Should we have a [big-list] tag similar to TeX.se?

Question

In the spirit of being bold, I created a tag parallelling TeX.se's big-list, where there are 85 questions tagged that way, some with over 100K views.

The clone I created on Emacs.SE is here: big-list, although I would suggest to peruse the TeX.SE version which is more developed and will show the kind of questions that are suitable for this tag, as well as some that were posted with that tag and then closed.

See also:

• one related question and some preliminary discussion here on meta: Big list question on writing mail with emacs and the corresponding question on the main site: Reading and writing email with emacs.

  "This question is intended to provide a canonical landing point for users wanting to deal with email using emacs. Please post only one package/solution per answer, with as much details as you can."

• Another test case on the main site (placed "on hold" shortly after I asked it, but the big-list tag was removed by a moderator before the question was placed on hold, so I used a stylized title to show the "big list" designation): https://emacs.stackexchange.com/questions/13197/big-list-what-do-you-use-org-mode-for

• An earlier question with a similar flavor that received a relatively high number of upvotes: What are all the ways of launching a shell from inside emacs and what are their various properties?

In making the case for, one thing I would point out is that Area 51 asks for our "answer ratio" to be above 2.5, and it is currently (as of June 16, 2015) only 1.7. This suggests that however we go about it, we should find ways to create multiple answers to more questions. This goal will be more challenging to reach if every question is constrained to have one definitive "right answer".

To give fair play to the case against, I note this recent exchange in chat:

"So, I wrote What are all the prepackaged Emacs configurations and what are their advantages [dead link] and apparently it's "opinion-based". I basically wanted to ask "Hey, what's the point of using one of those". Many similar questions are very valuable on stackoverflow, so why can't we have them on emacs.se?"

"Many similar questions fall through the cracks on SO, because there aren't enough people voting to close. That doesn't mean that they're useful questions. These list questions always end up being badly curated, grossly incomplete, and generally useless."

tl;dr I do think the case of TeX.SE shows that such questions can play a valuable place on some StackExchange sites -- what about Emacs.SE?

Answer 1

I'm for big-list sort of questions

First, to answer some nay-sayers. In something as complex as a site community, guidelines should be interpreted in terms of best effort, not as a legally binding document, whereas failure to comply must be punishable by the law. So, SO guidelines insist that questions be posed in a certain way, alright. But there are other questions, which cannot be posed in the format proscribed by the guidelines. Now, are such questions valuable?--From my experience, yes. So, if we cannot recast these questions into the proscribed format, then we cannot ask them at all, hence, we'll lose something valuable.

There are indeed other sites providing information about Emacs, but of course this doesn't mean that this site cannot do it (as long as it is technically feasible). To reuse an old joke: it would be like telling a poet that there already is a verse about love,--probably won't stop her from writing yet another one.

Other counter-arguments such as "it doesn't bode well for the statistical model of some site parameter"--well, obviously, you get it wrong :) The model should be adjusted to the content, if there's a mismatch. Trying to do it the other way is bound to bring up dispute and disappointment.

---

On the side note, my life's goal (for a very distant future) is to write a better search engine :) My disappointment with current state of search engines (not necessary Google, it's an overall trend) is that they are, essentially question answering programs, but they are too picky about the kinds of questions you can ask. Picky to the point where whoever asks the question learns to ask only those questions the program can answer. But there is no reason why only those should be the good questions! Those are just the limitations of the approach, and of the technical side of things.

"Big list" questions are the kind of questions that don't exactly fit the SO format, but they have a merit for the site, so, maybe a proper way to go about it is to think how to make the format match, rather than to avoid the topic and to pretend that the problem will go away, as long as nobody talks about it :)

Answer 2

Stack Exchange is not a good format for “big list” questions

Stack Exchange works best with questions that solve a problem. The problem can be of various types but they usually fall under one or both or two broad categories “how do I do this?” or “why is it done this way?”.¹ This second class of “why” questions is also about solving a problem, the problem of understanding a concept.

Answers to such questions generally aim at solving the problem. The best answers aim at being definitive: each answer should stand on its own and would be valuable even in the absence of other answers. Not all problems have definitive solutions — not only for “why” questions, for which multiple points of view are often a good thing, but even for “how” questions, there can be more than one good way to skin a cat. Still, definitiveness is a goal to strive for.

“Big list” questions, where anybody could add their own answer, don't fit in the Stack Exchange model. The problem with these questions is that by definition, each answer is not particularly valuable, it's only the collection of all the answers that is valuable. Stack Exchange is not good at organizing such threads with one item per answer. In particular, voting ends up reflecting primarily who posted first, secondarily the popularity of the proposed item or of the author, and only slightly how well the answer addresses the question.

https://emacs.stackexchange.com/questions/13197/big-list-what-do-you-use-org-mode-for is an example of the bad kind of big-list question. If you give it time, it will either sink into oblivion with many good items missing, or it will end up having multiple redundant answers and totally useless voting.

What are all the ways of launching a shell from inside emacs and what are their various properties? is a different kind of question. It isn't one where everybody can add their own item and all items are equally valid (if not necessarily all equally useful). Rather, it calls for an answer that not only lists the possibilities but also compares them. It doesn't just call for “eshell”, “shell” and “term” but also when you would use each of them. There is currently a single answer, not a big list of answers.

Inasmuch as “big list” questions can work on Stack Exchange, they must be structured not as polls, with one item per answer, but as wikis, with a single answer that people are encouraged to edit to add an item in relevant places inside the existing structure.

[big-list] is not a good tag

“big-list” is a meta tag: it's a tag that doesn't characterize what a question is about, but the form of a question or of its expected answers. This isn't useful: it doesn't help when searching for questions, it isn't a tag that someone might subscribe to (nobody is an expert in “big list”) or a tag that someone might ignore because they don't care about the subject (if people ignore “big-list” because it's a repository for bad questions, these questions should go).

Either a question is unsuitable for Stack Exchange, in which case tagging it is irrelevant, or it is suitable, in which case it just needs to be tagged to indicate what the question is about.

If there are “reference” questions on a topic, that you want to highlight because they are of general interest to anyone with interest in that topic, link them from the topic's tag wiki.

Other remarks

In making the case for, one thing I would point out is that Area 51 asks for our "answer ratio" to be above 2.5, and it is currently (as of June 16, 2015) only 1.7. This suggests that however we go about it, we should find ways to create multiple answers to more questions. This goal will be more challenging to reach if every question is constrained to have one definitive "right answer".

You're working for the metrics instead of working for the goal. Artificially making up questions with lots of useless answers defeats the purpose since it doesn't improve the site. Don't pay much attention to the Area 51 metrics, they aren't all that relevant; this has been officially recognized and new metrics may be provided in 6–8 years.²

Emacs has a relatively low answer/question ratio, but it has a lot of pretty good answers that evaluate the pros and cons of different solutions. That's better than having solutions spread over multiple answers and no way to tell which is appropriate when. This kind of quality is impossible to measure objectively (the best you can do is average the subjective opinion of suitably-selected people), so you won't find it in metrics.

¹ _{A third category is “what” questions, as in “what is this thing”; these are best answered by reference works such as dictionaries, encyclopedia, reference manuals and so on, but Stack Exchange can serve as a fallback when not exists.}
² _{The official time for something to change on Stack Exchange is 6–8 weeks but this one is taking far more time than it should.}

Answer 3

No, we should not have "big list" questions

As @Gilles's answer explains, "big list" questions are not well-suited to SE sites. Here's what the site says about itself:

Guidelines from the tour

Let's have a look at what the site tour has to say about questions and answers (all emphases from the site, not me):

This site is all about getting answers. It's not a discussion forum. There's no chit-chat.

Focus on questions about an actual problem you have faced. Include details about what you have tried and exactly what you are trying to do.

Not all questions work well in our format. Avoid questions that are primarily opinion-based, or that are likely to generate discussion rather than answers.

These are all pretty clear guidelines. We're not a discussion forum. We're not asking opinion-based questions, and we're not asking questions to generate discussion.

Criteria for closing questions

Two of the options listed for voting to close questions are:

too broad

There are either too many possible answers, or good answers would be too long for this format.

primarily opinion-based

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.

These are pretty clear criteria. "Big list" questions would almost invariably run afoul of one, the other, or both of them. If we do not apply these criteria to the questions posted, we are not doing our jobs.

"Big list" questions have a place, but that place is not here

I am not suggesting that there is no value in "big list" questions. I am saying that they do not fit into the Emacs.SE model. There are other sites to go if you want to do "big list" questions: reddit and the EmacsWiki come immediately to mind, as do various mailing lists.

Emacs.SE is not, and was never intended to be, a one-stop-shopping site for all things Emacs. It does not seek to supplant other sites. It has a discrete mandate: Q&A about Emacs.

Why can't we change the model?

We have some limited flexibility in the norms of how we use the site. However, the criteria listed in the tour and in the close-question options are pretty clear. It seems silly to adopt norms that countermand some pretty explicit criteria.

"But other site kittens.SE allows 'big list' questions"

"But mom, all the other kids are doing it" is not a valid justification for us to do it as well.

We're still in beta, so we still have the ability to set norms that avoid mission creep. In light of the site's stated description of itself, "big list" questions are mission creep.

Answer 4

SO is a perfect home for big lists. What else can be more natural for a discussion group to include responses that take a list form? Otherwise we are just talking about the format of a response and not the real content (or relevance) of the response. Furthermore, big lists are already happening and it is okay to find accommodation instead of finding yet another way to penalize questions. Big lists will only make SO more attractive.