Tag wikis need a size and functionality increase

Question

The existing tag wikis are not enough. They can only store a very small volume of information, and given the complexity of the topics we're discussing here, that's just not enough.

I propose to make them full wikis- that is, full community-edited sub-sites in their own right to host various reference material that does not suit the main site's Q&A format.

Edit: I mean that specifically, the community would be able to create, delete, and edit pages pertaining to the tag in question. Permissions to do this would be based on reputation in that specific tag, rather than in general.

Yes, the functionality, at its core, is similar to Wikipedia. But it's not really similar in usage. Consider some of the reference material due for deletion. The Definitive C++ Book Guide and List, for example. You can't find that kind of material on Wikipedia. The tag wikis would be much more specific.

Ultimately, not every piece of helpful material ever is in Q&A format. Some of it is reference. The question is whether we're here to help people, or just answer questions. For example, there are C++-FAQ questions on the site right now which could, more accurately, live in here.

Another edit:

More accurately, the problem is that we have existing material which is not really Q&A, it's reference, and doesn't belong. It can't go in the tag wiki because the thing is just a Giant Wall'o'Text. All we need is a little subdivision so that we can link to specific sections.

As for competition with Wikipedia, it's not even close. Consider operator overloading. Wikipedia has about 4 lines covering operator overloading in C++, which basically state "You can overload XYZ operators.". The C++ FAQ has about 400 lines covering best practices in-depth. It is not really the same problem domain.

Answer 1

Usually, when I want to find something on SO, I use [search ] textbox at top-right corner of browser. It's intuitive action. But it seems that SO search looks only in questions, not in tag wikis.

I'm not sure if SO search engine can be easily modified to support tag wiki.
So, for now, keeping or moving information in/to tag wikis will just hide it from users (at least from me).

Answer 2

I think the problem with tag wikis is not their functionality, but their purpose and scope. In that they don't really have defined purposes or scope.

Consider c#, a fairly typical "good wiki" list of resources. What's the point of that wiki?

Does it just define C#? Is it to teach you C#? Inform you of interesting resources? Teach you when to use the tag? Display some FAQs (how much of that should be the same as the FAQ tab?)?

That particular wiki seems have settled on defintion, "interesting resources", and some history.

Compare that to the .net wiki which is mostly a list of libraries, or php which is about 1/2 FAQs.

What's the common goal* of these pages?

I'd argue that there isn't one, and this is the flaw in tag wikis currently.

At best, the existing wikis are info dumps; they might help you solve your specific problem (but probably not), but you'll have to spend a lot of time sifting through them. Sort of like the bad old phpBB days now that I think about it.

_{*Concretely. Nebulous goals like "make the internet better" (or "help people with programming") are great mission statements, but talking about a single page or feature we really do need a laser focus.}

---

If you reach waaaay back in time to when we introduced tag wikis you'll find this tidbit:

Tag wikis help introduce newcomers to the tag. They contain frequently asked questions about the tag and guidelines on its usage.

Current uses of tag wikis as "list of links" or dumping grounds for off-topic (but popular) questions are contrary to the initial purpose of tag wikis

---

I think that in order to salvage tag wikis we really need to give them a purpose. Tag wikis can never be successful unless normal people have a reason to go to them.

A well defined X in "use tag wikis to accomplish X", in other words.

I am very much against just slapping new tools on top of the existing system

I think all we can accomplish by, say, adding sub pages is subdividing already poorly scoped content. It'd be reorganizing deck chairs on the Titanic, polishing a turd, etc.

---

As to that X... I'm not sure. I'm personally leaning toward just tossing** the long form of tag wikis, excerpts*** already accomplish their initial purpose more than adequately.

Of course, there could be a compelling reason to keep them around I can't think of. However the "tag homepage" that /info sort of is right now could be reclaimed for more useful things than tag wikis have become, at least IMO.

_{**To be clear, this is still just a personal feeling; not anything planned by the team.}

_{***Confusingly, excerpts are not actually excerpts of the larger tag wiki. They're a separate entity, so removing the "long form" tag wiki doesn't imply removing excerpts.}

Answer 3

Tag Wikis define the tag for the purposes of the site, not the general topic. This topic was throughly covered in the blog post Redesigned Tags Page. Tag Wikis and exceprts are to teach people how and when to apply a tag. They are tag wikis, not topic wikis.

Stack Exchange is a Question and Answer site, and all features should be developed to support that, not as some weird, feature creep addition. Tag Wikis currently function to support the Q&A role of sites. Making them into topic summaries is not what they are for.

When included, topic summaries in Tag Wikis are included so you know when to use the tag. The tag wiki really doesn't need to be a starting off point to start learning a language or topic. There are hundreds of sites for that, but a Q&A site is not one of them.

As pointed out in this post, putting subjective information in tag wikis is the worst possible place to do it. Tag wikis can't be closed or voted on, if tag wikis become reading lists or fun facts or any irrelevant junk like that, they're useless. They are meant to help understand how a tag should be applied on your site, not be a Gentleman's Club of neat tricks High Rep Johnny found out and decided to slap into a tag wiki.

Answer 4

Currently questions like The Definitive C++ Book Guide and List can be used as a target for vote to close as duplicate, this applies more generally though for cases where the question is essentially book-work. This is exactly the sort of content that might work better on a tag wiki. The current system works well for the person who asked - they get an answer and it works well for the content - it gets more exposure which in turn leads to better content.

Suggestion: deep links to tag wikis should be the basis of a new "close as book-work" close reason. Benefits:

1. Questions still get answered, Q&A format preserved
2. Tag wikis get increased exposure (particularly to new users, who I suspect aren't aware they exist)
3. Useful content gets an appropriate home

(If you don't like the term book-work read "basic reference" instead perhaps)

This is effectively a more specific case of general knowledge close reason - it's different though because it's on-site, not a generic off-site URL and the tag wiki is an appropriate (if not currently perfect) place to put "useful introduction" wiki style content.

Answer 5

I'm all for this idea. It's a big improvement over tag wikis.

Although we already have a system similar to this - the community wiki. Instead of messing with tag wikis, we should expand community wiki functionality. Specifically, "protected community wikis", as that's what most of the problematic questions are.

It's quite simple how we do this:

• Protected community wikis only prevent new posts, editing the main post is allowed and encouraged
• The answers to protected community wikis should be removed and reworked into the main post.
• A special message explaining it's a wiki and discouraging spurious flags
• Protected community wikis have a special information-wiki tag for searching

So it's very similar (and compatible) to the old system, with the difference being that we distinguish protected wiki posts from question posts and treat them differently.

I'm using "protected" as thats the old terminology. It's a bad term and I suggest something like "information wiki" instead.

Answer 6

A lot of the back and forth on this topic both here and on the original question that sparked off this one seems to be based around the lack of visibility of tag wiki's on the believer's side ( IMO they're not visible at all ) and the detractors saying that adding a proper wiki engine wouldn't really work ( meh, don't know ) or that tag wikis are for information about the tag not about the language etc behind the tag ( also true ).

My vote, is for the proper wiki, but if that can't be done there's another option that might satisfy both the delete all the massively up-voted, old, off topic questions people and the please leave the best questions alone people.

1. Add a wiki tab, either in questions or tags.
2. Let people ask questions.
3. Automatically make every question under it community wiki.
4. Create the tags c-wiki, sql-wiki etc, then link to the specific tags wiki-tag in the tag summary.

It creates the semblance of a proper wiki but without needing to build the engine to do it. It allows people to keep (some of) the old questions, but enables them to be removed from the main site - just allow for a new PostTypeId in Posts and then exclude or include based on that. It also means that people can ask relevant questions that might be useful in the wiki format. It allows for the "great library of canonical answers". Lastly if there was ever something a move to create a proper wiki a load of content would already be available.