Consider allowing internal >1 whitespaces in strictish-string

[sdruskat]

Reported by @hainesr via email:

I think it makes us backwards incompatible because in some cases, such as say the abstract field, someone may have used double spaces after a period (e.g. “This tool is great. Please use it.”) if they’ve copied from a manuscript or something. It may catch people out and we may be being too strict(ish). I see the value of “strictish-string” for things like postcodes though so perhaps we can only use it in those cases. In any case, I think if we’re going with Semantic Versioning then I think that means we have to go to 2.0.0.

A solution to this would be to define the pattern as ^(\S)(.*\S)?$ (^(\\S)(.*\\S)?$ in the schema):

• A string starting with a non-space character followed by an optional group of characters that ends on at least a single non-space character.

[hainesr]

I'd be more tempted to keep strict-ish string as it is, but not use it everywhere. Keep free-text fields as just strings?