I'm a big proponent of TeX. I want it to be accessible to a broad family of people, not just those who are sufficiently technically-minded to understand its inner workings, and/or understand the documentation (which, let's face it, is pretty poor much of the time unless you are already a TeX whizz).

TeX.StackExchange is an amazing resource for those who are not technically able, both in the existing answers to questions, and in the ability to ask a new question. It has, to a large extent, replaced the not-always-useful documentation supplied with different packages.

However, as is often the case in technical subjects, there can be a gap between what a user (asking a question) needs to know, and what an expert (answering the question) supplies. Some examples:

  1. Requesting an MWE. Often a valid request made by the expert before considering the matter any further, but does the user have the capacity to provide such a thing? For example, in "Draw this for me" etiquette, one comment is that all questions need an MWE. If, say, a user has never heard of TikZ, then how are they going to provide an MWE? Isn't a mockup of the picture sufficient in such a case?

  2. Referring to a similar (but not the same) post. Does the user understand the answer in the other post well enough to enact the modifications? As TeX.SE grows its library of questions and answers, perhaps the calls for help are for experts to carry out this modification?

  3. Directing someone to the manual. At least give a link and a page number (e.g. the TikZ/PGF manual is 880 pages long! (Edit: it's now 1321 pages long!)), but is the manual actually user-friendly, or only expert-friendly? Perhaps they came to TeX.SE in the first place precisely because they found the manual impenetrable?

  4. Criticizing users for not doing "research" before asking. Sure, there are lazy users out there, just wanting a quick answer and letting someone else do the donkey work. But there are also users out there who spent 2 hours trying to get something to work, but without knowing the search terms to plug into google/TeX.SE they don't know where to start.

  5. Not answering the question, but pointing out all the other mistakes the user has made. A user wants something fixed up, but they're using (e.g.) $$...$$ instead of \[...\]. As a user, how does it feel to ask a question, not receive an answer, but get told everything else you're doing wrong? In this case, I think the etiquette should be clear: answer the question first.

This is not a criticism of anyone in particular, but really just a question: Now that TeX.SE is, for many users and experts alike, the place that they learn many aspects of LaTeX (rather than in package documentation), do we need to adopt a more empathetic approach? Do we risk turning people off LaTeX for being too technical, simply because we sometimes expect something of the users that they are not in a position to give?

Of course, I can't guarantee that we will always be able to tell the difference between a lazy user and one who has literally told us everything they can, but much of the time I think it's fairly clear.

  • 11
    @AndréC Though quite often the asker is not at all interest in an explanation, only in the solution, so people (with people, I include myself) post a “look how brilliant I am” answer that solves OP's problem with no explanation whatsoever (what's the point of writing if no one's going to read?). However it also happens that sometimes people ask for an explanation, like here (see the edit history), and then the answerer usually provides an explanation. In my (not so) humble opinion, people in this site are in general quite kind and receptive. Jan 30, 2020 at 17:53
  • 10
    Commenting just on point 5: Sometimes the code given is so contrary to LaTeX conventions that it's all but impossible to tweak just to give the looked-for result. In such a case, I sometimes suggest changes to make it conform to conventions because then it will be easier to fix, but if I can't solve the actual problem I will leave that for someone else. (I do try to be polite.) If I do answer, I try to be thorough enough that the answer will be useful also to others with similar or related problems. A one-shot tweak might mislead a future reader. Jan 30, 2020 at 18:57
  • 1
    @barbarabeeton sure, sometimes that's certainly the case. Those weren't the sorts of case I had in mind, it was more the "you should be using X instead of Y" which may be fine, but may also just indicate a brusque approach to those who know less.
    – rbrignall
    Jan 31, 2020 at 9:15
  • 3
    There’s always danger that answers offered in a generous spirit will be read as brusque, and that questions following much effort will seem lazy. We come from different places, with different conventions of politeness, we’re more or less at home in English, we can’t see facial expressions or hear tones of voice, etc. My habit is to refrain from posting if I’m not in a good mood or if a question seems stupid (sometimes it’s that the question is over my head). That may reduce but not eliminate the problem.
    – Thérèse
    Feb 1, 2020 at 16:20
  • 1
    We all have our idea of what a TeX user is (or should be), mine is described in this paper I wrote as a sign of gratitude for D. E. Knuth: tug.org/TUGboat/tb39-1/tb121haralambous-knuth80.pdf I believe a TeX user is somebody caring about typography, document esthetics, precision, communicative efficiency of the document. That's our credo and obviously our interactions on SE are influenced by it. We should be able to transmit our care about the written word without discouraging the new user, without making everything look more complicated than it is… Thanks for raising the issue
    – yannis
    Feb 2, 2020 at 12:31
  • 2
    I just want to say that I really do not think that your (implicit) notion of "lack of empathy" is appropriate for the things you describe. You may very well have come across some users/behaviors that may be described that way but the actions you describe are performed mostly by users with a great deal of empathy and enthusiasm to support others.
    – user194703
    Feb 2, 2020 at 23:57
  • 1
    @Schrödinger'scat no I didn't say that. I'm really asking the following question: In order to be more helpful to newbies (and others), can we empathise with them more than we currently do? The math forum is quite different, as it features a lot of people wanting answers to their homework (and in math, having a go for yourself is how you improve).
    – rbrignall
    Feb 5, 2020 at 11:55
  • 1
    Who is "we"? How do you know how much any user empathises? My main problem is that you only discuss things from the point of view of the one asking a question. How about the many users who read through question and answer to solve their problem? This site is for those users, too.
    – user194703
    Feb 5, 2020 at 12:02
  • 1
    "We" is referring to the people who read the question and consider answering it (or, perhaps, don't answer it, but still feel empowered to say something on the post to make a newbie feel unwelcome and/or stupid).
    – rbrignall
    Feb 5, 2020 at 12:18
  • 2
    I am sorry, but I find these statements are really off. Because there are a handful of users who misbehave you make very general statements which criticize everyone.
    – user194703
    Feb 5, 2020 at 13:07
  • 3
    I have, quite explicitly, not cricitised anyone in particular, because I don't want this to be personal. We - me included - can and should always strive to do better. Don't you agree?
    – rbrignall
    Feb 5, 2020 at 13:09
  • 3
    You are trying to distract from the point. The question title "Empathy for those who don't know latex well" has a very clear message: there is not enough empathy. And it will be used against those who answer questions and provide sustainable help. But yes, I fully agree that we always should strive to do better. To start, you could edit your question to make it less biased and try to set up things in such a way that benefits all users, not just you.
    – user194703
    Feb 6, 2020 at 2:28
  • 2
    @Schrödinger'scat none of what I posted in my original question is about benefitting me: I'm not in the category of those who "do not know LaTeX well" (though there are plenty who know more than me), and I will always (e.g.) provide an MWE for any question I post. You may disagree with my points, but please don't resort to an ad hominem attack. There is no subtext to my question: I have cited a few examples of things which I have seen, as examples to illustrate my broad point, which is that we should be doing more to help newbies, so to avoid chasing them away.
    – rbrignall
    Feb 6, 2020 at 9:24
  • 3
    As I said, this is a broad attack that, in the way it is presented here, goes against everyone. I disagree with this way of dealing with the fact that there is a handful of users who overdo it. Plus, as I said, you miss to discuss the issues from the perspective of users answering questions. In the comments you seem to paint a very different picture from what is written in this above post. I stop here. Do what you want but do not think you help the majority of users.
    – user194703
    Feb 6, 2020 at 15:25
  • 2
    @Schrödingerscat it really is not what you describe, and I'm sorry you've interpreted it that way. Perhaps this is just a symptom of the fact that the tone intended when something is written is not always the tone that comes across, which actually fits rather well into the theme of this question.
    – rbrignall
    Feb 6, 2020 at 18:34

8 Answers 8


I'll take the opposite approach to the other answers (so far), addressing the question of empathy more broadly (rather than the specific 5 points), as I've been thinking about it for a while. I think there definitely is a shortage of "empathy" sometimes, though this site is much much better than most other sites on the Stack Exchange network, or even internet forums generally.

The problem IMO has not much to do with whether the user knows LaTeX well. Plenty of "good" questions from obviously novice users do get empathetic comments and answers that are appropriate to the user's level of knowledge. And there are also occasional "bad" questions from users who may have some expert level of knowledge, but (because of the way the questions are asked) are treated with not much empathy. Rather, the pattern I see, which is mostly just an inevitable consequence of people being human, is the following.

When someone "joins" this site and starts "participating" by looking at questions other than their own, what is their motivation? There could be many reasons why you (not specifically you who posted this question, I mean a generic "you") may start participating on this site: Maybe you just like helping people in need. Maybe your curiosity is piqued by the questions, or you like solving interesting problems. Maybe you have all this esoteric TeX/LaTeX knowledge, and want to put it to some use. Maybe you want to show off how brilliant you are. Maybe you want to "give back" to the community, having benefited in some way. Maybe you just like competing for internet points. All of these are valid motivations! They can all (in principle) be good for the site.

But over time, as you start looking at various people's questions, something happens to you: you become more discerning, and can clearly see that some questions are good and some are poor (by the standards of the site). This is because our brains do that with anything (xkcd 915: Connoisseur). There is reasonable consensus what a "good" question is:

  • The hover text on the upvote button says "This question shows research effort; it is useful and clear". This is generic across StackExchange sites.

  • On top of that, there are community-specific norms, like (here) the question includes a minimal example (wherever it makes sense), shows that the questioner has made an effort to solve the problem by reading relevant documentation, is otherwise well-written or well-stated except for the part the question is about, etc.

So far, all very good. These considerations influence (as they should) whether or not someone clicks the up-vote button. But beyond that, when it comes to actually interacting with the person who posted the question, that is where the question of empathy arises.

Because you see, you know what good questions and bad questions are. You have seen the entire spectrum of users. When you see a poor question (and that too again and again), it is natural to get annoyed, consciously or not. Grumble grumble, why can't this person state their question well -- look at the previous person: they asked their question so well? Questions like this are not what the site needs. So by now, especially if you strongly identify with this site or community, you may think to "improve the quality of the site" by telling the user how to make their question a "good" question.

But the person asking the question (typically not someone who has read enough of others' questions to have absorbed site norms) does not know or care. They are a separate human being from everyone else who came before and posted other questions, with their own goals and hopes and desires and attitudes and culture and everything. They have no thought or conception of "good questions", the quality of this site, or any of that. They are just here because they had a problem and thought they might get some help. You may think you are helping them to ask a better question "in future" or "the next time", but there likely won't be a next time -- more likely they'll just give up on the site, or even TeX/LaTeX itself, and move on to something else.

This IMO is the crux of the empathy problem: the disconnect between site regulars judging new questions and worrying about the site quality, and question-askers who are removed from all that context and just want help. From the asker's perspective, they came for help but got all this nitpicking about their question. From the site regulars' perspective, they gave the asker all this concrete advice on improving their question, but they showed no interest in improving the question and just went away. A mini-tragedy playing itself out day after day.

So what's the solution? I'm afraid there's no perfect solution, but I can think of a few things that we can keep in mind that may help. Well, I try to keep them in mind anyway:

  • While it's good to improve the quality of this site or questions thereon (after all, most of the site' usage/help over the long run will not be to the person who asked the question, but later visitors with the same or similar problems), never get so involved in this secondary goal that you forget your original motivation (helping people or whatever).

  • Always remember that the person you're responding to is a real person, different from everyone else, not just a generic "question-asker" or the same as their question itself. While evaluating questions as good and bad is fine and possibly important, mentally keep this evaluation a separate step from the act of communication. Before writing anything, try to invoke a feeling of empathy with another human being. If the only thing you can think of saying is something not nice, consider not saying it at all; someone else can deal with this question. One more sub-optimal question on a random internet website is not the end of the world.

  • This one is more subjective than most, but avoid pre-written text templates. I used to think they save time and can ensure everyone gets a consistent response, but I find I can manage better empathy (or at least be better aware of what I'm communicating) when choosing words afresh, in reaction to the question. The little details help.

  • Look for points of intersections between the user's needs on the one hand, and your/the site's needs on the other hand. Focus on and emphasize them. For example, instead of something to the effect of "No MWE. Bad question" (even if stated very politely) (remember, making questions "good" is not the user's immediate need), you could emphasize the benefits to the user (if you can honestly judge them to be true in this instance): "Welcome. It might be easier to answer your question if...", etc. By the way, from this point of view, five of your points can be either acts of generosity or of unhelpfulness; the difference is only in the intent and communication:

    1. Requesting MWE: Help the user get their question answered.

    2. Referring to similar post: Helpful if in the spirit of "in case you haven't noticed, have you seen this question? Does it help you? If not, why not? [let's fix that]"; unhelpful if in the sense of "already answered (even if you can't figure it out), go away"

    3. Pointing out the manual: Similar. Note the user may have already looked at the manual but just forgotten to mention it or say what the difficulty was (it ought to be clear that "nothing makes sense to me" is fine too); bringing up the manual focuses the question and makes it easier to be helpful.

    4. Criticizing users for not doing research: Well, merely criticizing per se is unlikely to be helpful, but pointing out how to do research, and that many of the people best positioned to answer questions are likely to respond better to well-researched questions, can be a way of helping.

    5. Pointing out stuff other than what the question is about: Again, can be either generous or nitpicking.

    The difference is all in the intent, and in the care taken to avoid coming off as rude, and sensitivity to what the user's immediate needs seem to be -- in short, in empathy.

  • Finally, retain your humour :-) Things won't be perfect: users will continue to post awful, lazy, poorly researched questions (by the standards of good questions on the site), and you yourself will often fail to be empathetic or helpful. It's fine, when you realize the situation laugh at human foibles and move on.

  • 6
    Nice analysis, one additional point might be that regular users get tired of pointing out the same issues with questions time and time again which makes the 1000th "don't post such long code" comment less nice (and more cryptic!). IMO this is mostly because the question interface on the site provides very limited guidance for askers - the MWE is not mentioned at all, the How to Ask box is very general, things like "show the code, show what the output is, show a mockup of what you want it to be" is not mentioned. Providing more guidance would improve the empathy issue a lot I think.
    – Marijn
    Feb 3, 2020 at 10:14
  • 2
    This is a great answer, and really gets to the key point: people who have been on here for a long time will have seen it all before, whereas those who are new are, well, new. It's not just the domain of this site either, of course: it seems to be human nature to do this.
    – rbrignall
    Feb 4, 2020 at 11:53
  • @Marijn That's a good point; it's harder to have empathy when one is dealing with the same crap again and again, and occasionally one snaps. (A possible solution to that is to just deal with less; I only look at questions from a few tags that receive not many questions.) Of course, it would be great to have more guidance to question-askers but one also wonders how much they'll read... Feb 7, 2020 at 0:42

Many of your sub-questions are addressed in other meta questions, but it's always good to have some further discussion. I'll post here some comments on each of them with links to the other discussions.

  1. Requesting a MWE This is an issue on which there is some partial disagreement. There are some who believe that (almost) every question should come with a MWE. These people sometimes reflexively ask for one even when the question could clearly be answered without one. Personally I don't think MWEs are always needed, but there are many cases in which it is simply impossible to help people without one. Furthermore, if the question seems to imply that the user has some code that's not working, it is entirely appropriate to tell them to turn that into a MWE. More discussion here:

  1. Potential duplicates The site is not helped by having the same question asked and answered over and over, which is why there is encouragement for users who know the site contents well to find duplicates where possible. The recent change in the wording of the automatic comment from "Possible duplicate of ..." to "Does this ... answer your question?" is a real step in the right direction, since it encourages users to check the potential duplicate but gives them a clear indication that if the question doesn't help them they can say so. Of course, closing happens a bit too quickly in many cases, but it's really not that hard to get a question reopened. There are almost always enough users in chat to provide the requisite votes if needed. Also, if you think a potential duplicate isn't really a duplicate, (for example, different questions can have similar answers, but are not duplicates) you can add a comment saying so, and vote to keep the question open (if you have such privileges). On the "closing too early" problem see:

  1. Pointing to the manual Frankly I don't see this as a big problem here in that I don't see many comments simply pointing to the manual. I agree, though, that such comments are likely inappropriate. That being said, it's definitely helpful to point out that the information the person sought was in the manual if you answer the question. It's also the case that many new users don't really know how to find documentation for the packages the are using. See e.g. How to find the documentation for a package?. Also, it's sometimes clear from people's questions that they are such a beginner that the only thing that will really help them is to point them to some beginner's resources. This is also entirely appropriate in my opinion.

  1. Criticizing users for not enough research This is fundamentally the "just do it for me problem" and people fall into three distinct camps here with respect to how to deal with it. Some people just answer the question. Some people first ask what the person has tried. Some people just ignore such questions most of the time. I'm in the last camp, generally. There's plenty of discussion about these types of questions on meta. See this question and the questions linked inside it.

  1. Not answering the question but pointing out other problems with posted code. This is a form of the "problem" of people answering in comments, and as has been discussed before, there are many good reasons for this. When people post comments about code errors, they usually say explicitly that the comments are not related directly to the problem at hand, but I think they help new users by not perpetuating bad code practices. As for not answering the question but commenting instead see:

Some comments on the user vs expert distinction

I think the 'user' vs 'expert' dichotomy is really a bit of a false one. Of course there are many of us on the site who have years of experience with TeX and/or LaTeX, and in fact many of the LaTeX core development team are active members of the site. But 'expert' is really a continuum. Personally I know quite a lot, but compared to others I am certainly no expert, yet I definitely have more knowledge and experience than many others too. But remember that the site is built on the goodwill and volunteer time of its users no matter how much or how little knowledge they have. That means that users with little knowledge do have a responsibility to try as best as they can to ask questions that have at least some effort put into them. Just because it's easy for me to answer a question doesn't mean that it doesn't take some time to do so. This fact alone is why MWEs are helpful: they reduce the time it takes an "expert" to produce a good answer, and this is a good thing for everyone.

Some comments about empathy

Overall, I'm not sure that we suffer too much from a lack of empathy for new/novice users. For one thing, from the very beginning of the site, there was agreement that downvoting should be used very sparingly, and comments should be used as much as possible to improve questions. See:

This practice is very much in effect still, and it's definitely a good thing, in my opinion.

I think also, that most of us do show quite a bit of patience with new users, as you can tell from the sometimes quite extensive comment threads asking for more details that arise on questions. On other sites these questions might be almost instantly closed as unclear, but we tend to only close as unclear if the questioner fails to answer the questions in the comments, or simply disappears.

That being said, I do think that we should all bear the points you have in mind when we deal with new/novice users. But there's one other thing that is also fundamental about making a community with many disparate views on interaction with others: you can't and shouldn't legislate behaviour, apart from the basic "be nice" edict. To see the diversity of views on this, however, see Should we "educate" (new) users by not answering until...? in particular this answer which reflects my views too.

So I think you're absolutely right in reminding people that they should take into account the novice user's perspective when they respond, but we should leave it as exactly that, a reminder.

  • 4
    Regarding "requesting a MWE", if the OP appears to be a really raw newbie, I instead say something like "How this is done depends on what document class you're using" and then a brief description of the contents of the "compilable example". It's a little less "institutional" than pointing to a canned response. Jan 30, 2020 at 23:39
  • 3
    @barbarabeeton I completely agree. I don't use any of the canned texts at all, and I don't ever mention the term MWE, but instead encourage people not to post code fragments but instead put the fragment into a compilable document.
    – Alan Munn
    Jan 30, 2020 at 23:42
  • 4
    Great answer! On duplicates, I'd add that the difference from a forum is that TeX-sx (or similar sites) are not about every person getting individual answers, but about collecting up ideas of general usefulness. That means that some questions are bound to be 'the same' as others, and comments may be needed to guide an inexperienced user to see this.
    – Joseph Wright Mod
    Jan 31, 2020 at 8:53
  • I like this answer, and thanks for taking the time to put it together. There are, of course, items on these separate issues, but my question was intended to be more "meta" than that: do we empathise with newbies? You're absolutely right there's no expert/user dichotomy, but there are a lot of newbies who visit this site once and never again. What they posted may look lazy and unresearchered, but may actually be the result of hours of agonising and not understanding...
    – rbrignall
    Jan 31, 2020 at 9:19
  • Great answer, Alan! Jan 31, 2020 at 13:00
  • I will add that the (somewhat understandable) obsession with demanding a MWE sometimes requires omitting context for which the poster is subsequently lectured for not having included or considered. I know tone doesn't translate well to ASCII text, but quite frankly I don't like the implication that I didn't consider X when I decided to omit mention of X in my MWE in the interest of minimalism. It's something I weigh every time I decide to post a question. I'm not an expert, but I'm also not an idiot (despite self deprecating humor at times). Jan 7, 2021 at 2:31

Taking the points separately..

1. MWE ... Isn't a mockup of the picture sufficient in such a case?

I would say here that a MWE is almost always an improvement, and should usually be asked for even though some people will helpfully provide a full drawing anyway.

It may be that the user is a beginner and does not know where to start. That is fine, but in that case they should ask "how do I draw a square in tikz" and they will get an answer or be pointed to a similar duplicate beginner question that has the answer they need.

It simply isn't reasonable to say "I am a complete beginner, how do I draw an entire technical diagram."

2. Closing duplicates.

It's just how the site mechanics work. It seems a bit odd at first, but it is not a criticism of the person asking the question. In fact rather the opposite, questions with lots of duplicates are weighted highly in searching the site. For future readers it is an improvement: when searching for their term rather than being shown dozens of more or less similar answers that they have to look at to realise they are similar, the site should have pre-filtered it, and show the one "canonical" example. That's the idea anyway, in practice it's always slightly subjective what is a duplicate (and whether someone answering wants to search for duplicates before answering).

3. Pointing at the manual

I think it's helpful to point at a relevant place in the manual if there is an obvious manual that has the information, but I don't think information being in the manual is reason for not answering the question here. More or less by definition all questions here should be covered by reading enough manuals, but that isn't the point. So manual references should be in addition to a specific answer.

4. Criticizing users for not doing "research" before asking.

There shouldn't be need to criticize anyone, basically there are no bad questions, only bad answers. But as commented in the first point, it can be the case that users are asking a too complicated question and they should start with simpler examples and just ask when they get an error. If doing that then they always have some code that they can show, and always have a specific tex question.

5. Not answering the question but pointing out other problems with posted code.

I commonly do that in comments (less so in answers). Seems no harm in being helpful, and sometimes the questioner has posted a lot of code, but missing important information like what document class is being used, the actual problem is unclear and the error described is due to code not shown. In such a case, you may as well comment on the code that is shown even if the comments are not related to the problem in the question title.

  • Comment on point 3: Pointing to a manual shouldn't be the only response, even if the answer is clear there. But if it's obvious that reading a manual would enhance the OP's education, then identifying that resource after providing something more specific is quite reasonable. Feb 2, 2020 at 21:18
  • 1
    @barbarabeeton yes that's what I intended it to mean, I'll reword a bit Feb 2, 2020 at 21:37
  • 2
    @barbarabeeton Pointing to the manual in a comment is IMHO perfectly fine as long as one is specific, e.g. the example at the bottom of p. 420 of pgfmanual v3.1.5.
    – user194703
    Feb 2, 2020 at 22:15
  • @Schrödinger'scat -- A newbie might not know how to find the manual, so a link or pointer would also be helpful. Feb 2, 2020 at 22:17
  • 2
    @barbarabeeton I do not disagree with that but I think a specific comment as above is not indicative of a lack of empathy. In a large fraction of the cases the comment is enough, and if not, one may reiterate.
    – user194703
    Feb 2, 2020 at 22:20

While I can see that there are communication problems, I feel that the question somewhat neglects one important aspect:

The answer and question should be useful for as many users as possible.

It may be true that one reason why some users ask for an MWE is that it saves them some time when writing an answer. I can speak only for myself, but this is hardly the really important reason. IMHO many of the points that you describe/criticize are not to save some time but really to make the thread useful. In more detail,

  1. Requesting a MWE is in many cases essential to define the question. It is very common that a potential answer clashes with some other package you load, and so on. An MWE will also allow future users to find more quickly the related post. So, in short, an MWE helps the one asking the question more than the one who answers it.
  2. Referring to a similar (but not the same) post. I cannot see anything wrong in pointing the OP to a related post. It happens quite often that this solves the problem. If it doesn't, all the OP has to do is to indicate that it doesn't, and why.
  3. Directing someone to the manual. I can't speak for everyone, but usually the comment does not just say "please have a look at the manual" but something like "see the example on p. xy of the manual". Very often (according to my experience) this solves the problem. Again, if the OP is not able to use this to solve their problem they can just indicate that it doesn't, and why.
  4. Criticizing users for not doing "research" before asking. This is a problem, yes.
  5. Not answering the question, but pointing out all the other mistakes the user has made. I guess this is a matter of perspective. While I understand that getting a comment that is not directly related to the main problem may be frustrating, you may want to look at the situation from the perspective of the one making the comment. In the case of most of the users the main intention is to help the OP (and not showing off, which applies to only a small number of individuals).

Overall I feel that your post is very lopsided in the sense that you only describe the things from the perspective of the one asking a question. Even more, you seem to assume that the purpose of this site is only to provide a quick fix for some problem. And you seem to complain that users do not explain things. However, explaining things requires some information about the knowledge of the one who is asking the question. To see what I mean you may want to watch this interview with Feynman. You can only give a good answer if the question is well-posed and it is clear what kind of information the one asking the question is really seeking.

To be honest, I find the title of your "question" really off, and also do not think that this is a true question. Rather, this seems to be a broad attack on those who answer questions. Why you suggest that anyone looking at the problems of the OP and giving feedback may lack empathy is really mysterious to me. It seems not to be a question, but (sorry to say that) only an attempt to show off and demonstrate how empathic you are. IMHO you are not. You do not seem to care at all about those who spend time trying to solve the problems of others, and try, using comments, to gauge which information the questioner is really seeking.

  • I don't disagree with anything you wrote. My question is about is how/when the above actions are appropriate, and recognising that there are times when (e.g.) requesting a MWE is simply going to send the user away never to return. This may be beyond them, so I'm just trying to ensure we all apply a "how much help does this user need" filter before we get annoyed with them.
    – rbrignall
    Jan 31, 2020 at 9:28
  • 4
    @rbrignall How am I supposed to interpret your comment? Are you suggesting not to ask for an MWE, even if it is crucial for being able to provide a meaningful answer? I am sorry to say this, but it seems to me that you insist on viewing at the issue only from one angle. How about those expert users who get driven away if there are too many low-quality questions? Shouldn't we also take into account their perspective in this calculation as well?
    – user194703
    Jan 31, 2020 at 21:23
  • I'm suggesting that, instead of immediately asking for an MWE, we consider whether a) this is really necessary, and b) whether this is something the user has the capacity to produce. In many cases the user should still give an MWE, but the accepted rule is that this should always happen, and that's what I'm questioning.
    – rbrignall
    Feb 1, 2020 at 9:49
  • 1
    @rbrignall I only ask for an MWE when it is necessary, in many/most cases I don't. I find "immediately asking for an MWE" really inappropriate because this is not what most users do. Really, try to see things also from the perspective of those answering questions.
    – user194703
    Feb 1, 2020 at 15:16
  • 4
    @rbrignall MWEs are crucial in many respects that I hope most of us know. But there is one that I believe is not mentioned often enough: the code is used as keywords for search engines. Let's take for instance a question whose crux is an incompatibility between two packages. If the OP provides a proper MWE, the whole page is likely to be well indexed by search engines; but if the OP includes tons of unrelated packages and text (as opposed to \lipsum calls), indexing will be polluted by many irrelevant keywords, especially if the people answering don't clean this up.
    – frougon
    Feb 1, 2020 at 17:32
  • 5
    @rbrignall And if the people answering waste time on cleaning up non-minimal MWEs (which I do all the time), this is wasted time doing something stupid, and is detrimental to other people whose sometimes non-trivial questions won't get answered. Not providing an MWE is very selfish, uncivil behavior—when you've heard at least once that it is useful for everyone.
    – frougon
    Feb 1, 2020 at 17:35
  • @frougon what you say is indeed the perceived wisdom, and has lots of merit. The search engine thing is curious: how does a user know what search terms to use in order to find what they need? Everything assumes they can do this, whereas in reality many posts nowadays are mild variants of existing ones, or people just wanting some very specific picture drawn for them. Also, if a question is answered, then presumably code will be provided in the answer, as well as hopefully an explanation what is going on. So, I don't think the search engine reason is as strong as you suggest.
    – rbrignall
    Feb 1, 2020 at 22:56
  • @frougon I think the argument for an MWE is quite simple: people asking questions are asking for the time of others, for free. If they want to increase their chances of a decent answer, or even any kind of answer, then providing an MWE is the right thing to do. I'm not sure that holds in a "draw this for me" or more open-ended question, or indeed someone who simply cannot provide a useful MWE because their skills are too limited even to know where to start. I'm pretty sure such people exist on here.
    – rbrignall
    Feb 1, 2020 at 22:59
  • 1
    @rbrignall You say: "how does a user know what search terms to use in order to find what they need". This is a problem, and this is why some of us link to related codes. According to how I read your post, doing so is indicative of "lack of empathy". My point is that this is not at all the case. Rather, if someone uses their time to search for related posts (by others) this is a high level of empathy. Low level would be to ignore the post.
    – user194703
    Feb 1, 2020 at 23:00
  • The empathy is in how that other decides to share that information with the user. Do they helpfully offer some possibilities and ask if that helps them to make progress, or do they wade in and criticise for lack of research?
    – rbrignall
    Feb 1, 2020 at 23:02
  • 4
    @rbrignall You will always be able to find abusive users. My point is that in the present form your post suggests that the majority of users answering questions and posting comments lack empathy. I disagree with that assertion most strongly. As I wrote in my answer, the purpose is to provide information that is useful for as many users as possible. I can only speak for myself but if I add a comment then I do that hoping that the OP and all others having a similar problem will benefit from that comment.
    – user194703
    Feb 1, 2020 at 23:06
  • 2
    @rbrignall You only gave the most selfish argument for providing an MWE (augmenting one's chances to receive an answer). There are tons of other good reasons to provide an MWE, which are not selfish. One of them is simply that when an expert is trimming down a document to prepare someone else's MWE, he (or she) isn't writing or maintaining LaTeX code or documentation (well, this argument boils down to wasting time of people who could use it for their family or in much better ways for LaTeX, etc.—that for easy things the OP could do himself).
    – frougon
    Feb 2, 2020 at 0:50
  • 1
    @rbrignall IMHO, if someone is absolutely unable to prepare an almost-MWE, LaTeX will always be a nightmare for this person. Better use a WYSIWIG tool in this case...
    – frougon
    Feb 2, 2020 at 0:50
  • 2
    @rbrignall "What I'm suggesting is that this may not be 100% of all questions": for instance I am only asking for an MWE when it is necessary, certainly in far less than 100%, and so do many/most of the users. I just happen to find your (implicit) notion of "lack of empathy" really out of place.You hit all those users who ask for an MWE when it its necessary only, who made hundreds of comments which solved questions (because the OP said so), and so on. Understand that this site is not about you, nor me, but about a large set of users who read through the posts.
    – user194703
    Feb 3, 2020 at 16:16
  • 1
    @rbrignall "LaTeX is not the domain of technical experts alone.": nobody says it is. Again, I feel you are really overdoing it here. You should go after the few users who are abusive rather than criticizing all users who answer questions.
    – user194703
    Feb 3, 2020 at 16:21

This answer will only deal with the 5th point "Not answering the question" and might perhaps just be an extended comment instead of a full answer.

With that being said the following is what I'd do if I had to meet a deadline and faced a problem I couldn't solve after investing several hours:

  1. Ask the experts (post on this or similar sites).

  2. Leave the question alone until there is response to it (be it comment or answer)

  3. In the meantime get as much work done as possible with the skills and knowledge I currently have.

Only if I get feedback on my question will I stop working on the project and care for the question asked. That approach is the most efficient way to spend my time.

What will be the outcome of all the work I do with my current skills and knowledge if no one points out that I shouldn't start displayed equations in LaTeX with $$? I'll continue to do so, resulting in inefficient time usage because I later have to revisit all those places I made this mistake. The same is true for many other of the many mistakes a beginner in LaTeX could do (e.g., using \\ all over the place).

And if the problem I asked is so hard that it takes a week until someone answers my question, I have done a weeks worth of work in the wrong way, only because someone felt it would be inappropriate to tell me my wrongs and rights before he fixed my hard problem.


Long story short: It is better for the questioner if those small mistakes are mentioned early, even if it might feel unkind. If a questioner is annoyed by honest help, then maybe he shouldn't have asked for help in the first place.

  • 1
    That's a good point: it is unlikely that a comment saying "don't use $$" is being done for any reason other than to help the user. What I wonder is whether it always comes across that way to the other person? And, let's be honest, if the user continues to use $$ for the rest of their document, probably nothing bad will happen...
    – rbrignall
    Feb 4, 2020 at 11:48
  • 1
    @rbrignall define bad. He might get inconsistent spacing, which I'd consider bad.
    – Skillmon
    Feb 4, 2020 at 12:17
  • 1
    Spacing does not seem to me to be something that the vast majority of LaTeX users worry or care about. Maybe you feel they should (after all, it is what LaTeX does best so you have a good point), but the reality is that they don't, or they leave it to an editor to fix up. In fact, this is exactly what I mean by "empathy": the user possibly doesn't care at all about such matters, and we need to be able to (a) understand what their priority is, and (b) think carefully about how we can help them to care about such matters in future.
    – rbrignall
    Feb 5, 2020 at 11:52

This is a much more complicated issue than it may appear, and by no means limited to SE or LaTeX.

I have seen many instances where someone has written "RTFM" and TFM actually does not contain the requested information. Even where it does, a helpful answer will indicate where in the manual the user should look.

Worse, the documentation is decentralized. If you point the user to the documentation, point him to all of the relevant files.

If you know of inconsistencies, warn the user of them.

Sometime a smaller document doesn't fail. That can make an MWE awkward.

Not only is good style subjective, but much of it is oral culture. I'm not aware of any single concise manual of good LaTeX style.

Try to explain why you are asking for additional information, or why a specific construct is bad form. Be a teacher, not a disciplinarian.

If you provide examples, explain them.

Sometimes the user knows things that you don't. Even when he doesn't, it won't hurt to give him the benefit of the doubt. Maybe there is a legitimate reason for something that strikes me as bad form.

If you give a man a fish, you fed him for a day. If you teach him how to fish, you feed him for life. If someone asks how do something, don't do it for him, explain ho to do it.

  • Very well put. I know that if I have some weird IT problem and I resort to google and technical forums, the chances are that I will find someone else with the same problem, and a load of replies to the posting that say absolutely nothing helpful. When you think of it that way, TeX.SE is a whole load better than the vast majority of such sites. The question is: can we make it even better?
    – rbrignall
    Feb 6, 2020 at 23:08
  • 2
    While I see your good intent, I'd like to draw your attention to the fact that "explaining" things is not a well-defined term. In more detail, in order to explain something you need to know what the other person does not know. I cannot explain this as well as Feynman did. You can only give a good answer to a well-posed question.
    – user194703
    Feb 7, 2020 at 22:47

As a person who asks questions on this site, part of the problem re trying to find answered questions is knowing what search text to use. For example, and obviously I'm not posting it here, I have a problem today.

It is a quite specific problem, and it's hard to know what search terms to use. So I search for every single phrase/set of words I can think of that might locate a solution. The upshot is spending over an hour reading irrelevant questions because they just might contain a solution to my problem. And this is what is happening every time I have a problem. I am unlikely to be the only one experiencing this.

There could be a great question with one or more great answers (giving viable alternatives to solving the problem). But unless the questioner has used the same words as exist in previously answered questions, the correct solution may not appear.

IMO and YMMV, LaTeX has so many packages, and some of them are quite complicated with hundreds of pages of manual, with an index that may not be helpful with respect to the problem. Using Rumsfeld, the problem is unknown unknowns: I don't know how to do this and I don't know what terms to use to find out how to do this.

  • 2
    Indeed, it is often much easier to find a similar question/answer if you know the answer to the question already. The TeX jargon that is a mix of terms from programming, computer science and typography as well as bits and pieces from the fields that use TeX for typesetting does not necessarily help here. But this site actually as a useful mechanic to deal with this: Duplicate closure. If you can't find the answer after some reasonable research efforts, ask a new question. ...
    – moewe
    Feb 9, 2020 at 12:32
  • 4
    ... If other people who know the answer or more about the jargon can find an existing question that resolves your issue, your question can be closed as a duplicate and will serve as a sign post for those who search for the keywords you mention in your question. If no one comes up with a duplicate you will hopefully get an answer. Of course this system requires that people understand that duplicate closure is not a bad thing. Those who suggest duplicates shouldn't do that in a way to suggest those asking questions did not do any research or are too stupid to find the right keywords. ...
    – moewe
    Feb 9, 2020 at 12:34
  • 2
    ... Those asking questions should not see duplicate closure as a negative thing and should not take it personally. In an ideal world those suggesting a duplicate would do that in a nice non-confrontational tone and should wait for a reply before closing a question directly. The asker would then reply in a short comment either saying that the duplicate helper and that the question can indeed be closed or that the duplicate didn't help and explaining why (often it is easier in this case to edit the question to explicitly mention the duplicate and explain why it didn't help).
    – moewe
    Feb 9, 2020 at 12:38
  • 2
    Finding a possible existing question based on keywords is not easy. The list of "related" questions in the right-hand panel is sometimes wildly irrelevant. There is a list of "Often referenced questions" here on meta; this might be helpful, but is likely not known by newbies. Suggestion: tex.sx veterans familiar with the subject matter of a question might edit in terms that are more likely to identify the topic for other readers; must go in the question -- comments aren't searched. Feb 9, 2020 at 14:32
  • 2
    @moewe I like the idea that closing down duplicates is simply viewed as a tidying exercise once the links have been established, rather than a reflection on the person asking the question in the first place. This is often how it ought to be viewed, I'd say, and comments should reflect this viewpoint.
    – rbrignall
    Feb 10, 2020 at 9:52
  • @barbarabeeton I did not realise the comments were excluded from the search. That's a PITA as the comments often contain important key words. Does that mean a Google search limited to the tex board is the better way to go compared to searching inside the board?
    – Michelle
    Feb 10, 2020 at 17:32
  • @Michelle -- Yes, it's definitely a PITA! I don't know whether Google is able to search the comments. I'll ask in the chat, and report if I get a definitive answer. Feb 10, 2020 at 17:38
  • @Michelle -- Just got a reply from Joseph Wright: "Search in general excludes comments by-design." Definitely discouraging. Feb 10, 2020 at 18:17
  • 2
    @barbarabeeton Joseph meant the site search, google will search the whole page, it doesn't know the difference between a comment and an answer Feb 10, 2020 at 19:12
  • @DavidCarlisle -- Thanks for the correction. Even if a general web search does search comments, there's still a disconnect trying to locate useful information. (That's why, when I find what I think is a good Q/A, I add it to the "often referenced questions" list -- so I can find it again when it's needed. Feb 10, 2020 at 19:16

The subject is "Empathy for those who don't know LaTeX well". You find a list of five key points that answerers of questions at TeX LaTeX Stack Exchange are asked to consider.

I like this list for various reasons. But I wonder whether these points really have to be mainly about empathy, or whether they could also be about behavior that, empathy or not, makes dealing with each other more pleasant and that makes dealing with problems more efficient.

Requesting an MWE

The phrase "Minimal Working Example" is misleading for many newcomers. Many newcomers think that it refers to an example that runs perfectly and without error messages during compilation. And that is exactly what they cannot deliver.
When I ask for code examples, I usually leave out the word "working" and explain that the point is that the code contains everything (and nothing else) that the person who wants to get to the bottom of the matter needs to reproduce the error-messages/the erroneous behavior during/by compilation and to be able to track down and fix things afterwards. Of course I also mention other "requisites" that make tracking things down easier—e.g., the messages on the console and in the .log-file using \listfiles (for package versions etc) and probably \errorcontextlines=....

When do I ask for code examples to reproduce erroneous behavior? Usually when someone wants to know why error messages occur and what the correct remedy is.

If a newcommer/beginner - due to a lack of familiarity with the "TeX/LaTeX-world" - does not know how to approach a problem related to accomplishing a specific typesetting-task, or related to organizing whatsoever data using LaTeX, or related to parsing some user-input for conditional-branching depending on that user-input, I do not ask for minimal examples but I explain how I would do it and provide examples of implementation on my part, and if I have the impression that it is necessary for understanding, I also explain what the individual code sections are for and how they intertwine when you put the whole thing together.

By providing code examples myself, I show that the problem can be solved using LaTeX. I show that things can be done in LaTeX. I do this on the assumption that people will be positively motivated to read in and to delve into the matter and to try things themselves when they see that you don't always just get stuck, but that things can actually be done.
Perhaps this is "answering do-it-for-me-questions" and therefore is not in the spirit of this site. But it is in the spirit of showing people that TeX is not terribly complicated and unusable, but that only the beginning of the learning curve may be terrible and that - after mastering the learning-curve - you can do things.
I hope that questioners will feel that they are not alone and abandoned when dealing with TeX/LaTeX. Many people shy away from Linux or LaTeX with the argument that they would be left alone with it as no one in their circle of acquaintances uses these things.

Directing someone to the manual/Impenetrability of manuals/instructions

When I read the TeXbook many years ago, I realized at some point that my problems were not necessarily caused by the "manual", but by the way I (ap)perceived and comprehended information while reading it.
Schoolbooks and introductory books are nice. There are a lot of nice filler words and things you can read over and still get the context/the gist. The TeXbook is written in the spirit of academic textbooks that convey information in a very condensed way. Once I got used to the fact that, as with legal contracts and rules, every sentence/syllable really does matter, and that slackening in attention easily leads to missing connections/implications that arise from passages/sentences placed far apart from each other, I no longer found the book impenetrable. But it requires one to be very concentrated and focused and picky. When I first read the book seriously, I very nitpickily made notes whenever I had the impression that terminology was used or a concept was taken up that would be explained later, and when I read the later chapters and the (double) dangerous bend paragraphs, I always checked my notes to see if something was being discussed that could refer to these notes.
When I refer to a manual, I usually assume that there were problems in understanding the manual and I quote the relevant sections and explain what is meant. For one, this shows how one puts the picture together from the individual terms. For another: The need for explanation often is due to the fact that instructions are written in English while English is not the native language of the reader/questioner and may not be the native language of the author of the manual, either.

I would like to take up the keyword "empathy" again

Empathy also has something to do with social competence. I am painfully aware that my own social competence is far from perfect. But as a rule, I try to give an answer that helps to solve the problem. And I expect questioners, even if they are not satisfied with the amount of empathy revealed by me, not to lose sight of what they are trying to achieve, which in the first place is not to find empathy but is to find a solution to the problem they present.

I used the term "reveal" because I often feel empathy and imagine how frustrated a questioner might be. Due to my own deficits in social competence I often ask myself what degree of showing the empathy I feel is appropriate.
The wrong ways of expressing empathy can give the other person the feeling of being helpless and unable to act themselves in the matter. And that in turn can be demotivating if not devastating.
I know from my own experience that it is frustrating to get stuck when working with TeX/LaTeX. But TeX/LaTeX is only a small aspect of life. So questioners, too, shouldn't let frustration dominate, but take things with humor.

As Donald E. Knuth writes in Chapter 6: Running TeX of the TeXbook:

Error messages can be terrifying when you aren’t prepared for them; but they can be fun when you have the right attitude.

Let's emphasize one of unknowns that you will probably always have to deal with in the question-answer scheme of the StackExchange-platform:

In most cases, the questioner and the respondent know nothing/know only little about each other's situation/circumstances. This in turn introduces a certain imponderability in assessing expectations.

This can lead, for example, to detailed, long answers being perceived as incompatible. For example, an answer written with a lot of commitment, peppered with academic reflections and digressions and references to many other sources, may show that the responder assumes that the questioner has the time and leisure to delve into the matter and to devote her-/himself to the fascination of the prospectus opening up. But the questioner is stressed at the moment and has no time.

Or the opposite: Someone who is delving into the matter and is fascinated and enthusiastic, and not only pragmatically wants to get rid of an application problem but also wants to gain deeper/fundamental insights into the matter, is disappointed when answers are quite pragmatically focused on slamming down the solution for the application case.

  • Not sure why this was downvoted, seems like a perfectly reasonable contribution to this Meta thread. On the matter of MWE (minimal working example), it seems the term in use on Stack Overflow is now updated to MRE — minimal reproducible example. There are also "magic links" that can be used in comments on SO to link to the corresponding help page, by typing any of [mcve], [reprex], [repro], [mre] and [example]. Nov 6, 2021 at 5:57
  • It seems that we do have an FAQ on this topic here on Meta, but maybe having a short and snappy help page to link to would also benefit new users? Nov 6, 2021 at 5:57
  • 4
    @TheAmplitwist Not sure why this was downvoted: Maybe someone disagreed with something I wrote. I understand that the possibility of downvoting should not be used in an inflationary way. But it is not a disaster at all if it is used: Downvoting is also an element of freedom of expression. I think it is important that downvoting is not seen as a personal attack with the aim of threatening one's existence. In this sense, however, it would be good if a downvote were always accompanied by a comment explaining what one considers worthy of criticism and why. Nov 6, 2021 at 12:58

You must log in to answer this question.

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