The question, Why not program our video text terminals/terminal emulators to use something JSON or XML on the backend instead of ANSI escape sequences? was closed yesterday by the admin, Robert Harvey. The close reason is "primarily opinion-based". The description of that close reason states, "Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise".

There are two answers to this question, neither of which has received any down-votes and neither of which in any way fits "[is] almost entirely based on opinions, rather than facts, references, or specific expertise". When I challenged Robert on this, he stated "I didn't evaluate the answers". To my mind, this is poor behaviour by a normal user, but for a mod to declare this, is inexcusable.

There are a small group of users that regularly abuse this close reason. I recently discovered that I can flag such abused questions, and thus in theory have a mod look at the question so they can use their good judgement and re-open them. But quite frankly if mods are going to also openly abuse this close reason, there'd be no point in me doing this.

A while ago, I asked the question Should the “primarily opinion-based” close-reason ever be used on Programmers, except on extreme occasions?. It has one non-controversial answer with no down-votes. So it looks like those that care about such things are all in agreement on when that vote should be used.

So why does it continue to be abused? And if mods are abusing it, is the battle to prevent its misuse lost already?

  • 4
    To be fair, the borderline between "using" and "abusing" features of this site is definitely to some degree opinionated.
    – Doc Brown
    Jun 14, 2019 at 8:23

3 Answers 3


My opinion about the "primarily opinion based" close-reason is probably a little bit different than yours about it, and from the fact your question from 2015 did get only one answer which did not directly oppose your question's statement, I would not conclude that you have broad consensus here on how this close reason should be used. In fact, my opinion about it is more in accordance to this meta answer, which says we need that close reason for questions

  • which are simply impossible to answer because there isn't a general answer

  • where suitable answers are all value judgments that will be different for each person, each company, each background.

(but are otherwise clear, on-topic and not too broad). This answer also suggest agressively edits to such questions to save them / get them reopened.

What does this mean for the question in stake?

Let's analyse the four main topics of the question one by one:

Is it purely due to legacy reasons why displaying text on our video terminals is done with ANSI escape sequences and not another framework such as JSON, Yaml, XML, or something else?

This is somewhat opinionated, but probably answerable. I gave a comment about ANSI being a standard, not writing a full answer, but at that time I already had my doubts if that question was answerable in full.

Is ANSI escape sequences in video terminal, simply an old technology similar to say, X11 that sticks around solely due to how embedded it is within the computing paradigm?

Same as above. Let me add, the wording of these sentences bothers me. It gives me the impression the OP thinks those "are only those bloody, crap, quite unimportant legacy reasons" hindering people to invent something better. To me, it looks the OP underestimates here the value of mature standards.

If not, why don't developers switch from an escape sequence style to something that would support nesting?

That is definitely something one would require a crystal ball to write an answer and falls under my personal definition of "too opinionated". We cannot read the minds of all developers all over the world, only speculate about it, but as you know, SE.SE is not a discussion site. The sentence contains also a pretty wrong allegation, that that this ANSI escape style is quite popular today, as we also discussed in the comments. Lots of developers use a kind of formatted text description which supports nesting, however the platforms for this are not called "terminal interpreters", they are called web browsers.

Are there any proposals to do away with ANSI escapes in terminals and replace it with something else?

That is something I don't know (and off-topic, since it is a request for 3rd party resources.)

Does looking at the two (good) answers change my mind on this? Not really - I can see the flaws very clearly with or without looking at the answers. Interestingly, none of the two answers adresses directly any of those four sentences above, they adress directly the wrong assumptions in the question, nothing else.

We actually don't know what would have happened when Robert would not have closed that question so quickly, maybe we would have gotten lots of more (opinionated) answers, interpreting the question differently. Maybe the community would have closed it, maybe not, who knows? But just because the explanation of the "primarily opinion based" close-reason mentions the tendency of attracting a certain kind of opinionated answers does not imply we (or the mods) always have to wait until this actually happens.

What about our expectations about how the mods should deal with such a question, which partly answerable?

I see four different ways of dealing with it, especially for a mod:

  • ask the OP kindly to edit its question and tell him/her what should be improved

  • edit the question agressively

  • close the question immediately.

  • wait what the community does (and maybe decide to close the question later).

All of them can be legitimate.

During the mod election phase, Robert Harvey clearly stated he would like to become a mod to get a binding vote to quickly close questions. Netherless he got most of the votes for his application - so it seems the community here prefers this moderation style where someone quickly closes even borderline questions.

This is actually not my preference, I would often prefer a different handling (as I have stated here). So closing the question so quickly is probably not something I would have done, since I think it contains a lot of interesting parts and may be improveable by an agressive edit to the point where it can stay open. However, I don't see Robert's action as a clear abuse of the system (as you called it more than once), it was a IMHO legitimate, understandable action - maybe not the best possible one.

To the point of not reading the answers here first: yes, I agree to you, a mod should have had a look at the answers first before closing a question, especially for questions like this one. But if they decide not to do it, it is IMHO mediocre, maybe weak moderation style, but still not an abuse of the system. And I doubt that for this case a look at the answers would have changed Robert's point of view (but let us wait what he has to say about it himself).

So is all what happened here ok?

No, there is one thing which I miss here: if mods get asked about the reasons for their actions, especially here at meta, I expect them in person to write an answer and give an explanation.

Advice #3 from stackexchange's theory of moderation says:

Whenever possible, try to leave frequent comments on posts where you’ve taken (or considered taking) a moderator action, explaining the reasoning. This is important so that community members can learn the norms of the community and the moderation policies.

So @RobertHarvey, you have been around here in the last 48 hours, but still no answer from your side? Same for this one where the OP asked for clarification about the closing, or suggestions for improvement, and you wrote not just one word about your motivations? Or this one, where you left it over to GlenH7 to give an explanation, but not even a comment if that is also your point of view? Also this one, where you left it over to gnat?

I would like to encourage you to rethink your style of (non-)communication over your mod actions. More detailed replies could actually prevent such emotional reactions as the one we saw here from David Arno which lead to such excessive accusations of "abusing the system".

  • "That is definitely something one would require a crystal ball to write an answer" I don't know; the fact that such a project would require cooperation from dozens of disparate console terminal projects would seem to be an entirely objective reason why it hasn't happened. It may not be "the reason", but it is certainly "a reason". Jun 14, 2019 at 17:06
  • @NicolBolas: your comment shows me that it is quite unclear to me what the OP actually meant or expected here: a paradigm shift for "the majority of terminal emulation projects all over the world", or just one platform for his debug console. Such an unclear wording has IMHO a high potential to attract opinionated answers - but leaves also some room for making the question salvageable by an edit.
    – Doc Brown
    Jun 14, 2019 at 22:08
  • I don't see how you could interpret the text of the question as being limited to their own platform. While the "background" of the question talks about one specific platform, the actual text of the question speaks very generically about "ANSI escape sequences in video terminal". As for salvaging the question... why? If it's asking about many terminal programs, then it has an obvious answer. And if it's talking about just one specific platform, then it's too narrow to be useful to anyone else. I don't see where the benefit is in salvaging this question. Jun 14, 2019 at 23:18
  • 1
    @NicolBolas: is the answer so obvious? As far as I can see, the question got two very different answer which are not too obvious to me, and if it had been worded differently, with less wrong assumptions and prejudice, I see still a chance for survival (not only of that question, but also of those answers). However, I think this rewording should be done or at least supported by the OP. Note also, this meta question is not really about that question about console terminals, it is about Robert Harvey's moderation style.
    – Doc Brown
    Jun 15, 2019 at 6:35
  • OP of the question here: X11 by the admission of its own developers, can be chalked up to sticking around due to legacy reasons, namely Netscape, and not mature standards. See: youtube.com/watch?v=-oFxhqYn-g0 -- By the year 2003, it was about "fixing" it more or less because they had to, because replacing, which they floated, it was untenable. Fast forward 15 years, and we see the developments of Mir and Wayland, and by those developers justifying their development due to limitations of X, with ever present complaint of dealing with the X API.
    – Anon
    Jun 15, 2019 at 21:29
  • 1
    I will also upvote this answer, because I do think your other criticisms on how I asked the question in the body are valid. Crystal Ball metaphor is quite useful.
    – Anon
    Jun 15, 2019 at 22:58
  • 1
    @Akiva: you can still edit your question and try to get it reopened that way. However, I fear if you edit all those problematic parts out, what remains will be a request for a library or tool which supports the usage of a "higher level language" for terminals - and such a request is ok for softwarerecommendations.SE, but not for softwareengineerung.SE. Maybe the lib I linked to in my last comment to Karl Bielefeld's answer is what you are looking for.
    – Doc Brown
    Jun 15, 2019 at 23:19
  • 1
    The other mods said everything I wanted to say. I have nothing further to add. Jun 17, 2019 at 18:08
  • @RobertHarvey: honestly, you were the one who told David Arno to bring this to meta, and then you leave the field to others? Seriously? Moreover, did you count the number of meta questions where someone asked for a clarification about your mod actions since you got the mod hat, and how few of them got an answer from you?
    – Doc Brown
    Jun 17, 2019 at 19:52
  • 1
    It hasn't been that many. I asked David to bring his discussion to Meta because that's the place where it belongs. Jun 17, 2019 at 19:53
  • clearly my answer is Objective Truth by definition and thus the close reason is incorrect
    – Ewan
    Jul 25, 2019 at 12:31

Firstly, the question was not put on hold just by a diamond moderator. Two other users had already cast close votes before the moderator saw the question and agreed with them. This is not a unilateral action by a moderator, but one taken in conjunction with other users.

Secondly, having been put on hold other users can come along and vote to reopen the question - which they might do if the question has been edited to address the reasons for closure. As I write this the question has two reopen votes, so it might well be reopened soon.

Also, there is nothing wrong with not evaluating any existing answers before casting a close vote. You are looking at and assessing the question, not any answers and think that the question as written doesn't fall within the scope of the site. You should consider editing, but not all questions are salvageable though edits. However, if you think the question is borderline, then you really should look at the answers to see if they're OK. It could well be that others are seeing something you're not.

  • 1
    Downvoting this as I fundamentally disagree with "Also, there is nothing wrong with not evaluating any existing answers before casting a close vote". If the close reason specifically addresses answers, it is downright disrespectful to the OP and those who've supplied answers to not read the answers before voting.
    – David Arno
    Jun 14, 2019 at 9:32
  • 3
    @DavidArno If you're presented with the question in the review queue you can't see the answers - only that there are some. I can't speak for Robert, but I'd only check existing answers if I thought the question was borderline.
    – ChrisF Mod
    Jun 14, 2019 at 9:53
  • @ChrisF: from the review queue, the full question and its answers is only "one click away", so this is IMHO a weak argument.
    – Doc Brown
    Jun 14, 2019 at 12:05
  • 4
    @DocBrown The vote to close is against the question. A great answer does not make a question appropriate for a site. That sort of an exception is the prime example that gets cited later on by someone posting a VLQ question and declaring that their question should remain open despite being off-topic. We have seen that occur here in meta countless times.
    – user53019
    Jun 14, 2019 at 12:21
  • 1
    @GlenH7: you got me wrong, I agree almost fully to you. I am just saying the "review queue" argument is weak, recommending to find a better one.
    – Doc Brown
    Jun 14, 2019 at 12:23
  • @DocBrown - it depends on whether you think the question is definitely off topic or only possibly off topic. In the former case you don't need to see the answers because they won't change your mind. In the latter case you should look at the answers because they might change your mind.
    – ChrisF Mod
    Jun 14, 2019 at 12:26
  • 5
    @ChrisF: yes, but one should not use the way the review queue is implemented as an excuse for avoiding to look at the answers.
    – Doc Brown
    Jun 14, 2019 at 12:44
  • 3
    @DocBrown point taken
    – ChrisF Mod
    Jun 14, 2019 at 12:47
  • "A great answer does not make a question appropriate for a site". That, I think captures my issue perfectly. In my view, it absolutely does, especially if we are considering whether a question is purely opinion-based. Clearly though my thoughts on this run counter to many here, especially amongst the mods. So I think I just need to accept that this site isn't moderated in the way I'd like it to be and that that is my problem. Thanks for the feedback.
    – David Arno
    Jun 14, 2019 at 18:57
  • Down-voting, as your post promotes group think inertia over analyzing a post based upon its own merits. Take heed of, en.wikipedia.org/wiki/Asch_conformity_experiments -- For all you know; the downvotes are trolls or the emotive dispositions of depressed or ostracized users who are projecting their own misery, bitterness, or ostracism unto other users. To the point, a takeaway from those experiments, are that those who speak up in truth to affirm their own perspective, give others the courage to do so as well, and I propose that Administrators should be chosen based upon that trait.
    – Anon
    Jun 15, 2019 at 19:11
  • 1
    @DavidArno: "... this site isn't moderated in the way I'd like it to be and that that is my problem" - well, here I am with you. As you may have noticed, I edited the last paragraphs of my answer for a more detailed explanation.
    – Doc Brown
    Jun 16, 2019 at 16:58

The standard of closing questions on:

primarily opinion-based

is extremely appropriate for Stack Overflow, and extremely inappropriate for Software Engineering, because for Software Engineering as a whole, the answer tends to find itself within a discussion, especially across generations of programmers (Where certain limitations forced different styles and philosophy of code), or even the temperaments of Software Engineers, say if one is more visually oriented over others. Given this:

primarily opinion-based

becomes a truism here. For example: At what point is brevity no longer a virtue? -- this extremely popular question could not be more opinion based, especially because you are asking about virtue. So much so that the accepted answer:


offers his personal explicit opinion:

enter image description here

Conversely, Why not program our video text terminals/terminal emulators to use something JSON or XML on the backend instead of ANSI escape sequences? is asking for technical and mechanical reasons not to program a terminal backend with JSON, XML, etc, and yet, it was argued that this was primarily opinion based.

But because Primarily opinion based is a truism, it is not technically wrong to say this, but it is certainly inconsistent, and I would argue, at the whims of social inertia. Had At what point is brevity no longer a virtue? been immediately down-voted twice, it would have probably been ultimately closed as users, and admins conform to the Asch paradigm, where initial conformity is weighed higher than objectivity.

As such, I propose two solutions:

1: Get rid of

Primarily opinion based

closes, and replace it with something along the lines of

Arbitrary Favouritism / Personal Inquiry

As such, you could not ask:

Why do you put your left sock on first?

But you could ask,

Why would you put your left sock on first?

and to that,

How did you decide that you wanted to write back-end in language x and not in y,z or w?

closes, where as

At what point is brevity no longer a virtue?

remains open.

2: Choose administrators who are readily cognizant of the Asch paradigm, and who swear to not let other votes influence their decisions.

  • 2
    I can assure you, conformity is not one of my problems. That said, you could ask "why" about almost anything. "Why is the sky blue?" "Because we have an atmosphere." "Why do we have an atmosphere?" And so on. Social inertia is not the problem; bikeshedding is the problem. Nor is the problem "why do" vs. "why would;" that's just a technicality. Jun 15, 2019 at 23:19
  • @RobertHarvey That's just a technicality << but it is a useful technicality that would delineate clearer lines in the sand. I have proven in my answer, that "Primarily Opinion Based" is an inconsistent standard, given the "Brevity Virtue" question and answer popularity. If you wanted to do away with Primarily opinion based, that would be the place to start.
    – Anon
    Jun 27, 2019 at 10:55
  • 2
    If you wanted to do away with Primarily opinion based, that would be the place to start. -- I'm quite happy with refraining from using that close reason, if it will make people feel better. Jun 27, 2019 at 15:39

You must log in to answer this question.

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