language specific classifiers

[kjdon]

User request. John Rose and Zineb Naji

Documents have metadata assigned in different languages. eg an English Title and a French Title. Want to have a title list that just has English titles in it, and one that just has french titles in it.

Two ways to achieve this metadata:

1. separate metadata elements for each language, eg dc.Title_en, dc.Title_fr. Can be used right now by modifying a metadata set.
2. GLI needs modifying to have a language option for each metadata element. So can add an english and a french version of dc.Title.

metadata.xml file would need a language attribute for each piece of metadata. Would be fairly easy to modify the metadata set dtd, but reasonably complicated to modify GLI.

To get the classifiers, in case 1, just build separate classifiers on each metadata element. In case 2, would need to modify the classifiers (probably just List as if you are working in different languages then List is the best) to have another language argument, and only use metadata in that language. The test would need to be on the language of the metadata, rather than the language of the document? Or maybe this would be another option?

To have language attributes for metadata would involve quite a lot of changes to perl code and runtime code.

Then, once there are separate classifiers for the languages, then in the navigation bar, we want to display the appropriate one for whatever language the user is in. ie in English, the titles list will be the english titles list. in French, the titles list will be the french list.

Can this be done semi dynamically, without the user having to write the navigation bar macro? They would need to indicate which classifiers belong together, which belongs to which language and which is the default. or something like this.