Reduce the contrast in the banding of tables

[pradyunsg]

Currently, the constrast difference between the banding for rows in a table is somewhat drastic, making it seem like the alternating rows have some sort of emphasis being applied on them. We should probably avoid that, since that's not the intent with this banding.

[pradyunsg]

Taking a few minutes to think about this, the approaches I can come up with are:

• Reduce the contrast. :)
• Remove the banding
  • in favor of light border between rows -- ALA https://squidfunk.github.io/mkdocs-material/reference/data-tables/#usage
  • in favor of collapsed borders around cells -- https://pradyunsg.me/furo/reference/tables/

And... that's the order of preference I have for them as well (i.e. reducing the contrast being my personally preferred option).

[hugovk]

Please see PRs #2740 and #2734 :)

[CAM-Gerlach]

And to summarize, this was inadvertently changed because of #2691

Reduce the contrast. :)

is #2740 (which basically restores the status quo), and

Remove the banding [...] in favor of light border between twos [I assume this is a typo for rows]

is #2734 (which does exactly that)

[pradyunsg]

Can we land one of those, and move forward on this?

[merwok]

The two PRs got comments from a few people, without clear consensus in my opinion.
I hate to say this, but I think you may need to start a discussion about PEP table styling on discourse.

Alternatively, a confident core dev could decide that if main docs don’t have stripes, PEPs don’t need stripes, and go merge that PR.

[pradyunsg]

I don't think tweaking colors to be better, should need a discourse discussion -- #2740 (comment).

Zebra stripes vs borders etc is a discussion I'm not interested in personally, so I won't be poking that forward -- IMO changing the complete presentation style of tables should not be a blocker to improving colors to make them more readable.

[AA-Turner]

I originally went with striping as it was used on the old /dev/peps site--count me amongst the supporters of #2734. We can look to merge #2740 in the near-term though as a straight improvement over the status quo.

A

[CAM-Gerlach]

Alternatively, a confident core dev could decide that if main docs don’t have stripes, PEPs don’t need stripes, and go merge that PR.

IMO, I'm not sure it would be entirely appropriate for another Python core dev to to swoop in and take unilateral action on an meta issue like that amidst ongoing discussion amongst the maintainers of this repository, as opposed to offering their opinion, suggestions and seeking to build consensus, which would of course be much appreciated.

We can look to merge #2740 in the near-term though as a straight improvement over the status quo.

Seems 👍 to me, I don't think anyone disagrees the current colors are too strong; #2740 effectively reverts to the status quo before the regression introduced in #2691 which made them that way, and the status quo wins in a stalemate (for now). To be clear, I'm not against #2734 , if that's where the consensus leads, but it is a bigger change that may need some further discussion, iteration and refinement, which suggests we should just fix the regression first and then have discussion rather than blocking on it.