@amiloradovsky@functional.cafe

Well, that's a bit out of context, but I can argue, of course, what is time.

I wouldn't lay any claims on description fields of your table(just like I wouldn't on contents of files in general), but if you have a filed/column for a name, I believe the contents of that fields would be as short as possible while remaining identifiable. If your fields/columns themselves would have names they would adhere to the same pattern, otherwise(not having names) they will be identified by decimal numbers, not because those are more natural than unary, but because they are shorter, which makes them more identifiable.

For example when building a table for a family, you might use first names instead of full names for convenience. Even building a table for a larger community, you might have first and last names as separate fields/columns, so that occasionally when narrowing the scope to a smaller group/family, you can discard the last name column. Surely you can nitpick on this, on the grounds that we are not a hive-mind(or that surnames have spaces, which is, like, put me in a canon and blast me out of context kind of move), but it is just an example of a coherent system, not a statement on a natural phenomenon.

To bring it back into original context, if you can accept shortness of names, the argument "I need space for my short names, more than others need space for long sentences they write in text interfaces" doesn't sound very appealing. Keeping in mind that nothing is stopping the text interface from accepting two names(first and last, or more) as a single identifier in general, but when it comes to file names (in common file systems) the directory structure is used for that, which is not exactly an n-part id, but that's a different discussion. (and the separator is not space, and it is forbidden in file names... I wonder why... can't have GNU/Linux as a file... preposterous!).

@tuxcrafting