Consider having one view for each unit of learning

[gustafl]

In issue #120, we described the possibility of showing the document in alternative views with different highlights. Now, with the four units of learning almost decided upon, we may actually need this feature to separate potentially overlapping highlights. For instance, when coming across a compound word, the user may want to register both the compound and its constituent lexemes, this would be an overlapping selection. Handling overlapping span elements would be too complex.

There are a number of downsides to this proposal. It adds a layer of complexity to the whole solution. It forces the user to move between different views. And there would be no unified view where the user could see the sum of known words for a certain document. These are serious downsides.

[gustafl]

Actually, there might be a compromize to be found here. We could have a unified (and default) view if we let some units of learning take precedence over others. For instance, we could let expressions take precedence over compounds, and compounds take precedence over single words (lexemes and inflections). In the case there is potential for two overlapping spans in the document, expressions take precedence over compounds and compounds over single words.

But if the user wants to concentrate solely on the lexemes, there is a view for that too. In theory, this translates to a minimum of 3 views:

• Unified view
• Single word view
• Compound word view

But separating compounds from expressions is probably not so important, so we should get by with only the first two.

[gustafl]

As noted in issue #4, we can't have a combined lexeme/inflection view. So, inflections have to go on the unified view. Thus, we end up with:

• Unified view
• Lexeme view