Is there really a downside? Could make a button that allows you to turn off seeing tags at all. Unless there was some kind of a you can have one thing but nothing else I don't see any downside to it.
I guess that might be going a bit overboard. But it's not completely unprecedented.
Slashdot, for instance, has a restricted tagging system for their comments. Comments can be tagged as "Interesting", "Insightful", or "Funny".
I've heard some people on Slashdot complain about the "Funny" comments, but it's quite easy to filter those out via Slashdot's interface. I'd say that was pretty useful.
I can think of two downsides. First, there might be a better system of organization than “tags”, and if PG implements tags, then even if a better organization system is found later, PG will not want to spend even more time implementing this better system, so we would be stuck with an inferior system. Also, even if tags turn out to be the best system of organization, we would use up goodwill if we asked PG to make them for us, and he might decline to spend time upgrading HN in other useful ways later.
Comments
Is there really a downside? Could make a button that allows you to turn off seeing tags at all. Unless there was some kind of a you can have one thing but nothing else I don't see any downside to it.
But you can't turn off comments arguing over whether an article was tagged properly or not :)
I can envision a lot arguments like that taking over the discussion.
Those comments could themselves be tagged "meta". And there should be an option to ignore all comments tagged as such.
So we're talking about tagging comments too (as opposed to just the article)? Not sure we need to go that far...
I guess that might be going a bit overboard. But it's not completely unprecedented.
Slashdot, for instance, has a restricted tagging system for their comments. Comments can be tagged as "Interesting", "Insightful", or "Funny".
I've heard some people on Slashdot complain about the "Funny" comments, but it's quite easy to filter those out via Slashdot's interface. I'd say that was pretty useful.
Collapsible trees could be added at the same time.
I can think of two downsides. First, there might be a better system of organization than “tags”, and if PG implements tags, then even if a better organization system is found later, PG will not want to spend even more time implementing this better system, so we would be stuck with an inferior system. Also, even if tags turn out to be the best system of organization, we would use up goodwill if we asked PG to make them for us, and he might decline to spend time upgrading HN in other useful ways later.
"there might be a better system of organization than tags"
We should cross that bridge when we get to it. If there is no better alternative to tagging now, this objection does not hold much weight.
How long are we to hold off implementing improvements in hopes that a better solution is found? A year? Ten years? A hundred years?
Remember, there are no guarantees a better solution will ever be found.
And your argument could be used to hold off implementing any feature, not just tags. After all, a better solution might someday be found.
We should make reasonable, incremental improvements like tags instead of waiting for someone to find the holy grail.