Legal characters in a full identifier

[timgdavies]

We have agreed that prefixes should stick to a limited set of characters.

But what about the identifier portion itself?

For example, what if a source identifier contains extended unicode characters, punctuation (e.g. ' or / or +) which might not be valid in some systems.

Should we propose a substitution approach for these cases? Or accept that such data is valid?

[timgdavies]

In this IATI thread the question of identifying sub-units of an organisation via their identifiers has also come up (e.g. identifying Programme Implementation Units)

This is likely to be a common need. One approach would be to introduce a further separator to the organisation identifier allowing a portion of the identifier for department.

E.g.

GB-COH-12345678/deptA