I’m trying to better understand how Dt initially determines/autosizes column widths and when manual resizing is retained. Here’s a specific example:
Opening a certain group containing a few dozen items in a fairly wide window was initially in List view with Name/Modified/Kind/Size/URL columns. The Name column was too narrow, with “extra” width given to the URL column. Other column widths had good defaults.
Double-clicking the column divider between Name and Modified resized the Name column to the width of the longest name, shrinking the URL column’s visibility to the window window but also making it horizontally scrollable. Unexpectedly, all columns were reverted after closing and reopening that window. But then …
Temporarily making the Name column narrower, dragging the now-visible URL column’s right divider to the left, double-clicking the Name/Modified divider again, and finally closing and reopening the window caused the Name column width to be retained with the URL auto-sized (and truncated) to fit a non-scrolling window width.
It seems there’s a higher priority being given to URL (at least) than Name field/column sizing/retaining, similar to the “Tab makes URL field editable” example in:
[DtPO 2.6.1] text field editing/focus misbehavior
By default I’d expect URL, not Name, fields to be truncated relative to column/window widths.
Something like that. 
I’m still befuddled trying to understand how Dt groups retain/reuse previous windows, columns, views. Analogous folder management with Finder is much more predictable and consistent.