Suggestion to add an additional layer in noSkE for top semantic domains only

[kristinapdm]

To further enhance the user-friendliness of the corpus metadata categories, it could be nice to have the opportunity to quickly look at the 21 top semantic domains only in noSkE by adding a separate text type layer to the following structure:

[TomazErjavec]

This probably makes sense but using USAS labels is somewhat tricky:

They come in 3 variants:

• "USAS tags" is how UCREL annotated it, with the tags quite complex, e.g. Df/A5.1+++mfnc
• "USAS glosses" simplifies the tags and adds their gloss
• "USAS categories" are the IDs that are used in ParlaMint in the USAS Taxonomy - these have the putative advantage that it is easier to search by them, as they don't have "+" but "p", so you don't have to quote it in RegEx.

They are marked up 2 levels: on structural phrases for multiword semantic units as well as on the token level.

So, any interpretations over this data need to be careful.
That said, if it would be of use, I think having just top level categories would make sense.
What about, rather than introducing yet another tag, to take - the maybe somewhat useless - "USAS categories" both on phrases and tokens and make it like USAS glosses, but for the top level?

[kristinapdm]

Yes, I believe that having easy access to top-level categories only would be a nice addition. I would imagine that USAS glosses are the most used variant anyway, so reusing USAS categories to provide a more easily interpretable output is probably a good suggestions. Thank you for considering this.