I have edited this question to bring it to the top of the meta-SA list, because it never seems to have come to consensus, and is relevant again (March 2013).

We have had two sourcing questions lately:

[Moved my argument as to why all such questions should be off topic into a new answer]

Is there any circumstance when a sourcing question is a good value add for SA?

Can we reconsider this question, and come to a consensus on how to handle sourcing questions?


Aaronut's original version of the question below:

By "sourcing" in the title, I am referring to questions along the lines of Where can I buy/find X, where X is any culinary item.

A recent question, What is a good place online to buy quality Sterling Silver Dinnerware?, picked up a few "off-topic" flags and a close vote. I have to confess that this puzzled me, because utensils are culinary items and our community has always allowed these "sourcing" questions in the past.

A search for the keywords "where find" or "where buy" in the title turn up 12 and 15 prior questions respectively, most of which seem to have been pretty uncontroversial. Some of them are practically identical in their format, such as Where can I buy online spices, sauces, extracts, etc. with international shipping? or Cheapest place to get bulk chocolate?.

I asked about it in chat and some of the comments revolved around these previous questions appearing to be low-quality or treated as polls, where everybody answers "here's where I buy them" as opposed to working toward a comprehensive list.

Those arguments are fine, and it's never too late to revisit a policy if we see a problem emerging - but we have to be consistent. We're already asking a lot of our newer members to understand the sometimes hazy distinction between good vs. bad recipe questions or "What can I add to X" - I don't think we'd be doing anybody a favour by closing "bad" sourcing questions based on a Stewart test.

So, things to think about:

  • Is there a fundamental, structural problem with sourcing questions like the one linked at the top of this post? Are they always polls, and therefore Not Constructive?

  • If we assert that these are polls, is there any valid formulation of the question? For example, we don't allow shopping questions of the form "What's the best X?", but it's OK to ask "What features are important in an X?" (equipment) or "How can I spot a good X?" (ingredient selection). Is there an equivalent when somebody is searching for a supplier?

  • Is it OK if the thing being sourced is rare or difficult for a layman describe, such as an automatic crepe maker? Can requests be too specific, such as some random tea set?

  • Is there a difference between asking for online suppliers vs. local suppliers? Why or why not?

  • One of our principles on Stack Exchange is that questions should demonstrate some preliminary effort to solve the problem, which is inherent in our rules about recipe requests and general reference questions. Does that apply in any way to sourcing questions - could it be the deciding factor in whether to close or leave open? If so, how is somebody supposed to demonstrate this effort?

P.S. Please note that I am not advocating a policy change here or suggesting what it should be. I'm simply pointing out what appears to be a current state of inconsistency and confusion, and opening up the discussion for others to comment on.

If nobody has anything to say, we (hopefully speaking on behalf of all the mods here) will just continue to leave these questions alone, unless there's something else seriously wrong with them and/or the community starts consistently voting to close them.

Thanks for posting this. You've said it much better than I could have. –  yossarian Dec 31 '11 at 21:21
Dunno about the overarching question, but the specific silverware question seems off-topic to me because it has nothing to do with food or cooking. As far as the food is concerned, it doesn't matter if you eat it with a plastic spoon, a stainless-steel fork, or those organic doohickies attached to your hand. –  Marti Nov 1 '12 at 21:09
If a question like this did lead to polling and discussion, could it not be converted to a wiki in some cases? –  Preston Fitzgerald Mar 14 '13 at 22:54
@PrestonFitzgerald: That is an ancient loophole, long since closed on I think pretty much every Stack Exchange site by now. Wikis aren't used for polls, they're used for wikis, which is why the checkbox isn't there for questions anymore. –  Aaronut Mar 14 '13 at 23:54
Got it, thanks! –  Preston Fitzgerald Mar 15 '13 at 0:57
add comment

3 Answers

It turns out that most people aren't any good at searching the web, especially once what they're looking for is becomes non-trivial. This, I suppose, isn't a surprise to anyone who's spent any time answering questions on the SE Network. And, although I feel I'm usually pretty good at finding stuff on the Internet myself (citation needed, certainly), there are times when, no matter how hard I try, I just can't find something. But I'm pretty sure it exists. Somewhere. But I know not the magic Google incantation to find it.

I think there are a couple of ways an expert can add real value by answering a sourcing question:

  • "Yeah, in the industry, we don't call those X, we call them Y. Search for that, and you'll find them."
  • "It turns out, you don't actually want an X. If you use Y and Z together, it does that so much better, and is cheaper too."
  • "Find a local home brew store, they carry that type of thing."
  • "No one makes that, because whatever"

Questions that provoke those types of answers are, I think, clearly valuable, and work well in the SE model. They also do not suffer from any excess localization, poll-like qualities, etc.

Honestly, I think we ought to accept any sourcing question where its clear the asker has at least made reasonable attempts to find the item. Especially since, it seems, most people really do have a hard time STFW and its not like our question volume is anywhere near overwhelming. But I think we ought to encourage answers which tell a person how to find a vendor. We can do this by editing the question and—in case I haven't been controversial enough yet—deleting answers that do otherwise.

(Exception: I think its fine to point to a particular vendor if its basically the only option.)

... and, to give my opinion on the borosilicate question, allow it. I've tried that quest before, and indeed, its impossible to find. I gave up when the only answer I found was labware.

If you accept the principal that the quality of the answers are not relevant to whether the question is on or off topic, how do you use these critiera to distinguish the good sourcing questions from the bad? It seems it is the answer that governs not the question? –  SAJ14SAJ Mar 19 '13 at 19:48
@SAJ14SAJ Well, I'd distinguish off/on-topic. "Where can I buy lumber?" is off-topic; "Where can I buy flour" isn't. [Both are bad questions, of course.] But for many other things, I think its primarily the quality of the answers that matters. And then we close (e.g.,) poll questions because we expect, from experience, the answers to be bad (and if we see something that looks like a poll question, but the answers disagree, then we don't close it—because reality wins—but of course we edit it to make it less likely to get poll-type answers in the future). –  derobert Mar 19 '13 at 19:59
Okay, personally I have to disagree--to me Aaronuts point that the quality of the answer should not determine the on-topic status of the question is pretty persuasive. –  SAJ14SAJ Mar 19 '13 at 20:31
@SAJ14SAJ then do you want to close the storage lifetime question, as its clearly either a poll or vastly over-broad? –  derobert Mar 19 '13 at 20:34
Actually, yes, under current guidelines as I understand it, we wouldn't allow this question today as a new question. I think (if available) the thing they do on SO where they mark it as "of historical interest" would be the most appropriate. On the other hand, I would be very open to changing our guidelines on open, wiki, and list type questions when they are culinary relevant and add value--which to me means doing more than substituting for google. There was a question on best vegetarian cookbooks which similarly isn't allowed today, but I would like to allow. Its a list, too. –  SAJ14SAJ Mar 19 '13 at 20:37
@SAJ14SAJ Sounds like we need to open a new meta question for this, we're both on chat so we should probably figure out its scope there. –  derobert Mar 19 '13 at 20:40
It occurs to me that 3 out of 4 of your example "value-added" answers are actually not direct answers to the canonical sourcing question, and the 4th is only going to apply in very limited scenarios (i.e. ingredients or equipment normally used for things other than cooking). That's usually a hint that the essence of the question is important but the formulation is wrong. I think if we are to seriously consider the opinions expressed in this answer, we need to be looking for phrasings which clearly request these types of answers and exclude localities and brands. –  Aaronut Mar 31 '13 at 3:05
add comment

I would argue that all sourcing questions fall into one of two categories:

  • They are extremely local and require extremely local knowledge
  • They are better served by a general internet search with google, bing, or whatever

None of them require specific culinary or cooking knowledge, as opposed to a question on how to choose an ingredient or piece of equipment. It seems to me that allowing what are at their heart "scavenger hunt" questions to SA is not a value add.

Additionally, all sourcing questions--at least in the abstract--have an unbounded number of answers of which none of them is clearly a "better" or "right' answer. Everything is relative to the reader, so they become either lists, or subjective.

While they are probably only asked for more rare or hard to find items, that doesn't mean that there is a right answer in general--and we already have rejected the notion (in comments to answers to the original question) that the quality of the answers matters to the relevance of the question.

I would submit that therefore, sourcing questions should be off topic as a general rule, unless there is some compelling reason (which I cannot think of) why they would not be.

In the older comments, there is a suggestion that a store that is open for six months is not a specific moment in time. While I would argue that in fact it is a specific moment in time compared to the lifetime that the question and answer will live on SA, it is also only relevant to the people who live near that store, and is thus very local. Either way, "too local" applies in my mind.

I would tend to agree, although sometimes (rarely) there are not-too-localized versions potentially dealing with what kind of store to get something from (just off the top of my head, for example, I seem to recall at one point someone recommending to get a crème brûlée torch at a hardware store, not a kitchen/appliance store, which was good advice). But for the most part, this is all true, the salient point being that they largely don't require any culinary knowledge or skill at all to answer. –  Aaronut Mar 14 '13 at 23:48
Along the same lines as Aaronut's exceptions: questions about how to search for something ("what are these vertical-walled small tart 'pans'?") which do have an answer. –  Jefromi Mar 23 '13 at 11:38
add comment

I would say that it depends from the answers. In What is a good place online to buy quality Sterling Silver Dinnerware? I would expect the answers to just report a particular place where the user goes, and the answers would just be a list of stores, which could make the question too localized. If you report a store you find in USA, that would not help me, as I live in Italy; vice versa, if I ask a question about a store I can find in Italy, that would just help who is in Italy (and maybe in the area close to where I live).

Suppose that somebody asks where to find an ingredient, and an (hypothetical) answer is the following one:

That ingredient is used in Japanese cuisine; if you ask for that in a Japanese restaurant, they will give it you for free (or paying a few dollars).

That answer is generic enough not to report the name of a store, or a restaurant, and it would be valid for more than one place. You could argue the question is too localized, if the same doesn't apply for other countries.

These questions remind me of the single-word questions asked on English Language & Usage where somebody asks for a single word to use to express a concept; sometimes the questions seem absurd, and you would bet there isn't a word that can be used, but then there is somebody who suggests a word to use, and that can be really used.
To decide if the question should be closed, you need to wait for any answer. If the OP keeps rejecting the proposed answers, then the question could be closed as too localized; if nobody gives an answer, then the question could be closed as too localized, assuming that nobody answered because there isn't a word that matches the requirements, and knowing that there isn't a word for those criteria is not useful to anybody.

If you are looking for criteria that is always valid, then every subjective question where all the answers are equally valid are non constructive. Questions asking for "the cheaper," "the best," "the more convenient" are then too localized, IMO. Even the question about where to buy buttermilk in France is localized; where to find it depends from where you live in France, and even in that case it could change, if a specific shop stops to sell it.
I don't think that the topic of a "the best thing" question makes the question more constructive than another "the best thing" question; it would also be difficult to explain to the users why a question is accepted, and another one is not accepted, when both the topics are fine for the site.

Thank you for your insight, although I can't agree with the recommendation. The criteria for whether or not a question is valid should never depend on the answers it receives; it's unfair to askers and unworkable as a policy. If a certain type of question frequently receives very poor answers then that may be a red flag (and it was, in fact, part of the motivation for this meta discussion); however, I don't agree with that being a valid test on a case-by-case basis. –  Aaronut Jan 1 '12 at 21:39
I don't think there are more restrictive criteria from the "not constructive" and "too localized" ones. Or you are going to close all the "not constructive" and "too localized" questions, or you should accept that some questions the community of Seasoned Advice accepts some questions, and dislikes other questions. The comparison with EL&U is because that is another SE site where subjectivity is part of it; in the same way a word to use in a specific context is subjective (which means there are more than one word that could be used), there is more than one way to do something while cooking. –  kiamlaluno Jan 1 '12 at 22:33
I think you're misunderstanding "too localized". Questions are allowed to be localized, on every site; they only become too localized when they are only relevant to the one person asking it, or maybe a smattering of other people in close proximity. A major metropolitan city is not "too localized"; an entire country is definitely not. I also don't really see what subjectivity has to do with it - that's never been a problem - although the issue of all answers being equally valid is one worth considering (and would therefore imply that the questions should be treated as polls). –  Aaronut Jan 2 '12 at 1:08
"Too localized" is not only referring to the place, but also to the time. If the question is about a store where to buy something, the answer is localized in time; it is valid until that store is not closed. "Too localized" is then not defined as "relevant to one person, or other people in the close proximity." (Close proximity is then a relative concept; what is close to somebody is not close enough to somebody else.) –  kiamlaluno Jan 2 '12 at 13:12
Subjectivity is not a problem. In some sites subjective questions are not allowed at all; in other sites subjective questions are partially allowed. In the latter, there are still subjective questions the FAQ suggest not to ask, including the ones where every answer is equally valid; in those sites, the users will have to accept that not all the subjective questions are accepted. –  kiamlaluno Jan 2 '12 at 13:14
Again, the key word in "Too Localized" is too. If a store is only open for 6 months, that's still far longer than a "specific moment in time". As far as every answer being equally valid, that's not exactly true; more like every correct answer is equally valid, and also independently verifiable, which barely registers on the subjectivity meter at all. I think the issue seems to be that questions of the form "Which stores sell X?" receive the same answers as "Where do you like to buy X?" in practice, whereas in theory they are different questions, the former requesting a full list. –  Aaronut Jan 2 '12 at 19:26
This discussion goes all the way back to the inception of the Gaming site, where Grace Note referred to these as "Repository" questions, which every site still sort of handles differently. The best argument I've heard against the class of question is not subjectivity, nor localization, but simply the fact that the SE engine really sucks at dealing with questions where there are more than about a dozen correct answers. –  Aaronut Jan 2 '12 at 19:29
The only questions that should not be acceptable are subjective questions where all the answers are equally valid, not constrictive, and too localized questions. To notice that the description for too localized ends with the following phrase: "an extraordinarily narrow situation that is not generally applicable to the worldwide audience of the internet." –  kiamlaluno Jan 2 '12 at 20:24
I don't see any problem with non subjective questions where there are more than one correct answer. With Drupal there is more than one way to resolve a problem; I don't think that non accepting a question because there are more than a correct answer is something I would like to see done. So far, users have been handled questions with more than one correct answer: Users vote the answers, and the OP accepts the one that helped him more. –  kiamlaluno Jan 2 '12 at 20:28
add comment

You must log in to answer this question.

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