8

Here's version 2.0, which cuts a lot of the flavor in favor of more bullet points :)

Problem: Tags != Organization

Organization requires:

  • Planning, to understand the best way to organize
  • Structure, to match the complexity of the items being organized
  • Policy, to ensure items are properly sorted as they are added
  • Maintenance, to fix issues created over time

Stack Exchange Tags are:

  • Easy to create and apply, low rep requirement, few checks/barriers
  • Applied by question askers (who frequently are unaware of policy)
  • No checks or methods of policy enforcement
  • Hard to maintain (page-by-page reviewing 2000+ tags on 20,000+ questions, edits spam the front page, etc)
  • No concept of structure any more complex than a single layer deep

Tags are not good for organization. We fight/overload the system, and use it in a way that is not compatible with how it was designed when we try to use them for organization. The system is designed in a way that makes our jobs hard.

Current State: Brute force and spot checking

Our current solutions to this (periodic tag review and destruction, occasional discussions on policy) are good, but are overly brute force and could be refined/improved/extended.

  • Vigilance (attempting to process each entry as created) is important, but at the rate questions are asked and tags are created, we are bound to miss tags now and again. Also, a system where, by design, the power users/administrators must touch many, many bits of user generated content in order to ensure organization is enforced is inefficient and/or poorly designed.
  • Maintenance (after the fact cleanup) is essential, but is not enough, and it is currently too hard. The tools we have are not designed to make maintenance easy. The data exists, so helpful tools could exist.
  • Policy is good, but it requires either vigilance or maintenance to enforce - there is no policy enforcement built into the system. It's hard to make coherent policy that works within the system, because, as stated above, tags are bad for organization.

Possible routes of attack:

  • try to campaign to fix the system (thereby preventing the mess), which is difficult/complicated and requires SE intervention, and their inertia on the subject is great
  • try to make identifying and cleaning up the mess easier (perhaps something we can do ourselves?)

In the meantime, we have to acknowledge that things are going to continue to get messy, and require more effort than it really should take in order to clean up.

  • 9
    This is less a feature request than a rant. – Raven Dreamer May 31 '12 at 19:09
  • 1
    @RavenDreamer, defining the problem is step one. Defining the solution comes next. The intent here is to define the problem. – agent86 May 31 '12 at 19:10
  • 6
    @agent86 and I agree with that. But I don't think you've done an adequate job of defining the problem. This seems less like defining a problem, and more like a series of complaints topped off by "we can't succeed the way things are currently". – Raven Dreamer May 31 '12 at 19:10
  • 2
    I have to agree with @Raven. I don't quite get the point you're trying to make here. OK, tags are not perfect and until they are changed we should lower our expectations. And? I think I'm just missing a line or two explaining what responses you are expecting to read here. What is the feature you are requesting? In a sense, your answer below gives a hint to what you're expecting, but I'd like to see it in the question itself. – Oak May 31 '12 at 19:33
  • 2
    Calling it a rant is awfully dismissive - here I'm trying to get us to rethink. I think we're focusing our efforts on retagging every month, getting some small subset of the bad stuff out, and then watching as it fills back up again with bad stuff. Can't we think about this from a different perspective for a bit? – agent86 May 31 '12 at 19:46
  • 7
    Mods fighting scares me. If the mods get divorced, will I have to go live with grandma and grandpa? :( – Dave McClelland May 31 '12 at 19:52
  • "Bad tags would still crop up constantly, and they'd frequently slip through our cracks." How many bad tags have been created recently, versus created a long time ago and never cleaned up? I feel like we've done much better about this more recently. – bwarner May 31 '12 at 20:38
  • @bwarner "directx-3d" "secret" "skill" "books" "user-interface" - just a quick review of the first couple of pages of the "new" tag list. – agent86 May 31 '12 at 21:18
  • 1
    @DaveMcClelland You can come live with me. BEEP BOOP – user9983 May 31 '12 at 21:43
  • @OrigamiRobot Can you make human food, or will I have to learn to live on oil and alcohol? – Dave McClelland May 31 '12 at 21:46
  • 2
    @DaveMcClelland Yes, I can make human food. Humans like swarms of things, right? – user9983 May 31 '12 at 22:01
  • @Dave naa, we fight internally all the time. Whenever someone loses he just deletes questions and answers randomly to vent his frustration, which is why you sometimes see unexplained deletes :) – Oak Jun 1 '12 at 6:18
  • 2
    @Oak Nah, when I lose a fight, I'm going nuclear. Mass merge all the skyrim, diablo-3, and minecraft questions into murder. That'll show them. grumblegrumble. ;) – Raven Dreamer Jun 1 '12 at 20:14
  • 1
    i'm literally crying at how much truth this question presents – Mr. November Jun 2 '12 at 13:34
7

In writing this, I came up with a sample (incomplete, wish list, etc) "feature set" for things I'd want to see in Tags 2.0. Didn't really seem to fit into the question, but perhaps it will make sense as an answer:


We seem to have problems with:

  • Tag editing bumps questions. Trying to fix any more than ~5 questions with a bad tag spams the front page.
  • Tag creation is very low friction - there's no way to have coherent policy that is at all complex, as it requires people to have been exposed to the policy already. There's no way to enforce policy on users.
  • Getting tags applied consistently is a chore. We have to go back to tons of questions and evaluate them, and then attempt to recategorize them.
  • Tags disappear off of questions when there's just one question tagged, and we solve this by retagging one-question tags on a regular basis.
  • Tag length is too short for our use case of "tag the game name" - games frequently have long subtitles that are useful for SEO.
  • There's no relation between tags, although we fake it.

Some of this could be solved through different software, and some through different tag attitudes/policy/usage. Not all of this can be solved through different usage, though, unless we abandon some of our policies wholesale, or special case everything frequently.

Some nice to have features:

  • Subdivisions of a given organizational structure. X is an expansion to Y. A is a class or unit type in game B. Some level of association or organization would be nice for times when we don't want to have to spell out the structure in the tag, or create some sort of assumption of hierarchy between tags.
  • Review of or tracking of new divisions. If someone's creating a new tag, we should know when that happens and be able to say "no, this does not fit" or "yes, this may continue" - retroactive removal of new bad tags would be a single click for some level of user.
  • Ability to mass-retag with maximum flexibility. For instance, if "achievement" is a good tag, and should be applied to questions involving said game mechanics, we should have a system that allows us to search, mark, and re-tag questions with minimal effort. Bonus points if it's able to look at the questions under a given tag, and suggest similar questions that might benefit from it.

I have the ability to create API tools, which might provide some stopgap measures, as well as we can make API N+1 requests, which if we can come up with a really great idea that we don't particularly think the dev's are willing to invest time in, I'm voulenteering my own time to mock up ideas or tools if we can make something that works.

  • I feel like the reply to the third point from the staff will be around the lines of "sure, tell me what you need to do and a dev will try and squeeze a moment to figure out the SQL incantation to make it happen." – badp May 31 '12 at 19:05
  • 10
    Quite honestly, you should have lead with this instead of the rant above. – Raven Dreamer May 31 '12 at 19:07
  • You request subdivisions, but you never specify what functionality they should drive. What difference would marking a tag as a subdivision of another tag make? – bwarner May 31 '12 at 20:36
  • 1
    It's mostly spitballing an idea at this point. What I'm thinking is something like "tagging binding of isaac suggests wrath of the lamb" or "you can't tag wrath of the lamb without binding of isaac" or "each tag has subtags, ie, 'zerg' is a subtag of 'starcraft' as well as 'starcraft-2' and those two things are separate, even though they have the same name. Mostly, I'm loathe to suggest solutions - I want to report problems and request the designers, the people with knowledge of the code/system/goals design the improvements. – agent86 May 31 '12 at 21:29
  • The rep limit is the friction in your second bullet; when I first came here I couldn't make an Earthbound tag for example. But on Gaming it's really easy to earn that much rep without learning the system – Ben Brocka Jun 4 '12 at 13:09
4

Tags may suck, but getting this response (to be fair, not just from you) virtually any time anyone attempts to fix it got tedious a long time ago.

To everyone: if you're not willing to help with tag projects in good faith because you think the tag system is fundamentally broken, that's fine. I get the argument and am not going to begrudge your right to hold that position. But it's both demoralizing and counterproductive to get this when people who are actually trying to fix tag problems one step at a time are dismissed with "LOL TAGS SUCK GIVE UP NOW" every. single. time.

It's natural for each Stack Exchange site to think they're not like any of the others and their problems are unique and unmatched, but the fact is we're not special snowflakes and every single site on the network grapples with tagging issues. Moreover, most sites have been able to make it work, tagging problems and all.

So we can either complain every time a tag discussion comes up about how we can't have tag hierarchies, or that platform tags mean X to Y and Z to A, or we can make it work.

To that end, I propose a new guideline: if anyone gets wrapped up in a tagging discussion/argument and they feel the urge to write the whole thing off because Stack Exchange's tagging system doesn't do X or Y: they should feel free to walk away from it and let others handle it.

I don't think anyone would be offended if someone opted out of a tagging project—we don't all need to be involved in every single discussion—and I think everyone's heard the arguments about why tagging sucks on SE before, and if they haven't, there's more than enough back history on meta to get them up to speed.

Anyway, that's my thoughts on this: consider yourself counter-ranted.

  • 1
    When software doesn't work the way it's being used, you have three options. One - do work to counteract the software's shortcomings. Two - attempt to adjust your use case to the software's pattern. Three - attempt to change the software to make it work more like the use case in practice. We almost always go for one. I apologize if this seems like a rant, but what I want to consider is either two or three. Accept that we can't have what we want, and/or request that what we want become the software's use case. – agent86 Jun 1 '12 at 1:05
  • 2
    @agent86 Given the same arguments get trotted out time after time for the past 2 years any time anyone starts a tagging project, I think it's safe to say the ship sailed on "let's get SE to radically change tagging" a long time ago. Rehashing this every single time people who have accepted that try to fix things is not helpful at all. The only two options are get on board or get out of the way. – user3389 Jun 1 '12 at 1:08
  • 2
    So "counter rant" if you like, but this is less productive than outlining the problems and campaigning for long-term change. Saying "I give up on changing, and we just have to continue work with the broken system in the broken way we've always done it" is worse than "hey guys, this loop we're stuck in needs to have a break statement at some point" – agent86 Jun 1 '12 at 1:09
  • @agent86 there's a difference between having a rational argument, presenting that to the decision makers—letting the chips fall where they may—and repeating the same thing over and over in hopes that you'll wear them down and just agree after months to going on years of them saying "No, this is how it works". It's insanity we have the "tags are broken" talk every single time we start a tag cleanup. – user3389 Jun 1 '12 at 1:11
  • 1
    If you think I haven't put forth the use case, or suggested some potential fixes, then help! Taking what you've said and turning it back around again, if you just want to bitch about my supposed bitching, then why are you bothering? I'm not anti-tag cleanup; I'm pro tag cleanup. I'm just pro-let's-not-keep-doing-this-constantly. – agent86 Jun 1 '12 at 1:11
  • @agent86 I think the premise of the argument presented in this question is flawed: tags are good enough and other sites have made them work, so let's just cut out the meta-discussions like this about the nature of tags and focus on the merits of specific cleanup projects. Look, if people want to keep beating the "tags are broken" drum like it's going to do something, there's not much I can do to stop them: but at certain point, the broken tag situation becomes a self-fullfilling prophecy. – user3389 Jun 1 '12 at 1:23
  • I guess then you can take your energy and invest it in that project and I in mine - long term we want the same thing, which is a better, cleaner system. If you think we can achieve that with the way things stand, I wish you good luck and I'm rooting for you. – agent86 Jun 1 '12 at 1:25
  • +1 I agree with everything you've said. Almost everyone agrees the Tag system sucks, but most of us have also come to accept the fact that the SE just won't do anything about it. We just need to work with what we have. – Wipqozn Jun 1 '12 at 11:34
  • 1
    @Wipqozn It's all software, operating on data. We have access to the underlying data in several different ways - even if SE is refusing to take action, that doesn't mean we shouldn't cry out in pain at their poor software choices, and try to do what we can to munge that data better. Even small fixes to our use cases would be a big help. I guess part of what doesn't make sense to me is that I read Joel's work, and then look at what he's produced here, and the two things don't line up - maybe that's what gives me hope for change. – agent86 Jun 1 '12 at 12:02
  • 1
    @agent86 We have cried out in pain though, and it just never accomplishes anything. In any case, there's obviously no point debating this. You and I obviously have opposite views on the issue. – Wipqozn Jun 1 '12 at 12:06
  • Sometimes SE gives in on these things, and with Jeff gone they are much less likely to cling to a specific view as if it's the only one that exists. Personally I'm all for saying "Can you not see how broken this is?" until they can't ignore or rationalize it any more. I'm fine with the "Not worth the effort" response, not so much with the failure to admit there are better solutions. – Matthew Read Jun 1 '12 at 15:47
  • @MatthewRead Nobody's suggesting there aren't any better solutions. The better solution is to learn from other SE sites who have been able to use the tag system successfully and actually follow through on a tagging project instead of sticking our head in the sand, writing them off every single time they come up, and complaining this is all SE's fault for not listening because people have differences of opinion about what tags are good and bad. 80+ sites that run the gamut of topics figured it out: we can too. – user3389 Jun 1 '12 at 18:53
  • @MarkTrapp I have seen it said that there aren't better solutions, yes, thanks; I wasn't referring to your post. And plenty other SE sites have nontrivial tagging issues. I'm all for making the best of the system we have, and will continue to work within it as best I can, but I'm also simply not going to stop pushing for better alternatives in any aspect of the site. I really don't see how you can say we're sticking out heads in the sand when you're the one saying you're tired of hearing this and acting like working within the system is incompatible with wanting a better system. – Matthew Read Jun 1 '12 at 19:01
  • 1
    @MatthewRead That's not my position; that's the position presented in this post and why I wrote this answer: "We have failed so frequently in the past, and failure in this system is so easy that fixing it would be not only difficult, but largely pointless." As I said in my answer, I don't begrudge anyone's right to believe the tag system is broken and I get the argument. My issue is when people write questions like this that say it's impossible right when we're in the middle of a major tag cleanup. – user3389 Jun 1 '12 at 19:31
  • OK, my apologies for the misunderstanding. – Matthew Read Jun 1 '12 at 19:32
1

Would it be a huge stretch to ask SE to change the tag placeholder to something else for gaming?

I think the current wording promotes using more than one tag, however one tag is usually all that is desired here. I can't speak for others, but I know when I first started asking questions on SE I used to try and use as many tags as possible since I thought it would get my question more visibility.

It would be nice if said something like

enter the tag of the game you are asking about (diablo-3, skyrim, minecraft, etc)

instead of

Screenshot of Tags in Ask Questions page

You could make your case that tags on Gaming are different than tags on any other site, so should be handled slightly differently.

Perhaps even make the Tags heading be a link (or contain a (more info) link next to it) that goes to a meta post instructing users how they should be using the tags for gaming.

-1

One thing I thought of this morning which may help things is restricting new users from using new tags. Bad tags are created by 300+ rep users, but I have a feeling that it's the new users who may perpetuate them. To curb this behaviour, I suggest that new users can only use established tags. The only detail would be to determine what "established" means. I think there are 2 options there: post count and time it's been active. I'd shy away from the later simply because it's already used to auto delete new tags that aren't active.

  • That's an interesting suggestion, but it's not that easy - what if it's a relatively-uncommon game tag? It would be very appropriate... but I guess a smarter heuristic could be implemented, e.g. disallowing a tag if it's relatively rare and used by a new user and there's already an established tag on the question. Still, before we suggest something like that, I think it would be worthwhile to monitor new posts, see how the tags are misused, and see what effect would this policy have had (?), had it was implemented. – Oak Jun 4 '12 at 12:52
  • @Oak: I definitely agree that we should Do Science before we implement something like this. However, since this is a spit-balling post, I thought I'd throw the idea out there. It may only be one medication in the tag treatment plan, but every little bit helps. Another thing to realize is that there's nothing in this plan stopping 500+ rep users from coming in and adding restricted tags if they're appropriate. – MBraedley Jun 4 '12 at 13:16
  • I think the main problem is misuse of existing tags, really. Once we get the massive cleanup of old crap tags out of the way, we just need to watch for new bad tags. But inconsistent and non-useful application of tags is impossible to prevent with the current system. – Matthew Read Jun 5 '12 at 17:15

You must log in to answer this question.

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