Inspired by bad-tag-cleanup project by our awesome mod, here I suggest the idea to cleanup duplicate tags:

This idea was duplicate of inspired by this thread in MSO: Clean up of run-together/hyphenated and singular/plural tags.

Currently, SE has mechanism to prevent tag creation that differ only in hyphenation or pluralization. But some tags was slipped before this idea was implemented.

In the past, we have some merge request for these tags for example:

Now, we can use powerful query from Stack Exchange Data Explorer (forked from the original query) to search all the tags and clean them.

For guidelines, I'll quote it from original thread

  1. First decide whether the tags really mean the same thing. There are false positives in this list. windows and window have nothing to do with each other. Most of the hits are good, though.

  2. If the tag names really mean the same thing, decide which name to keep. There are several general principles:

    • For multi-word tags, a hyphen is preferred.
    • There is no consensus for plural vs. singular for nouns. This old thread very weakly suggest the singular, and among popular pairs the singular is dominant, but recent practice tends towards plural.
  3. If it has been decided that a tag T2 should be merged into a tag T1, then:

    • If there are a lot of occurrences of T2, suggest or vote for a tag synonym, which will eventually lead to the tags being merged.
    • If there are only a handful of occurrences of T1, then we can get rid of T1 manually.

Several things that I would like to discuss are

  • For point 2, how do we decide which plural vs. singular tag that should stay? By opening a meta question for every tag pairs? By post count? Or what?
  • For Point 3, doing self mass-retagging would be cluttered the front page. And looks like our mods have a tool to do mass retagging (without bump the question) and synonymizer tool. So, did regular user allowed to do retagging or can we rely on our mods to do it?

I've listed all the affected tag pairs in two answers below. Posting the list in this question will make this question become TLDR-type. Feel free to edit it or bring it to meta to discussing whether the tag pairs should be kept or not.

  • I noticed you tweaking the query to get these results in data.stackexchange. One thing that stands out is that about 10-15% of the results had wikis only for one of the two tags. Here you could prioritize the ones that have a wiki, and make the ones without a wiki a synonym of the first.
    – Reaces
    Commented Feb 19, 2015 at 14:07
  • 2
    Just as an FYI, many tags with a singular and a plural form mean different things in their different form (or the questions tagged with them do), so there's that to look out for, both when merging, and when deciding on which form to keep as the master tag. To point 3, the rule of thumb is keep edits that bump old content to around 2 an hour (or less). If you can keep it to that rate, feel free. If you can't, leave it to the tag-bitch mod. (Me.) Commented Feb 19, 2015 at 15:11
  • @HopelessN00b, that's right. We should aware there is some false positive from the query above. For that, I have do some basic filtering by looking at tag name. Maybe other users can add feedback about tag pairs that should been kept or give feedback how we should decide master tag
    – masegaloeh
    Commented Feb 19, 2015 at 15:18
  • As a point of interest to any badge-hunters out there, there are at least 3 badges for working with tags in general, one of which is so rare, that only 3 people on all of Server Fault have it. So, that might be worth considering. On that note, I'm not 100% clear on the mechanism behind who gets credit for a tag's creation, but sometimes creating the tag wiki, if it doesn't have one already, will trigger credit to the person who created the wiki (and sometimes it won't), but something worth considering if you want a taxonomist badge. Commented Feb 19, 2015 at 15:20
  • 1
    @masegaloeh On that note, many of these dual tags with different forms still exist because one form has a high concentration of questions which should be assigned to a specific, different existing tag instead (host is a great example). It usually makes my life easier to keep avoid merging these tags together into a bigger group until I'm done filtering through for questions that need to be moved out... so if it take a while, well, that's because it's going to take a while to get the site's tags into any semblance of usefulness. As they say, Rome wasn't tagged in a day. Commented Feb 19, 2015 at 15:25
  • @HopelessN00b Looking at it further this is apparently by design. Tag creators are obfuscated To prevent people from gaming the system for badges. Which seems odd to me, considering the limited importance of badges them warranting censoring the creators names. Regardless... thx for the link.
    – Reaces
    Commented Feb 19, 2015 at 19:31
  • @HopelessN00b what badges are those? Which one is the rarest?
    – pepoluan
    Commented Feb 22, 2015 at 15:56
  • 2
    @pepoluan, maybe he means Taxonomist, Tag Editor and Research Assistant. The rarest one is Research Assistant.
    – masegaloeh
    Commented Feb 22, 2015 at 16:05

2 Answers 2


This list of tag pairs that differ by pluralization. The number inside brackets was number of question tagged with that tags.

Untouched tags

Post-cleaning tags

False positive results

Keep one tag, burn the other

Burn both tags

Other action

✓ = tag will stay; ✗ = tag have been removed

  • FYI, I've unwikied your answers (and there's a reason for that), but more importantly, keep in mind that the data you get from the SE data query is a few days stale, and I'm pounding away at this, so your lists will probably lag behind reality by a little bit. Commented Feb 20, 2015 at 7:02
  • Yeah, I'm already checking every tag by clicking it one by one :). I'll try updating the cleanup progress here.
    – masegaloeh
    Commented Feb 20, 2015 at 7:05
  • 1
    I'm not sure guides and guid should be considered synonymous… Commented Feb 23, 2015 at 16:17
  • Hi @owen, as I said above, 'guides' tags was deleted and 'guid' tag stay.
    – masegaloeh
    Commented Feb 23, 2015 at 21:41
  • 1
    @masegaloeh I can see that. But guides is not a plural for guid; they are unrelated words. Commented Feb 24, 2015 at 12:13
  • @owen, yes I'm aware of that. The query was too naive and stupid to detect this. I've edited above post to make it clear about false positive results
    – masegaloeh
    Commented Feb 24, 2015 at 12:50

This list of tag pairs that differ by hyphen. The number inside brackets was number of question tagged with that tags.

Untouched tags

Post-cleaning tags

Synonimzed tags

Burn both tags

Keep one tag, burn the other

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .