Hacker News Reader

Comment by jhgb | Hacker News Reader

It would seem reasonable to me to use the latinized form of the Chinese word as part of the identifier instead of the awful descriptive identifier.

View on HN

I honestly don't know if I would prefer CNLabelContactRelationYoungerCousinMothersSiblingsDaughterOrFathersSistersDaughter or CNLabelContactRelationBiǎoMèi (is unicode even allowed?). And I probably butchered that romanization and upset half the internet in the process.

If unicode is allowed, you can also use the Chinese characters

Then about 80% of the world wouldn't know how to pronounce it.

A quick test in google would suggest that CNLabelContactRelationBiaoMei doesn't introduce too much ambiguity (first result for Biao Mei is still about the correct term), so even without unicode it might be fine.

It's not specific to one Chinese Language or even just Chinese.

Surely it has a more generic name, then.

think about what you’ve just said

If you’re suggesting the current identifier is already latinised, it’s not, it’s translated.

I've basically just said that, e.g., instead of the identifier EmployeeTheNumberThatIdentifiesATupleInARelation, you should really be using the identifier EmployeePrimaryKey. That is, you usually refer to things using the name of the respective concept, not by describing what the thing's concept is every time you're trying to refer to the thing. It's called "coining a term". Or, in programming, abstraction (as per SICP, that is giving a name to some compound construct so as not to be forced to repeat it endlessly).

I would like you to figure out what that would be and post it here.

Reply on news.ycombinator.com