23
$\begingroup$

In general, most Stack Exchange sites do not deal with recommendation questions such as:

  • What is a good book to teach me x
  • Please recommend a widget to help me foo my bar
  • Where can I find more information about z
  • Where can I buy part x for my y

Excluding sites in which recommendations are the entire scope, there are a few sites in which question of this nature are acceptable.

So there are questions here?

  1. Do we want these types of questions to be on-topic or not?
  2. If so, what limits do we want to place on such questions?

This question was inspired by Magnetic levitation. After the background, the question boils down to "where can I find literature about this".

$\endgroup$
  • 1
    $\begingroup$ Here is a great example where a question that was not phrased as a recommendation request received a thorough answer that included some recommendations anyway. Readers take note: If your question is fundamentally interesting and useful, and you ask how and why, a good answer will give you enough to get started. $\endgroup$ – Air Feb 26 '15 at 20:27
18
$\begingroup$

No

Recommendation and sourcing questions end up attracting the following types of answers:

  • SPAM - including both blatant and thinly disguised
  • Ephemeral answers that will quickly become dated
  • Answers that are meaningful only to the asker and no other future visitor
  • link only answers

Generally, the answers those types of questions attract are very low quality.

Programmers SE has a really well laid out explanation for why those types of questions aren't welcome.

$\endgroup$
  • 3
    $\begingroup$ +1 for second bullet point. The vast majority of answers to these questions will go out of date. Book will be out of print, Widget will no longer be available, website will be down, etc. $\endgroup$ – user133 Jan 21 '15 at 7:55
16
$\begingroup$

No.

As engineers and students of the discipline, we ought to know how to do some basic investigation into X. When we need clarification on something, we can ask "What is Y?" or "How do I Z?" Specific, answerable questions about engineering problems will provide answers useful to an engineer. 90% of recommendation questions give instead terrible answers like "A is my favorite", unexplained links, lists, etc.

$\endgroup$
15
$\begingroup$

Qualified No: Exception for questions about relevant code(s)/standard(s)

The reasons (copied below) not to allow these kinds of questions outlined by GlenH7's answer are well considered.


Recommendation and sourcing questions end up attracting the following types of answers:

  • SPAM - including both blatant and thinly disguised
  • Ephemeral answers that will quickly become dated
  • Answers that are meaningful only to the asker and no other future visitor link only answers

The bottom line is these questions do not attract the kinds of answers that make for a world-class Q and A site designed for the maximum utility of questioners.

However, I'd like to submit that there is at least one type of Finding Things Question (FTQ) that should be allowed for an engineering website:

What is/are the relevant standard(s)/code(s) / where can I find the relevant code(s)/standard(s) for X kind of problem/project in Y location?

I don't see how it's possible to be a world-class engineering Q and A website without allowing very important FTQs such as the below, for example:

I am a bridge design engineer in country-X and have always used codes X-1, X-2, and X-3. However for personal reasons I am considering a move and would like to become familiar with the bridge design codes used in country-Y. What are the relevant codes and organizations I need to become familiar with?

Some of the answers to these specific FTQs might be ephemeral, however:

  1. I think that this should be a price the site is willing to pay in order to provide extremely relevant information in questions which DO have a correct answer, and are NOT subject to the other two bulleted problems, i.e. SPAM or not relevant to many people.
  2. The answer to this type of FTQ can be updated at the point in time the information goes out of date.

It is also true that many of the answers to this type of FTQ may simply be a link or a list of links (to the code(s), or simply to the relevant organizations/professional societies). However, again, I really don't see how the utility of a Q and A website for engineering is maximized by not allowing the answer (which might be a simple link!) to the question (for example):

What code do I need to consult for the design of a buried metal culvert in New Zealand?

This is a fundamental engineering question. Why - in 2015 - should I have to pick up the phone and call New Zealand, when I could just get an answer to my question on Eng.SE? It's true that the answer may only be the answer today. So what? The relevant code is probably among the most important things, as an engineer, I need to know.

$\endgroup$
  • $\begingroup$ How do we draw the line between providing answers for relevant standards versus becoming reference / research librarians? And how do we avoid link-only answers which become subject to link rot? $\endgroup$ – user16 Feb 2 '15 at 15:55
  • 3
    $\begingroup$ A code or standard is a very specific kind of reference, and it isn't just research. It is, in essence, an integral part of the question "I am in x-locale. How do I design widget-y?", to which an essential part of the answer must be "Your design has to conform to this code and these standards." I suggest the line be when the thing being asked for is a code or standard that the engineering profession itself (locally or otherwise), or a governing entity (or both) has determined to be THE governing standard that defines quality of care when practicing. $\endgroup$ – Rick Teachey Feb 2 '15 at 16:08
  • 1
    $\begingroup$ As for link rot, two things. First, I suppose you don't have to provide a link to answer such a question, you could just give the name of the code (although, again, links can be updated). Second, the applicable code or governing body itself isn't likely to change even though the contents of the code might change as it is updated. So for example, the applicable code for determination of wind load on a building in the US is, was, and probably always will be (for the forseeable future) ASCE-7, but there are of course different versions: ASCE-7 2005, ASCE-7 2010, etc etc. $\endgroup$ – Rick Teachey Feb 2 '15 at 16:14
  • $\begingroup$ I'd just like to add: there is nothing particularly ephemeral about codes and standards in engineering; design standards themselves (from which codes are derived) change all the time. As I said before, they are part of the answer to "How do I design X?", and if asking for the relevant code is considered too ephemeral, the answer to "How do I design X?" types of questions would often be as ephemeral due to the same changes in code. $\endgroup$ – Rick Teachey Feb 2 '15 at 16:26
  • 1
    $\begingroup$ Yes, yes and yes. Helping people find the relevant code is such an important thing. $\endgroup$ – mart Feb 2 '15 at 20:11
8
$\begingroup$

Yes

Many helpful discussions I found on eng-tips.com featured this: What is a supplier for piggable valves? What's a code/textbook/article that explains how to size a widget?

To take a look at the reasons for no given in the (as of now) highest voted answer:

SPAM - including both blatant and thinly disguised

Can be a problem, yes. We need good criteria for what's spam.

Ephemeral answers that will quickly become dated

Like how? I daresay that most physical stuff produced now was produced in a similiar way 5 or 10 years ago, and will exist in a highly similiar or identical way in 5 or 10 years. Answer to "what is a device that does X" will not age faster than answers to questions like "What industry code governs X" (and the latter are, IMHO, highly relevant for this site).

Answers that are meaningful only to the asker and no other future visitor

Again, I don't see this.

link only answers

Are obviously bad and should be deleted. But I don't see this as a reason to ban this sort of question.

To look at the types of questions

Ultimately, when I ask a question, I often don't know what shape the answer will have. My question on sizing heat exchangers for shear thinnging fluids could have a straightforward (and short) answer. Or the best answer to this question could be "get the text book y from Z, chapters bla and blo explain in depth" because two book chapters is the minimum length required to answer my question. I don't know.
In my work, when I look for a solution for a problem the solution often takes the shape of a product. Why reinvent a widget to foo my bar, when I have neither the time nor the expertise to invent foobaringwidgets?
Likewise, some widgets or textbooks may be pretty unique. How many textbooks specifically on repurposing refractury columns in oil refineries are there? I think I stumbled across one, once, and I wouldnt be surprised if it's the only one. Again, the asker might not know before asking.
Finding more information about z - Very, very often you have only vendor information about Z. I often wonder where to get un- (or just differently) biased information on a number of Z.
Where can I buy ... - I'm in favor of banning those. Answers will be advertising. Still, I often wonder where I can buy a certain type of pump or other machine. I want us to find a way in supporting this kind of question, without making the answers vendor specific.

Don't just think about the quality of the questions for this site, think about the relevance of the questions for your work.

$\endgroup$
  • 1
    $\begingroup$ I've asked this question as a testing balloon: engineering.stackexchange.com/questions/402/… (I also want an answer) $\endgroup$ – mart Feb 2 '15 at 10:44
  • 3
    $\begingroup$ +1 (I don't know why someone voted you down with no comment.) But I have no problem with shopping type questions. (I guess we are in the minority.) Finding a good vendor for a product is something that sometimes takes years. If someone else has a recommendation, I'm all for hearing about it and chasing down the details myself. I also don't mind "good book" questions. IMO good books don't go away, they get a new edition. $\endgroup$ – George Herold Feb 5 '15 at 18:22
8
$\begingroup$

Recommendations are not bad

Be sure to distinguish between recommendation answers and recommendation questions -- getting recommendations in answers is natural and normal for many types of questions, and is definitely a good thing. If someone is trying to reinvent the wheel in Javascript when they can just use jQuery or the like to do the same, pointing out that this can be easily done in jQuery may be a great help to the next 5,000 people who are trying to do the same thing and don't know about jQuery.

The problem with recommendation questions is that they usually aren't trying to solve a specific problem, which means that there isn't a specific answer. Often someone asking a recommendation question actually has a problem that can be figured out through clarifications.

Resource Requests

If someone asks, "What is the best book to learn about large-scale heat exchangers for oil refineries?" they may have a problem with a large-scale heat exchanger in an oil refinery that can be gleaned by asking questions like:

  • Which process?
  • What type of heat exchanger?
  • Information on installing, maintaining existing exchangers, or troubleshooting them?, etc.

If you ask those clarification questions, it may very well turn out that "What is the best book to learn about large-scale heat exchangers for oil refineries?" was really a question about "How can I troubleshoot heat exchanger tube corrosion on a large-scale heat exchanger used in process X?" And maybe part of the answer to that will be "Heat exchanger corrosion is often caused by problems with water treatment. The first things you want to look at are X, Y, Z. For a more general overview of heat exchangers in that specific process, take a look at this book/this website."

That is a good recommendation, and one that will help people in the future who have the same problem (or a similar one).

Product Recommendation Requests

A product recommendation request may also be something similar. Someone may ask, "What is the best valve to discharge condensate from a 50 psi steam main in batch operation?" Again, there is probably a question hidden in there:

  • What are you currently using?
  • Are you having trouble with condensate backup?
  • What schedule do you run batches on?
  • Do you shut down overnight? etc.

So after clarifying, it may turn out that the question is actually, "I am having trouble with condensate backup in my 50 psi 1/2" steam pipe on first start-up every morning after 8 hours with the equipment off. How can I prevent condensate backup after a long shut down?" And again, part of a good answer may be "You want to look at a manufacturer's steam trap sizing diagram like this or like this, and match it with the amount of condensate built up in your pipes after a shut down using equations like this or this." Or alternatively, "This steam trap manufacturer recommends installing a bypass valve after long shut downs to manually purge the initial condensate"

These are also good recommendations, even though they are recommendations.

When recommendations are good

What both of these examples have in common is that they provide useful resources for further information on how to solve a specific problem. They do that while also answering the question for the person here on the site even if those links do end up dead because they explain why they are also recommending those resources.

I hope that when you see a recommendation question you don't reflexively close it as off-topic, but that you think about what sorts of underlying problems the person asking for the recommendation may have, and try to clarify the problem to salvage the question and leave a better resource for people going forward.

I wouldn't worry about product recommendations in answers too much either. If I ask a friend for advice on an engineering problem, and that friend gives me a specific product that has worked for them on the same issue in the past, that is a good thing. What we want to make sure is that those recommendations are organic, and not people trying to sell more of their product. See how not to be a spammer.

$\endgroup$
  • $\begingroup$ I think that your examples show that recommendation questions are bad. They are only good if there is an underlying question that should have been asked instead of the "too broad" or "recommendation" one. This sounds like the exact reason why questions should be closed and improved. $\endgroup$ – hazzey Jul 10 '15 at 23:18
  • $\begingroup$ @hazzey I think we agree on everything you just said. Recommendation questions are generally bad as-is, but usually have a good question lurking. Just want to make sure we don't lose sight of that even if we do agree that the question as-is are bad. $\endgroup$ – jmac Jul 13 '15 at 0:46
1
$\begingroup$
  • No: Opinion.
  • No: Opinion.
  • Maybe: Yes, if it's an unambiguous fact-based question such as where to find/purchase certain engineering standards or learn more about certain engineering societies. No, if the answer would be "Google it" or "there's this good book I once read."
  • No: Opinion. ...unless they want to buy something truly obscure such as a plutonium rod for a nuclear reactor... but I think that may lead to different issues ;)
$\endgroup$
  • 1
    $\begingroup$ Agreed. Codes/standards and professional societies are SO IMPORTANT to the profession, it would be a massive lacuna in the utility of the website to not allow requests for this information. $\endgroup$ – Rick Teachey Feb 3 '15 at 4:51

You must log in to answer this question.

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