Suggestion: allow filter on tracklist values please

I would like to be able to filter on any values I see in the tracklist. E.g. I have a computed column "status" and I would like to filter: status IS ok. Thanks.

You can do that if you copy the computation expression from the column definition and paste it as filter expression.
Using an arbitrary column header would lead to the necessity to only allow unique labels for columns.

Like this?

Not at all. It would lead only to filtering being unavailable on the rare column names that are duplicates.

Yes.

Which is not quite "not at all".
It would also apply to user-defined fields which may have the same name as one of the columns and then you cannot even filter for the files that might have the field of that name as filtering would be disabled.
If you allowed filtering then it is not really clear which data would lead to hits.
An algorithm that automatically renames columns to make them unique again would probably be irritating.
So, from my point of view such a function would produce more hassle than ease of use.
Although copying and pasting may be a little cumbersome, it at least is not ambiguous.

Than any idea why that example is failing?

To few inverted commas?
Try a

"$if ...." IS "ok"

Thanks. It seems the first two quotes are needed, and the second two not.

And if there quotes in the expression, it fails.

I think the user should not have to mess about in this way just to filter on what is already shown.

You can use the filter also for criteria that are in fields that are not shown.
To use displayed data you have the option to sort by one or more columns.
If you write the computed data to a user-defined field, you can easily apply a filter that is as short as the one you mentioned in the first post.
The difference is that data in fields is more or less a constant string, while the computed result only exists at runtime. And this has to be computed every time

Sure I am not suggesting that be changed.

That's no substitute for a filter. Especially when viewing thousands of tracks.

I'll pass thanks, because the other difference is that it takes about 100x more time to write that data to the files than to compute and display it. And I have to rewrite it every time an input field might change. I don't want display-only data embedded in my music files.