advantage of the string body is that it has a completely flat structure with no sub-resources (objects) or lists (arrays). By introducing complexity in the simplest case, we reduce the simplicity and thereby the benefit of allowing this pattern at all.

[iherman]

At the moment, there is a discrepancy. For body the text says

There SHOULD be 1 or more body relationships associated with an Annotation but there MAY be 0.

whereas for bodyText, it says:

There MAY be exactly 1 bodyText for an Annotation,

The only argument I have seen in favour of this restriction on bodyText is simplicity. However, due to the possibilities for body, implementations already have to deal with the duality of array vs. single object when handling the bodies, so this handling can easily be transferred for multiple bodyText, too. In other words, the restrictions seems to be fairly arbitrary to me.

It is not a major issue because I do not believe there will be tons of annotations with multiple textual bodies. Ie, if the majority wants to keep the restriction, I will not lie down on the road:-). But spec wise it looks fairly asymmetric to me.

[hugomanguinhas]

Hi both,

A bit side question, what happens to the oa:hasBody, will it still be possible to have both text (Literal) or a Resource?

[iherman]

Decision on telco, 2016-02-19: keep to 0-1

See http://www.w3.org/2016/02/19-annotation-irc#T16-34-51

[akuckartz]

For the record: I think this restriction is arbitrary.

[azaroth42]

It is intentionally arbitrary to encourage people who think they need this functionality to actually use the TextualBody approach :)