Better algorithm for column width

[platypii]

We should immediately reduce the default max column width, its way too wide now.

But in general I would like if:

• columns default to a reasonable width, but the user can make them narrower or wider as needed
• size differently for a file with a single column, vs a file with 40 columns (exact details tbd)
• screen-size aware

[severo]

columns default to a reasonable width, but the user can make them narrower or wider as needed

it's already the case, with #161 we limit the initial (or autoresize) width to a max of 40rem (640px for default font size) instead of 2000px.

[severo]

I'm interested in more details about what you would expect, maybe with some example datasets and an idea of the expected widths.
The current way of computing the width is not bad, I think, but happy to work on it to improve.

[platypii]

I don't have exact algorithm in mind, but would be probably something like:

• If there is 1 column, let it be up to the screen width wide by default (but expandable by user)
• If there are 3 wide columns, default to each being ~33% screen width
• If there are 10 wide columns, default to each being say 25% screen width and user would need to scroll right to see all the columns.

So basically try to fit stuff in view, but if there is simply too much content, then let it overflow. The factor missing from the current sizing algorithm is that there is no "pressure" squeezing the columns from the viewport width. And there is no difference in handling between 1 column and 10. 1 column should expand to show content, 10 should aim to fit more columns content for the user without horizontal scrolling.

[severo]

hightable 0.17.0 is published, and I updated hyperparam, the demos and the HF space