Identify groups

[blueCat11]

The composers in our corpus are not all formatted the same. An example for Beethoven:

While this is not necessarily a problem in our search results, because the musicologist user can choose which pieces are relevant for their analysis, in the distant hearing analysis, this will result in many "different composers" when comparing by composer, which are not actually different. Instead they are the same, but specified in a different way.

We need to decide how to act about this!

[blueCat11]

My idea would be to present the user with the possibility to specify groups that they want to compare. That way, we kill two birds with one stone:

• We make our tool very versatile, because musicologists can specify groups according to categories we don't know
• We avoid comparing multiple composers which are actually the same

[Maxikilliane]

I think this is generally a good idea, but could you clarify this a bit more. How do users define groups? Like do they give a group a title and then add the relevant music pieces? That would result in a bigger change of the Choice interface, right?

[blueCat11]

I have several ideas, which all use the infrastructure we have in the choice view:

1. We could add x number of columns to the choice-table, one for each possible group. We would define the maximum number of groups per analysis beforehand (that's x). Users would select one group per music piece they want to analyse. (+) avoids errors (+) relatively easy to implement (-) not very flexible (-) need to have an extra possibility to give the groups personalized names, to make reading the analysis results later, easier

2. We could add a textfield per row to the choice-table, and allow the user to specify group names by typing a group identifier/name. (+) more flexible (+) groups get names in the procedure (+) relatively easy to implement, in a basic version (-) very error prone (typos...)

3. We could add an extra section to the choice page to allow the user to define groups, that is to say to specify what groups the want for the duration of the analysis. Probably a text field, plus list of already defined groups, would suffice for that purpose. Then, in the choice-table, we would have a dropdown-select in each table row, allowing the users to select a group for the music piece in that row. (+) flexible (+) no typo errors (+/-) group names have to be defined beforehand, but are included naturally in the selection process (-) more work to implement than the other two possibilities

[blueCat11]

What do you think? Personally, I think the third variant might be the most easily understandable. I could get hold of some users tomorrow evening and test that though.