search

wtchg-kwiatkowski / observatory-web

0 stars 0 forks source link

Article category checkboxes: all-checked has the same effect as none-checked #288

Open leehart opened 6 years ago

leehart commented 6 years ago

Maybe the default state should have all the checkboxes checked, i.e. articles shown from all categories. When no checkboxes are checked, no articles should be shown (Unless there is an uncategorized category?), perhaps with a message explaining why no articles are shown.

Maybe the categories that each article belongs to could appear as a sort of boxed-label on the bottom-right of each article's card. (Sounds like a separate feature request, but related.)

benjeffery commented 6 years ago

I thought this sort of behaviour was very common place and expected, I deliberately coded it that way.

leehart commented 6 years ago

Any big-player examples we can look at?

I wouldn't be completely surprised if the current behaviour is conventional/popular, and that would certainly trump my personal opinion of how I would expect it to behave, which I thought seemed perfectly logical, but hey... strange world!

leehart commented 5 years ago

Similar behaviour is common-place in some shopping and comparison websites, where lists of checkboxes are used for filters, where all-unchecked = no filters = all-checked = no filters, e.g. https://www.uswitch.com/mobiles/iphone_deals/

The reason seems to be that it's easier for the user to check-mark the things they're looking for, after being shown everything, rather than unchecking everything except the things they're looking for, after being shown everything. Even though it's not logically consistent (all-unchecked != all-checked), users are generally happy to overlook that quirk (or generally don't notice) and instead gain the convenience. (It seems to pay to show everything first AND start with all categories unchecked. A kinda SFAQL approach.)

[I should add: the interfaces that show-all when all-unchecked are probably steering away from showing the user nothing at all, which isn't so good for the user or the sales/data-funnelling.]

I'll look more closely at this, to check alternatives (I'm still bugged by the logical inconsistency), and I'll investigate the second part, i.e. adding category-labels to each article summary.