Work out how to deal with inflected versions in English
Chinese is easy for lookup and translation. Inflected languages not so much. There are several ways and places to implement the logic for searching here, including:
- only ever using lemmatised forms
- having multiple versions stored in data storage
- searching for all forms and then merging before sending to client, or let clients decide how to present multiple versions
The problem is that various dictionaries treat this differently.
ABC Eng to Zh, for example, has both the 'infl'... ' of ' which indicates it is an inflection of the word. Alas, these entries can also have their own, separate definitions, so that is problematic. Ditto with 'subof' for ABC.
Bing will return largely identical results for an inflected form ('speak' vs 'speaks'), at least on the lookup API.
All this needs to thought - there are many ways to do it but the way it is done needs to be optimised for what the learner actually needs. And as this may depend on the level of the learner or even the particular word or context (brocrobes vs notrobes), this is a particularly hard (and critical) issue to get right.