36

Questions on the "Hot Network" list are like travel billboards: they advertise sites to users elsewhere, to maybe get those users interested in participating. But I have noticed that those questions don't seem to get edited any more than others; bad titles pop up in the list, and they stay. I studied this over the course of a few days, and I collected a fair number of samples.

This makes me sad. One of the great things about this platform is the ease with which experienced users can improve others' posts. These posts have interactions from those experienced users, but they're left looking a bit shabby.

I'd like to propose an automatic banner be placed on questions that are featured on the HNQ list. The banner would be visible to anyone with the editing privilege, and would say something to the tune of

This post is on the Hot Network Questions list! That means lots of people are seeing its title and deciding whether to come visit the page. Is the title grammatical and descriptive? Is there anything else that can be [edited] here? Help make your site look good for visitors!

It might also be worthwhile to give experienced answers a little pop-up after they post an answer

Thanks for answering! Please take a minute to look at the question and [edit] if necessary to make it as good as it can be.

Here's a few examples. I've noted the title's problems and taken a picture of the HNQ entry. I visited the questions as I saw them in the list and looked at the circumstances: how much attention the question had gotten and from whom, and how old it was. (Images link to the questions if you want to visit them in their current state.)

While I was collecting these, I entirely ignored sites with deliberately "weird" titles: Arqade and Puzzling, and to a lesser extent Code Golf. This is just about objectively bad writing: mostly grammatical or spelling problems, and also titles that don't provide much in the way of information.


Capitalization; grammar/descriptiveness

Five answers, two by top users. No edits (and the body also needs grammatical help).


Utterly undescriptive

Edited by a high-rep user, title untouched; well-voted answer.


Undescriptive

Up for 4 hours, 1 answer (accepted), edited by two moderators, no title edits


Grammar

Up for 6 hours, answered by half a dozen site regulars. Edited by one person to add a tag and remove "Thanks"



Then there's the really low-hanging fruit. This stuff is so easy to fix up; it's almost mechanical.

Capitalization

22 hours, several answers by regulars, one highly upvoted. No edits.


More are captured in this Gist.


Now, this isn't a huge problem by any means. While I was actively looking for these things, I never saw more than three awful titles in a list of approximately twenty. At the absolute maximum, 5% of Hot Network Questions have titles like this.

On the other hand, I think it could have an important impact on individual users' attitudes towards their sites, opening their viewpoints a little to consider what the place looks like to inbound visitors. And if more improvement does happen, it will only make the network as a whole look better.

  • 1
    Thanks for the post. I edited one HNQ thanks to your post. But as I suggested here, meta.stackexchange.com/questions/284993/…, I think Most Recommended Question should be introduced as HNQs sometimes don't represent top-quality quesitons and answers on each site. – Rathony Oct 24 '16 at 16:56
  • "First thing to keep in mind is that purpose of hot questions feature is established pretty firmly and that it's not really about helping in your community growth..." (How can I get people to join a site and not simply glance and pass it by?) – gnat Oct 24 '16 at 17:48
  • 5
    Related: Let mods and 10kers know when questions go "hot" – jscs Oct 24 '16 at 19:41
  • (@JoshCaswell, wow, you really made an effort there with the example list (extent)! +1) – Aaron Thoma Jan 2 '17 at 22:35
  • Heh, I may have gone a little overboard with the list @accolade... – jscs Jan 3 '17 at 2:29
  • 1
    I think the question would benefit from dropping any concern over capitalization and caring less about grammar. Useless and nondescriptive titles on HNQ are a problem. If nothing else, could you reorganize so that the truly problematic ones (you can't tell what they are without clicking to the question) come first and the "shrug" ones (eg your current second and third from workplace and security respectively) come way further down the list? In this current order I started to disagree with you due to trivial examples, but hung in there and ended up agreeing. – Kate Gregory Jan 14 '17 at 17:05
  • 4
    To my mind, the grammar/spelling ones are, if anything, more important, since they're the lowest of the fruit. An editor fixes up the body of the question, but can't be arsed to capitalize the first letter of the title and add an apostrophe? Your clubhouse may have velvet seatcushions and the finest cocktails in town once you get inside, but from the street it looks like a burnt-out auto body shop. However, I take your point about organization: the list is long and could probably stand some rearrangement. – jscs Jan 14 '17 at 17:31
  • I try to do this for the site I frequent, on one, I also casted along a close/down vote. – Braiam Jan 15 '17 at 15:17
  • 5
    There has also been a problem where Hot Network Questions with unedited titles get turned into ads with grammatical errors. Solving this problem would solve this one too. – Thunderforge Jan 18 '17 at 16:36
  • How would you handle things that are borderline nitpicking? Also, do you think editing a title to be good does not require being an active user of said community and knowing how titles there ideally should be? – Buffer Over Read Jan 21 '17 at 8:26
  • There's no reason at all that drive-by visitors shouldn't fix HNQ titles too, @TheBitByte. I often do so myself. But site regulars who are already interacting with the question are really perfectly positioned to do it and should ideally be thinking about this. – jscs Jan 21 '17 at 10:56
  • 1
    @TheBitByte, re nitpicking: I think perfectionism is fine for content that will get viewed a lot. – Aaron Thoma Jan 21 '17 at 14:37
  • @effective.altruistUtoo However, there is a limit. I do not think posts should be excessively edited (even I am usually guilty of this), unless they make a feature where title-changes don't make a post rise to the top. – Buffer Over Read Jan 21 '17 at 14:43
  • 1
    a related post. – Nobody Oct 31 '18 at 1:36
  • I like this, but one issue I see is that it further feeds in to the HNQ self-sustaining cycle: HNQs attract attention which attracts votes which increases their hotness which keeps them on HNQ longer. Plus, the attention attracts answers which bumps it which attracts attention, feeding into the loop. If people are also encouraged to edit the HNQ, it adds even more bumping which attracts attention, feeding into the loop yet moreso. To that end I like the review queue system since it has us do the edits beforehand. – doppelgreener Nov 27 '18 at 13:42
24
+350

An alternative approach would be having Hot Questions go through a review queue for cleanup before joining the Network list.

If for some reason the timely publishing is more important than prior cleanup, the Hot Questions could go on the Network list and into the cleanup queue at the same time.


The existing Help and Improvement review queue could be reused for this, with Hot Network Questions getting top priority.


If the queue is not processed timely, notifications could be issued to mods and/or experienced/ high-rep users.


(­A review system would also enable providing a function to discard Hot questions if they are not good for global publishing, although I can't think of any very good example where that would be useful.

Less good examples:

  • A question has properties that the site does not want to be known for.
    Quoting a comment from anime.SE:

    Seems like every time there's a question about sex on our site, it hits the Hot Network Questions. Like when we had Why are anime girls drawn without pubic hair? all over the sidebar on every site from Puzzling to Parenting.

  • A question is significantly less interesting than its title sounds and would thus be disappointing, creating a less favorable (potentially first) impression of its site.

)

  • 6
    This is a really, really good idea. An excellent use for a queue. Thank you for adding this to my proposal! – jscs Jan 3 '17 at 2:28
  • 2
    I proposed a similar idea here for just reviewing HNQ in general before other sites see them so the sites have a say on what actually goes on the HNQ list. – Dom Jan 19 '17 at 18:39
  • 1
    I think, if HNQ simply lists trending posts, then all trending posts should appear in it. If HNQ is to be a curated, queue-based list, then non-trending questions should also be selectable for curation, with the title maybe changed from HNQ to, say, "Curated Network Questions", or something like that. – Buffer Over Read Jan 21 '17 at 8:37
  • 1
    Yeah, I think there could be use cases for manually selecting questions for the xNQ list. +1 @TheBitByte – Aaron Thoma Jan 21 '17 at 14:32
  • (@TheBitByte, while I neither see a reason for censorship (although it doesn't seem problematic to me in this case), I don't think it's important for most users (to know) how the xNQ list was selected. "Top Network Questions" would be easier to read than "Curated", and consistent with both manual curation and automatic generation.) – Aaron Thoma Jan 21 '17 at 14:32
  • @effective.altruistUtoo I think the word "top" may imply "ranking" or "position". The word "curated" seems to imply better the idea of "selected regardless of ranking" – Buffer Over Read Jan 21 '17 at 14:35
  • @TheBitByte, yeah. It's probably reasonable to take advantage of automatic ways of identifying (potential) xNQs though, regardless of whether there's a manual part in the selection process. ("Selected" would also be easier to read than "Curated", but more strongly implying manual curation than "Top", although it wouldn't exactly exclude automatic selection.) – Aaron Thoma Jan 21 '17 at 14:49
  • 1
    @effective.altruistUtoo What kind of automatic process? Maybe training a neural network on some data could be possibility, but I'm not sure. Besides, if it's going to be automated, why not stick with the original HNQ, which is already automated anyways? About the choice of wording, it may be a bit subjective, but I think the word "selected" does not make it clear how the selection happened, while the word "curated" may imply better that some humans were involved in a manual selection process. – Buffer Over Read Jan 21 '17 at 14:52
  • @TheBitByte, >What kind of automatic process?↦ Well, the current one, by default. >Maybe training a neural network on some data?↦ I like that idea!^^ +1 The manual curation could be used as feedback for a supervised machine learning system. – Aaron Thoma Jan 21 '17 at 17:32
  • @TheBitByte, >if it's going to be automated, why not stick with the original HNQ which is already automated anyways?↦ I don't know. Although I did come up with 2 "less good examples" and added them to my answer. I didn't mean the discard functionality as a recommendation, just pointed out a possible feature with low implementation effort, given the case. Indeed, as of now I would tentatively recommend against it. I moved the point to the bottom of my answer. – Aaron Thoma Jan 21 '17 at 17:32
-4

A reputation bonus for getting a accepted edit on the title would make me edit them. Like effective.altruistUtoo said have them get reviewed. The rep bonus could be from 2-5 rep for the title 2-5 rep for tags and 2-10 rep for the body. That would get me editing a lot more and looking for edits.

Or you can just give a better edit rep boost

  • (+1 Good idea!) a) Although I'm not sure it would be a good idea to distinguish between editing those 3 areas for rep calculation, simply b/c it might be an undue complication. So possibly, simplifying the idea to a rep bonus for editing HQs would be good. … – Aaron Thoma Jan 21 '17 at 14:07
  • … b) An alternative would be to make the rep dependent on the size of the change (measured in bytes I guess), but that might likewise be an undue complication. – Aaron Thoma Jan 21 '17 at 14:07
  • c) The bonus should not go away when a user has gained the Edits are applied immediately privilege. … – Aaron Thoma Jan 21 '17 at 14:07
  • … d) Matching that, it is also an option to put edits to HQs in a review queue even when the edit is from a privileged user. But that's probably not worth it since these experienced users' edits are probably usually good. (→ … – Aaron Thoma Jan 21 '17 at 14:07
  • … ( e) → Sub choice: Whether to publish privileged users' edits to HQs immediately (like the First Posts queue) or only after review (like for non-privileged users). I think immediately would be better, since the edits will usually be good and to prevent conflicting edits – a risk that may be higher when the edit incentivization works.) – Aaron Thoma Jan 21 '17 at 14:08
  • ( f) An edit should probably earn strictly more than 2 rep to provide additional incentive, since 2 rep is what (unprivileged) users would get for such an edit anyway. – Aaron Thoma Jan 21 '17 at 14:08
  • 6
    You, as a low-rep user, are not really what I'm focused on here. I want site regulars to be aware of how these questions look and think about how that reflects on their site from the outside. (Plus you already get rep for edit suggestions.) – jscs Jan 21 '17 at 14:57
  • A red bonus sounds like a great idea, I mean, who doesn't like rep (hahaha), but, how would very small or unnecessary edits made for the purposes of gaining a little rep be prevented? Imagine a theoretical situation where I capitalize, say. 3 or 4 HNQ questions, add a comma here, a question mark there, the small amount of rep then adds up. How would we establish a way to differentiate between honest editors and rep farmers? – Buffer Over Read Jan 21 '17 at 14:58
  • @TheBitByte there Is a min edit amount – Christopher Jan 21 '17 at 16:28
  • 2
    @TheBitByte, I think perfectionism is fine for content that will get viewed a lot (and may be the first impression users get from an SE site). Therefore I personally would even deactivate the min edit requirement for HNQs. (Adapted from my comment in the above thread.) – Aaron Thoma Jan 21 '17 at 16:33
  • @TheBitByte, for dealing with rep farmers: I don't think that would happen a lot, but when it does, a mod could be informed (e.g. through flagging). They could send the offending farmer a warning PM and probably somehow deduct the rep. – Aaron Thoma Jan 21 '17 at 16:37
  • @effective.altruistUtoo Your first comment, I've already replied to it before. Your second comment however, it's not about whether or not it can happen or will happen a lot, prevention is better than cure anyways. And I wasn't asking what to do when rep farmers are eventually found, but how to detect them in the first place. SE may already have some tools........ – Buffer Over Read Jan 23 '17 at 2:46
  • ....... And perhaps they could use those, but I meant, is there something more applicable to HNQ? Or can the general methods they already use be used for HNQ as well? – Buffer Over Read Jan 23 '17 at 2:49

You must log in to answer this question.

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