Summary: SE explictly encourages having lots of questions ("millions of questions... [so] someone types a question into Google and finds their exact question already answered"). It also allows "canonical questions" where lots of related problems share a common pattern and can all be answered by one answer that explains the pattern.
When someone has a problem not covered by the standard approach covered by the canonical question, they are allowed to ask the question. There's no precedent for trying to shoehorn every problem within an entire topic area into one question. That's not what canonical questions are for: SE is "not meant to be a library of reference manuals".
Here's a real-life example of the correct use of canonical questions, from our mother site, programming Q&A Stackoverflow:
- They have a canonical question about a common problem, how to respond to NullReferenceException errors. There are lots of types and causes of these errors, but there's a pattern to most of them, so therefore there's one standard approach and method that fixes most of them. The answer details that approach, in extensive detail. One pattern, one question, one +1270 accepted answer. Alternate answers offer alternate general approaches.
- This is not intended to be a complete be-all and end-all reference manual on NullReferenceException errors, constantly expanding to cover every single case and every possible exception to the pattern. They have an entire tag for those.
- For example, that site's #1 ranking user Jon Skeet encountered an unusual NullReferenceException error, and asked Why would finding a type's initializer throw a NullReferenceException?. It's a problem in the same area: he's encountered a NullReferenceException error, and is, in his words, "stumped". But it doesn't fit the standard pattern covered by the canonical. This is a much more difficult NullReferenceException problem than most.
- Therefore, as an atypical problem, it is not a duplicate of the canonical answer designed to cover typical problems, even though it is an example of someone trying to understand and fix a particular NullReferenceException error. It's not closed, Jon Skeet is not told to go away and place a bounty or hope someone by chance thinks of his problem and edits it into the canonical. It doesn't fit the standard pattern, so it's not a duplicate. There are hundreds of questions about fixing non-typical NullReferenceException errors that don't fit the standard pattern, and that's fine - that's what a Q&A site is for.
How this applies to the question under discussion: the linked canonical question should be a canonical question, but by trying to bloat it with an answer for every country on earth then duping every related question at it, we're going about it the wrong way. Reading all the answers and comments, there does appear to be a normal canonical pattern that does actually cover most countries, but is mislabelled. The SE-standard 'canonical' approach would be, I think:
- To have an answer that describes the 'standard method', of turning 123456 into 12345, 123 into 12300, SW1A 2AB into 12000, and trying 00000 if all else fails
- Say "this has been tested and appears to work for: Canada, UK, [list]" and encourage people to comment or edit in new countries that fit the pattern
- Allow new questions when a particular country or bank card doesn't fit the pattern, like how Jon Skeet was allowed to ask about his particularly tricky NullReferenceException error even though there was already a canonical question about fixing typical NullReferenceException errors. If Korean cards need a different trick to the standard trick, leave it open. If they fit the standard pattern, dupe.
- If in doubt, if someone asks and you don't know if they've tried the standard approach, ask. "Have you tried this approach? [link]". If "Yes that works thanks" or no reply after a reasonable time, dupe. If "Yes and the standard method doesn't work in my case", leave open.
I'm not sure where exactly this idea of a few massive encyclopedia-style articles trying to cover every problem and every circumstance within an entire topic area came from, but it explicitly goes against stated aims of Stack Exchange: which is to:
- ...give people specific answers to all their specific problems, "in the form of millions of questions and answers"...
- ...because the "more chance that someone types a question into Google and finds their exact question already answered, the better a job we've done"...
- ...and we are "not meant to be a library of reference manuals".
This official blog post is from 2011 and so is specific to StackOverflow, the programming site that Travel.SE grew out of, but the same principle applies. Many answers on the official meta re-iterate it (example from SE staff). My bold, and I've added notes in italics:
The Wikipedia of Long Tail Programming Questions
["long tail" was the idea, considered novel in the 2000s, that providing thousands of niche products or services to small numbers of people each can add up to a lucrative business model]
...[introduces and explains importance of duplicates and community editing, in contrast to old sites and tools like Usenet where such things were impossible]...
If you’re going to close a user’s question as a duplicate, it has to be a real duplicate. For example, if a user asks, “What does the IP address 128.0.1.1/24 mean?” it’s OK to close that as a duplicate of a more general question like “What do IP addresses of the form a.b.c.d/e mean?” [NOTE for non-techies: these follow a common pattern that can be explained once, and are therefore NOT like adding a new paragraph to an old answer for every country on earth] But it’s not OK to close it as a duplicate of a twenty-seven page guide to netmasks. That’s the moral equivalent of saying “RTFM.”
Stack Overflow is not meant to be a library of reference manuals. It’s supposed to contain the same information as a library of reference manuals, in the form of millions of questions and answers. Combined with Google, that gives us the magical power of a library of reference manuals you never have to read! It’s like, you got to the library, and there’s a wizard there at the door, and you ask your question, and, instead of being told to read a book, you just got (are you sitting down?) the actual answer!
That's why we actually don't mind having several versions of every question, where there are variations in wording or circumstances. The more chance that someone types a question into Google and finds their exact question already answered, the better a job we've done.
It goes on to say that there's nothing wrong with a reduced version of the encyclopedia idea:
It is OK to edit a question to make it more general. With the power of editing comes the power to take someone's selfish, very specific question, and edit it a little bit until they're asking the more general question that hundreds of people encounter.
Another official blog post makes the point that the creation of canonical questions needs to be done with good judgement and restraint:
...but if every variation in circumstance requires adding another paragraph to the answer, those are different questions, and the aim is suited better by letting people google straight to it instead of following a dupe link then digging through dozens of paragraphs.
There's a very recent (June 2016) contribution on this topic from SE staffer Shog9 on the SciFi site describing a similar debate to ours:
The idea of a massive canonical Topic FAQ is persistently alluring; instead of having 50 questions about a "sorting hat", you could just have one question with a few dozen huge answers! However, this quickly becomes impractical:
- Massively broad questions don't rank as well in search results as specific, focused questions.
- Folks with specific, focused questions tend to not read massively broad FAQs even if they do find them.
- Finding specific information among multiple answers to massively broad FAQs is troublesome.
- Remembering which information is even contained in these tomes is difficult; eventually, folks just start to assume that they contain everything and close new questions without worrying whether they're actually answered or not... This chokes out new information.
Canonical questions should be simple outlines of patterns, equivalent to “What does the IP address 128.0.1.1/24 mean?”, not "massive canonical Topic FAQ" "tomes" that are bad for SEO, not user friendly because the relevant bit is buried, and "chokes out new information".
The policy is to not optimise aggressively for as few questions as possible. It's to ensure as many genuine distinct problems and "variations... in circumstance" are solved as possible, to the highest standard, and then make them easy to find and keep up to date.
I understand the urge (popular in academia) to ruthlessly prioritise neatness and consistency, even at the cost of usefulness, but it goes against the stated aims of the site, and hasn't worked. Wikitravel, Wikivoyage and Wikipedia already exist, we don't need to duplicate them.
As the asker points out, no-one added new countries to the 'canonical' US credit cards/gas station question in 3.5 years. Why would they? There was plenty of time to add South Korea, but no-one did. Why would they? SE doesn't have any mechanism to encourage users to do that. Those Wiki sites do - such as talk pages behind each page where people can organise themselves and say things like "Guys, we're missing South Korea, Russia, Isreal, someone get on it". Reading the comments above, one of the few users who strongly supports the encyclopedia model literally said that because this system has failed, people who want answers for countries not covered should take their (answerable, on-topic) travel questions to a different travel site. That's not what SE is about.
We're a Q&A site with a little wiki functionality. Our mechanisms primarily encourage people to answer, edit and cross-link new and unanswered questions. We have nothing to encourage people to dig up some years-old answer that the asker feels is adequately answered, add another paragraph for another country, and go against the stated aim of millions of Q&As where people can cut straight to their answer with a google search.
It's not surprising therefore that people don't do that, and we shouldn't have a policy that tells people "You're not allowed to get the answer to your question, because several chatroom regulars have convinced each other that people should have put the answer here already, even though no-one has", when we know the whole SE model is not optimised or designed for that.
One of the comments under that question expresses a theoretical concern:
I could create 150+ questions on paying with a credit card of each country on Earth then. Voting to close as dupe
...but that's not how real-life Q&A sites work. People ask when they need an answer (or want to self-answer). People answer when they see a question. The fact an unhinged individual could, theoretically, write "Can I use credit cards from [X country] in US gas pumps" 150+ times is no more relevant than the fact someone could go to, say, StackOverflow and ask about iterating through arrays recursively "in [X language]" 150+ times. It's a Q&A site, people ask when they need an answer. It works.
It's the alternative that doesn't work. Are we really supposed to fill a years-old question with 150+ answers, one for each country, not even knowing if there's any demand for info for each specific country? That's clearly a bad model. Why not help real people by answering real questions when they're asked? So that person who follows the Zambia tag because they know a lot about Zambia can see the Zambian version of the question when it's useful to someone who is interested in Zambia.
If you've ever used StackOverflow, just imagine if they used this model. Someone asks a question about some detail of memory assignment when passing objects by reference in PHP. They get an answer. Someone edits out 'PHP' and decides this is the 'canonical' memory-assignment-when-passing-objects-by-reference question. Someone adds a Javascript answer, because it's easy. People add rare esoteric languages like brainf**k to show off their knowledge. No-one happens to bother adding, say, Python. Someone asks the same question about Python. They're told to plough through the canonical and, if their case isn't covered, go away or place a bounty (if they can afford it) with a message "I hope someone reads this. I'd like to know about python please" that no-one who follows the python tag can see.
That wouldn't be a good model for a successful Q&A site, and it is not how SE works.
Related: A friendly reminder that duplicates should be the same QUESTION, not different questions with similar answers. A large majority clearly agree with this position, but unfortunately it only takes 5 users determined to go against the consensus to prevent people from getting the answers they need before they travel.
