Why don't we block all tags that we remove?

Question

Currently, on Stack Exchange, it is customary to burninate a tag first, and block it from creation only if it becomes a recurring issue. I'm sure there's a good reason for this, but it eludes me.

Why do we do it this way?

---

If we burninate a tag, we're saying the tag is completely unwanted. It's either a meta tag, generates no helpful context, or has no meaning. The community wants it gone.

Burnination without blocking implies that it's okay if the tag returns, because we didn't take action to remove it permanently when we very easily could have. The question in my mind is: If a community has decided that a tag has no merit whatsoever, then what is the purpose of allowing it the luxury of returning?

This is what's confusing me. We burninate tags all the time, and they're always tags we never want to see again. If they keep coming back, we block them. This doesn't make sense to me - if we're just going to keep killing them until we decide to stop them from coming back, why not just stop them from coming back in the first place?

Answer 1

The short answer here is that tag blocking was not designed as an integral part of the tag system. It's an add-on, intended for rare problems that cannot be handled any other way. That may not remain true forever, but that's how it is right now.

The long answer is that there's effectively no difference between burninating a tag and... never using the tag in the first place. Oh, there are lots of little details that vary, but if you imagine a scenario where someone creates a tag by using it on a question and then immediately removes it within the grace period so there's no record of it in the revision history... Well, that tag will be deleted by the system at the end of the day, and as far as the system is concerned it looks exactly the same as an intentional mass-removal (burnination) would.

The end result? A lot of tags get deleted, for all sorts of reasons. Blocking all of them would be a lot of overhead at best, and potentially harmful at worst. For every tag that gets removed with prejudice, there are easily hundreds that are removed because they were simple typographical errors... Or simply don't have enough questions to warrant keeping around.

One might imagine a system where blocking - and burnination - work rather differently: instead of wiping all traces of the tag from the system, the tag would merely be marked either "inactive" or "prohibited", and either ignored or ignored and blocked as a result. In such a system, blocking would be a superset of burnination, and any tag that required conscious effort to remove would naturally qualify for it.

But, that is not how the current system was designed, nor is adding such functionality a trivial extension of what exists today. When we talk about "blocking", we're not even referring to a system specific to tags - rather, it is a system intended to block harmful or malicious input from users, completely divorced from the tag system. Indeed, the most common tag that are blocked are tags that never existed: each new site comes with a block list entry for that site's topic, thus preventing the [chess] tag on https://chess.stackexchange.com/ and the [rebase] tag on https://worldbuilding.stackexchange.com/ (don't ask).

Answer 2

One major reason is that it is simply not necessary in most cases to blacklist. Not all tags come back, and if you have an attentive community they'll often quickly remove bad tags if they appear again.

Blacklists are useful for tags that appear to be a good idea at first, but that are problematic if you think about it a bit longer. Those are likely to be recreated by well-meaning users, and blacklisting them makes sense.

There are a few drawbacks to blacklisting a lot of terms:

• you need an SE employee for each change
• a very large list of blacklisted terms makes it harder to review it
• the blacklist is based on regular expressions, a small mistake can block unexpected things
• I doubt the blacklist is optimized for large lists, there could be a performance penalty for checking a few hundred regular expressions every time someone uses a tag

Blocking users automatically from easily recreating tags that were intentionally removed before might be useful, but blacklists are not the correct tool for that. I could see a less invasive mechanism that e.g. just warns users if the tag they are about to create was previously removed, or increases the reputation required to recreate tags.

Answer 3

Only SE devs can blacklist a tag, for obvious reasons. The procedure of blacklisting a tag is exhaustive and takes time, as explained by this answer

Burninating a tag simply means that a tag is not needed in the current time. But that doesn't mean that it should be disallowed forever. A word that is unwanted one day can become the subject of many questions later.

Tag blacklist should only be done when we are sure that such a tag will never return. Blacklisting generally means that a tag is bad.

That said, there is no need to blacklist a tag just because it is a typo, it is unlikely to come back, and you don't want to annoy the devs with such requests that don't really help the site in the first place.