My view is that 5e has too many skills to really be lightweight (especially once the tool proficiencies are counted), but it also has both some awkward gaps and some overlaps (notably Athletics/Acrobatics, Perception/Investigation, and Perform/musical instrument proficiencies).Say more… what number would be better for 5e and why?
But the weakness that really bugs me are the Charisma-based skills. Because either a character has one (which means all interactions follow a fixed pattern) or they have more than one (in which case there's no meaningful difference); and also because in most parties one PC will specialise in these skills and so gain an effective monopoly on one pillar of the game. I'm inclined to think that it would be better if each (and every) PC was proficient in talking to some groups of people, with Expertise being granted in various social circumstances. That at least spreads the fun out a bit.
As for a better number...
- Six. One very general skill for each ability score.
- Twelve. Each ability score is split into two (as in the really old Player Option days). A PC can be proficient in either side, but never both skills for any given ability.
- Twenty-ish. About the same number as now, but a revised set.
- Fifty-ish. Lots of very little skills, with an attempt to be comprehensive.
- Undefined. PCs just declare any skill they want, with the details left intentionally vague (and no attempt to be comprehensive). They then apply their Prof Bonus if any of their 'skills' apply.
- Undefined II. As above, but each PC only declares one or two - they're also assumed to be proficient in anything related to their race, class, or background.
Zero: The game doesn't actually absolutely require them, of course...