Let's get the facts straight

Per SE policy, as outlined here the following applies

On our main sites "the best answer" is somewhat fluidly defined as the superposition of the highest voted answer and the accepted answer. [...] Practically speaking, accepted answers to meta-questions are no different than any other answer. - Jon Ericson ♦

thus giving a definition of what a consensus is and how it is reached, namely the highest voted answer regardless of accepted answers things, with the only caveat being

Size of the site determines how many people could be voting on questions. - Jon Ericson ♦

Which is reiterated by the answer

There's no hard number or ratio that will guarantee that you're doing the right thing; let folks discuss, when they've had a chance, make a decision and implement it. - Shog9 ♦

When applying that to this discussion that means this answer is defined as the consensus, if it garnered enough votes. At 29 votes at the moment this does seem to be the case. (and even the second highest voted answer is higher voted than the accepted answer) PS. If you want to read more about the fact that accepted answers hold no special meaning on meta sites read here.

The 'problem'

Now, finally to the 'problem' at hand. On this post Yori corrected the answer as per meta consensus. After an invalid roll back by Paul Gessler I didn't want to get in a rollback war, so I added a comment saying

@PaulGessler: I am just an outsider, but the consensus in the linked meta discussion is that changing grammatical errors (including capitalisation) is acceptable as per the highest voted answer: meta.tex.stackexchange.com/a/2103/48077 The fact that egreg picked a different accepted answer means nothing per SE functioning. Your duty to the community and future readers is greater than to Barbara.

after which I received the follow up comment

@DavidMulder We don't follow such things here for the sake of SE network. More in this post, since this pops up every now and then meta.tex.stackexchange.com/questions/3802/… Also there is no duty involved on TeX-SX, it is all community driven voluntary work. – percusse 1 hour ago

Now, the problem is two fold:

  1. SE policies and even internal consensuses seem to not be clearly communicated to users here on TeX.SE (two technically incorrect comments both received upvotes) (This btw is a larger problem with a couple of other communities as well, however this is one of the oddest cases I have found)
  2. The fact that moderators haven't intercepted here to enforce the consensus on this and other answers by Barbara despite this seeming to happen on other answers written by her as well (invalid roll backs linking to the consensus, but looking at the accepted answer rather than the highest voted answer).

And just for the record, I am perfectly fine if you guys want to change StackExchange policy or mass vote now on the relevant meta question and redefine this specific consesus. Like seriously, if you feel like it, go for it. However, at least at the moment, this does not seem relevant.

PS. Yes, I am just an outsider (worked with TeX enough that I will be happy each day I don't need to use it O:) ), but I do love the functioning of the SE network as a whole and this answer came up in the hot network questions, and something like this just reflects extremely poorly on this sub-site.

  • 2
    See also Do we follow StackExchange rules? Commented Jul 13, 2014 at 12:20
  • @MartinSchröder: There are policies which are above all specific sites. For example the way a consensus should be reached (community driven) is SE specific, not *.SE specific (if somebody is going to hang around as a dictator you can bet they won't accept it). Then there are rules like which questions are on topic and which are not, those are absolutely *.SE specific and it's fine that TeX.SE is deviating in that regard. In this case however the TeX.SE consensus (barbara's posts may be fixed) aligns with SE policy, however is misenforced due to users looking at the accepted answer by mistake. Commented Jul 13, 2014 at 13:07
  • 1
    And whoever voted to close this as a duplicate, this post is all about the answer given in the linked post not being enforced! I know what the answer is, I am linking to it myself in this post. Commented Jul 13, 2014 at 13:08
  • 12
    I think that to quote the current vote on a meta question as determining current "consensus" on the main site is really quite a bizarre position to take. Attendance in meta is low so most of the users, even regular ones, wouldn't be be aware of the vote. If it were really to be taken seriously then we could easily as you note yourself, highlight that answer on the main tex.sx site and encourage people to vote to re-establish the voting figures to match what is clearly the working consensus of the active users on the site, that in the main, the authoring style of the OP should be left intact. Commented Jul 13, 2014 at 15:41
  • 3
    What makes you think that the post you mention is considered "SE policy"?
    – Werner Mod
    Commented Jul 13, 2014 at 15:50
  • 4
    I think @DavidCarlisle makes a good point. Moreover, given that there was already a bit of a "debacle" about this, I wouldn't be surprised if a good number of the votes in Respecting poster's style come from those who are not (or no longer) part of the TeX.SX community. (And we don't downvote that often here.) Meaning that, even were this a good method for determining consensus, it's not clear that it would be a good method in this case.
    – Adam Liter
    Commented Jul 13, 2014 at 16:50
  • 8
    Since at the current time your question has 0 votes and 4 answers with 11 votes arguing against the position that you take, do you accept the consensus position that there is no problem that needs fixing here? Commented Jul 13, 2014 at 18:07
  • 5
    Since you have defaced your question twice I'm not going to fix it again, I have flagged it for moderator attention. Commented Jul 13, 2014 at 20:56
  • 7
    @DavidMulder: Now it is going to get really ridiculous, behaving like a small child not getting his toy
    – user31729
    Commented Jul 13, 2014 at 21:04
  • 1
    @DavidMulder I've written an answer that I think more accurately characterizes the "consensus" on TeX.SX. We do respect posters' styles, but not without qualification. The qualification being whether the style makes the post less comprehensible or readable. The, quite honestly, immature edit that you've made is an instance of a style making a post less comprehensible and less readable. Lastly, before I leave this discussion and go do something productive, I hope you have a good day! :) This seems to have turned unnecessarily toxic ... sigh.
    – Adam Liter
    Commented Jul 14, 2014 at 0:24
  • 3
    All: Please bear in mind that the core question here ('How is a decision reached on 'policy' questions on the meta site?') is perfectly reasonable. The fact that it may be a somewhat flexible process and can be argued as dependent on the policy item under discussion makes giving good answers potentially tricky, but that does not detract from the question.
    – Joseph Wright Mod
    Commented Jul 14, 2014 at 10:39
  • 3
    @JosephWright or rather that would have been a reasonable question had it been asked. One of the main problems with this question is that (even before the silly edit) it was not phrased as a question but as an assertion. Commented Jul 14, 2014 at 11:24
  • 4
    @DavidMulder as someone who professes a desire to follow the stackexchange model, you should know that questions should include a question It could have been closed as "unclear what you are asking" (as it does not ask anything) but because we are lenient with new users it was let go and people answered whatever question they thought might be implied by your statements (the tone of which is to be honest most rare and unwelcome on this site) Until Martin used the same rollback that I had used this was almost the lowest voted question on the site (I haven't voted) fame is not the issue here. Commented Jul 15, 2014 at 12:30
  • 1
    I've voted to leave this question open again. I agree with @JosephWright that there is actually an interesting question here (or, at least there is, had that question actually been asked). Minimally, I don't think it should be closed as a duplicate of "Respecting poster's style". Perhaps David's suggestion of closing it as "unclear what you're asking" might be best if the community does really want to close it. If this is what the community wants, it could then be reopened if the OP actually formulates a question about how we reach "consensus", rather than telling us that we have a problem.
    – Adam Liter
    Commented Jul 19, 2014 at 5:11
  • 3
    I, personally, had no idea that a vote here was a vote on policy as opposed to a way of recognising a good answer. If I had not started from this question, I probably would have (+1)ed at least 3 answers in the linked thread, 2 of which are being interpreted here as endorsing entirely incompatible policies. Now maybe that's just my ignorance but I wonder if The One Way, Truth & Light might be rather less justified than it is claimed. A vote is only an expression of the will of the people if the people know what they are voting on. (Not a sufficient condition but surely a necessary one.)
    – cfr
    Commented Jul 21, 2014 at 3:06

5 Answers 5


The first thing to bear in mind here is that while there are places where there are StackExchange rules/polices, on editorial matters 'guidelines' is a much more accurate term. (Rules are clearly appropriate for legal matters, for example.) Each StackExchange network site has its own conventions, and those are to some extent 'organic'. They also need to work with the fact that each question and answer involves a 'real' person: people are funny things, and rules don't always work so well with them! So there's got to be some flexibility. (I'll come back to that below.)

As noted in comments, voting on the meta site involves a subset of people 'routinely active' on the site, and that in itself is a subset of people who visit the site. As such, the best one can say about votes on meta is that they give some indication of the feeling of some of the site users. At the same time, there's no way of seeing what the history of voting is. Asking a 'how do we deal with ...' question requires that at some stage there is a decision: presumably there is some form of implementation to be done. However, votes can continue after this, so what gets implemented might end up not being what has the most votes. One way of indicating which answer seems to have 'won' at some arbitrary point is to accept it, but as we can't see the history we can't actually tell if that was the case here (at least, I can't be certain).

While the network emphasis the wiki/collaborative element to questions and answers, each one has someone's name on it. That's something we should, I think, be mindful about. So it's one thing to have a convention to 'fix' things, but another to ignore the OP when the roll such edits back. For example, there is a weak convention on the site to remove 'thanks' from questions, but if the OP rolls back such an edit then I'd usually anticipate that it would be 'left alone'.

I think it's important to note here that as noted in comments on the question, there are a reasonable number of votes in both directions on the 'What should the guideline be?' question, and also quite a number in favour of a 'light touch' approach laid out in response to a previous question on this area. As such, and in the context of 'How are decisions reached?', it seems to me that this is not an area where there is a clear consensus one way or the other. As such, we as a 'community' and in particular the mod team are left having to find a workable balance. To me, that 'balance' seems to be to edit where it will be seen by the OP as helpful whilst avoiding editing 'wars' as much as possible. (That of course does mean having to accept some things that as an individual I may not be 100% happy about.)

In the end, my feeling is that we do better having a wider range of contributors, even with the odd eccentricity, than rigidly enforcing a set of 'rules'. To date, we've not had any serious confusion from people looking for help, at least that I know of.

  • That's what happens on a dynamic site as SE. But the nice thing is, old posts can be left alone, and if someone feels like fixing them they may, if they don't they don't. As I said before, fine with me if everyone now upvotes the 18 vote posts, but the thing is, even after raising the issue here that has not happened. Meta is here for a reason, not just as a random forum for users without a life ;-) Commented Jul 13, 2014 at 17:30
  • 2
    @DavidMulder what if the vote count keeps changing? Should we edit things now, but then, say, in a few weeks, go roll everything back if the votes change again? And then re-edit things again if the leading answer changes again?
    – Adam Liter
    Commented Jul 13, 2014 at 17:44
  • @DavidMulder I can certainly see that (as a mod I really need the guidance from meta to know how to act in the 'spirit of the community'). The problem is, as noted by others, that meta is a rather small self-selecting place, and so it's what happens in practice on the main site that is most likely the guide to what the 'community' wants. ...
    – Joseph Wright Mod
    Commented Jul 15, 2014 at 7:52
  • ... I see that this can be something of a chicken-and-egg problem (some people take a line on the main site, that's then referenced as the standard behaviour so others follow it, ...), but at the same time a simple 'most votes wins' situation for meta posts doesn't really work that well either. The mod team are very keen that everyone can contribute to the site: we are trying to find a balance between competing desires (e.g. individual's 'ownership' of their posts versus wider need for this material to be useful/standardised).
    – Joseph Wright Mod
    Commented Jul 15, 2014 at 7:52

I see my name there with this;

SE policies and even internal consensuses seem to not be clearly communicated to users here on TeX.SE (two technically incorrect comments both received upvotes) (This btw is a larger problem with a couple of other communities as well, however this is one of the oddest cases I have found)

Well this is very simple from my point of view, any TeX guru or TeX contributor who made this typesetting system possible for me to use TeX and to write stupid answers on this site, gets the priority. If barbara beeton wants to retain her style which has been established way before TeX-SE and known to the TeX world then I say aye aye. Because instead of only hanging around here, they made the real contribution to countless people around the world, and I won't accept forcing a user comply with a rule set that is not even self-consistent.

If people here are OK with it, I'm OK with it. When you are signing up to the site you don't see a rock in the mount StackOlympus they said, laws you should abide written on it. It's a TeX site and we have our nice environment. So why should we introduce some questionable policy that is not even relevant to our case. There is no mention of SO whatsoever. I'm not signing up for the whole network, I just care about TeX.

You might wonder why we are defending our inconsistent behavior, well because we don't care about consistency for the sake of OCD. Here are two more that would probably make you even more disappointed about us;

Our Do-it-for-me and Draw-it-for-me comments don't reflect our hypocrisy. Can they be improved?

What to do with opinion based questions which are accepted by the community?

And here is a teaser for you;

So far I have seen at least seven users (that I remember) coming here and trying to impose those things that they are used to from SO. They always start from removing thanks and grammar.

If you really say that it is the content that should be protected and crystallized why on earth any single of you don't start from MWE requests or answering questions but rather pick up on the most ridiculous detail to start contributing?

And finally, get on the chatrooms of recently graduated successful sites and ask them where they got their welcoming behavior or downvoting schemes or block comments, you'll have a different picture.


As commented in the posting that you quote initially, any assessment of whether consensus has been reached has to take account of the possibility of low participation. That clearly applies in the case of the meta site here which has far lower participation than the main site. So a simple appeal to the current voting figures on a question doesn't really have much effect: common practice, and common sense, need to be used to moderate the established norms, not just transient numerical supremacy.

However it is not really controversial that questions may be edited to fix grammar and other problems, but that does not mean that all edits are good.

A newcomer to this site might for example be surprised at the strange capitalisation of TeX and "fix" it. Any such edit is likely to be rolled back. The strange capitalisation of Barbara's posts is a convention as old as TeX and explicitly documented in her profile on the site. As a newcomer you may not be aware of it, but on having it pointed out and having the edit rolled back your response should have been to say "oh sorry" and move on, not to post calling in to question the behaviour of the members of this site and the competence of its moderators.

  • 2
    Regardless of how old the convention is, 28 votes on a meta post for a not too big site is quite a lot. You as an old user have no right to ignore the majority vote on such a question, even if that majority vote is down by both new comers and old users. I have compared it to other meta posts and that's seriously quite a lot. Tradition does not overrule democracy on SE. Commented Jul 13, 2014 at 17:25
  • 7
    @DavidMulder You are ignoring the fact that your edit was simply a bad edit, irrespective of any general conventions about fixing grammar. It changes the meaning of the answer (given the statement on the user profile). Perhaps you think that's a bad convention for a user to have but making all following readers believe that a statement is an official policy statement of the American Mathematical Society when it was explicitly marked as not being so in its original form is a bad edit and should be reverted. Commented Jul 13, 2014 at 17:51
  • 2
    @DavidMulder "comparing to other meta posts" rather misses the point: you should compare it to posts on the main site to see if numbers on meta shoudl be taken as meaning anything on the main site. Commented Jul 13, 2014 at 17:56

Some very good points have already been made, but I would like to add one thing. As noted in other answers, not a lot of the site membership visits Meta very often. So I propose to look at the behavior of users in the main site itself, as a rudimentary gauge of the prevailing view on the site.

I did these tabulations by hand, and there is certainly a possibility that I have made an error. Please, feel free to check my work if you so desire. :-)

At the time of writing, user barbara beeton had 517 answers and 2 questions posted on our main site (I did not consider any Meta activity in these figures).

Both questions (both with a modest number of upvotes) have no beginning-of-sentence capitalization, with no edits by other community members.

Of the 517 answers, 493 have no beginning-of-sentence capitalization. Many of these answers have been around for a long time and have received large numbers of upvotes. Just 24 of her answers have some beginning-of-sentence capitalization. Thus, the community as a whole has accepted her convention in over 95 % of cases. Furthermore, this figure slightly under-represents the actual acceptance of her convention, since the answers with beginning-of-sentence capitalization include community wiki posts with multiple contributors and "official" statements from her employer(s) where she capitalized on her own, in addition to posts edited to correct her capitalization.

So it appears that in a vast majority of cases, users of the site support (or at the very least accept) her convention. In addition, I personally view the convention as helpful. This is because I can distinguish a canonical recommendation from a personal one immediately.

  • 6
    I would also like to point out that the original editor of barbara's post (whose edit I rolled back) thanked me in a comment (which has since been deleted) for informing him of barbara's convention. Commented Jul 13, 2014 at 18:19
  • 1
    I, at least, am far more gung ho about editing questions than I am answers. Barbara's unreadable "style" might be accepted in answers, but it really shouldn't be in questions. The fact that it is accepted shows that you're not judging posts on their individual merits, as you should. The post is getting undeserved respect because of the history of the poster. That's a problem.
    – TRiG
    Commented Oct 5, 2014 at 14:33

This is more of an extended comment than an answer, but I think it is quite important to clear something up, here.

We are not talking about consensus.

The Stack Exchange network is not set up to facilitate consensus-based decision making processes. Speaking from both personal experience and from having read theories of consensus, consensus is often hard to achieve. It takes long face-to-face conversations and a lot of work and energy. This is something that the Stack Exchange model neither provides nor facilitates.

Unless a network is still in beta and has fewer than 50 (20?) users, I find it particularly unlikely that consensus will ever be reached about anything that is even slightly controversial.

So this means that we're talking about (at best) majority-based decision making processes,1 and this inherently entails that some users will be unhappy.2

Now, what's the point of this discussion?

Well, the Stack Exchange model has been setup so as to facilitate the creation of communities, and these communities are, within limits, free to determine some of their own policies, practices, and oddities. Given that some of these policies will perhaps be controversial and given that the Stack Exchange model does not allow for consensus-based decision making processes, we have to accept that some users will be unhappy. So while this does not address the question about how "consensus" is actually determined (and I think David has already raised some good points about this), it is important to recognize that this is not "consensus" and just because some individuals (often individuals who do not participate regularly in the TeX.SX community, if at all) get upset, does not mean that we should immediately stop our practice and adopt theirs because the other way of doing things will not necessarily be any better. That is, it will just make other users upset.

  1. It's not clear that this even constitutes majority-based decision making. See @cfr's comment.

  2. This is why comments like this, I think, are ridiculous. Why doesn't SO stop running so roughshod with new users? That obviously repels a lot of users. So SO should obviously stop doing that, right?! Of course not. There is good reason that they do things the way that they do.

  • 1
    I am really sorry, but consensus on SE is used to refer to majority decisions reached on meta sites. Commented Jul 13, 2014 at 17:26
  • 1
    @DavidMulder you/they can call it what you want. Anyone is always free to stipulate their own technical terms. The only point I am making is that this is not a process wherein everyone actually agrees with the decision that gets made, which is what consensus usually means. If you/they want to keep calling it consensus, that's fine. I find that confusing because the usual meaning of the actual term obscures the fact it is really just majority-based decision making, which I think is important to recognize precisely because we have to expect some users to not be happy with the decision.
    – Adam Liter
    Commented Jul 13, 2014 at 17:31
  • 3
    (+1) for not consensus. (Maybe consensus\textsubscript{SO} but that is different.) However, I don't think it counts as majority voting either unless everyone is a lot clearer than I was about what a vote on an answer was 'supposed' to mean on meta. As I've been proceeding, there is absolutely no reason not to vote for two incompatible answers if they are both good answers. (Or, indeed, for more pairwise inconsistent sets of three or more answers.)
    – cfr
    Commented Jul 22, 2014 at 1:53
  • 1
    That's a good point, too, @cfr. And it's also not even the case that that many people participate in meta to begin with. Anyway, I agree with your point, and I've edited my answer to reflect that.
    – Adam Liter
    Commented Jul 22, 2014 at 3:36

You must log in to answer this question.

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