I'm working on refactoring the Decode As dialog into using model/view functionality (
https://code.wireshark.org/review/22625). The model appears to be "forced" to provide the names of the column headers that are part of the "table" in the Decode As dialog. Currently the names of the column headers are stored in the .ui file, but with the model the data of the .ui file isn't used. Is there any reason to keep the column info in the .ui file? Does it help provide context to translators? (Although my understanding is that the developers that have provided translation are familiar with Wireshark, unlike some general software translation services) If it's removed, do the translators have to do some rework even those the column names haven't changed? I'm just wrapping the existing English names with the tr() function.
Any guidance on how to approach this (and future) refactors would be appreciated.