Language code in description prefix should always be ISO639

[Twan Goosen]

Follow-up of #778

Some records have content language prefixes with codes that are not lower case ISO639 three letter codes. For example ​this record has

{lang='en-US'}various arabic dialects
{lang='en-GB'}Dictionary Gate is a growing collection of lexical resources
{lang='en-US'}a simple search interface to the dictionaries in the dict-gate

The web app used to assume a 3 letter code but this was relaxed in r6404. This fixes the problem in the view, but it would be nicer to use a unified code scheme, as is the case with the values of the languageCode facet, for easier processing in the front end (note: currently no such processing takes place).

Note: as far as I'm concerned, this scheme does not necessarily have to be exactly the 3 letters scheme that was assumed first. Maybe we want to represent language variaties as illustrated by the example above?

[teckart@informatik.uni-leipzig.de]

I don't see a reason why the region codes would be useful in the webapp. Therefore I think providing only ISO 639-3 codes is the best solution here.
That would leave the question if we want to stay with this value schema ("{lang='eng'}") or use the schema of the languageCode facet (including something like "{name:SOME_LANGUAGE_NAME}" for values that couldn't be mapped to ISO 639-3).

[Twan Goosen]

Replying to teckart@…:

I don't see a reason why the region codes would be useful in the webapp. Therefore I think providing only ISO 639-3 codes is the best solution here.

Agreed.

That would leave the question if we want to stay with this value schema ("{lang='eng'}") or use the schema of the languageCode facet (including something like "{name:SOME_LANGUAGE_NAME}" for values that couldn't be mapped to ISO 639-3).

I think that would be nice for the sake of consistency if nothing else. We should be able to reuse some code here (at both ends) :)

[teckart@informatik.uni-leipzig.de]

Fixed in r6653.

The output for the example record above is now:

{code:eng}various arabic dialects
{code:eng}Dictionary Gate is a growing collection of lexical resources
{code:eng}a simple search interface to the dictionaries in the dict-gate

If mapping to ISO 639-3 fails, the output would look like:

{name:XYZ}Description

Moving this ticket to VLO web app for adaptations there.

[Twan Goosen]

Since this change is in trunk and the adaptation in the webapp is expected to be minor, I'm moving this to the 3.4 milestone.

[Twan Goosen]

Marking this as criticial since the importer has been modified but the web app has not been adapted

[Twan Goosen]

Fixed in ​ee658bc.