5

Background: I've been doing a lot of comment flagging on Stack Overflow recently to clean up comment sections of highly-viewed and/or highly-scored posts. I flag "no longer needed" for things like "thanks!", "works!", "you saved my life, I am eternally grateful", "I love you", "I'd give you my firstborn but I already gave it to another user" etc. (and have developed a great empathy for iOS devs in the process) I haven't verified this, but I can imagine this isn't specific to just Stack Overflow.

In my comment flagging, I've encountered an interesting category of flocky comments (comments that have (/attract?) many similar comments): "this also works in version X" comments- that for reasons I'll explain, I feel like I could/should do something about, but don't know what that something is, and doubt all my ideas.

Here are some examples from Stack Overflow (if some get deleted, you'll just have to take my word for the counts):

Should such comments stay undisturbed as comments? It does seem to me that they fall under the "Add relevant but minor or transient information to a post" category of "When should I comment?".

I've been leaving them be, or just flagging ones that duplicate existing ones (same version), but I'm wondering if there's anything more constructive that I can/should be doing about them.

Ex. Should I be editing to inline that information into the post? I'd feel uncertain doing that, since I typically am not able (or just not willing to spend the time and effort) to verify that information myself. I wouldn't want to inline incorrect information into a post.

I also noticed that some of the posts these comments belong to seem to "invite" such comments by the post saying "this worked for me in version X", which tempted my mind to think "... would it be better if I just deleted that?" Of course- I have no way of knowing, and I have no SME in most of these areas, so I just don't touch the post either.

I'm sure that this information holds some amount of value. I think that makes the current situation somewhat non-ideal:

  • This info is not very visible in the comments. Comments only show a fixed, small number of the top-voted comments by default, and I've seen comment sections with much more than five such version comments.

  • Having one comment per version makes the information spread out and fragmented. It would be more ideal to have that information gathered together. (The issue with that is verifying which info is true to put it in some "good pile")

To reiterate: Is there anything constructive that I can/should be doing about these "flocks" of version-relevancy comments? (Taking into consideration that I don't have the ability or time/motivation to verify their correctness)

7
  • 3
    Valid discussion, but you'll get more relevant answers in MSO, this is marginally off topic here because it's 99% about SO so better asked on its site meta. Dec 26, 2022 at 11:05
  • 1
    @ShadowWizard Good point. I'll do some digging tomorrow on other technology sites like askubuntu and superuser just to see if it's actually a wider problem. It might be the case that it's better to ask on a per-site basis if it ends up being a really technicality-based discussion, which it definitely could be. I'll leave this here for now.
    – starball
    Dec 26, 2022 at 11:09
  • Thanks for this question! I know I've made such comments myself in the past and am now rethinking whether I should do so. Dec 26, 2022 at 11:20
  • 3
    Version comments are relevant to multiple sites, so I think this question is ok here on MSE. OTOH, if you're primarily concerned about such comments on SO, it would be a good idea to ask about it on MSO. FWIW, there are lots of questions on MSO regarding "the version problem", eg meta.stackoverflow.com/q/405302 There might be existing discussion there on version comments, but it's not easy to search for.
    – PM 2Ring
    Dec 26, 2022 at 17:54
  • I've found both Nij and Journey's answers helpful, but for the sake of choosing one, I have currently accepted Nij's.
    – starball
    Dec 26, 2022 at 21:29
  • 1
    So, after looking at your examples, specifically about Linux versions, this is one of the cases where users do not use backports as recommended by the developers in the case of phpmyadmin (which is the first thing you should do on a "stable" distro) and irrelevant for people running updated software on the java one. In both cases, the entire question is obsolete as a common user would never find these issues unless they go specifically looking for it.
    – Braiam
    Jan 7, 2023 at 11:24
  • 1
    Going off into the other questions there's a common theme: most questions are caused by bugs (and apple being apple). So, I doubt you can make a generic stance on this. You need to find examples that are more varied for this. I have extensive experience answering all kinds of questions, as you can see in my profile, and rarely if ever these comments are useful, since if I'm affected by an issue I go first to the upstream, rather than ask a question.
    – Braiam
    Jan 7, 2023 at 11:28

3 Answers 3

5

Knowing the limitations of an answer is just as important as the answer itself.

Particularly older versions of software tend to have less cross-compatibility, and major overhauls can suddenly break a lot of work.

Removing existing information about version validity would be vandalism. Do not make such edits unless you know surely that it is false. Even then, consider carefully whether the original author has made this statement with a context or other clarifying details in mind.

Comments that add useful information about version validity are constructive and could be the basis of an edit to the answer (or question, in the case that it's about an error or bug or a function etc. that may have changed behaviours) especially if they are parenthetical, for example, changing

In FooBar 2.7.18 you can add the Widget under the Woogle menu to correct the Wiggle factor...

to

In FooBar 2.7.18 (and later versions to at least 3.1.4) you can add the Widget under the Woogle menu to correct the Wiggle factor...

If the constructive feedback has not been accounted for, the comment should not be flagged as No longer needed, nor should it be removed.

If you can't verify the information yourself, either as a subject expert or as a person reading documentation and other answers, avoid these edits unless you are submitting them from a low-rep account so they can be reviewed.

If you can't firmly reject the information as false or inaccurate, avoid seeking its removal. Consider requests to the commenter for supporting information, but bear in mind they're 1. unlikely to still be around, 2. unlikely to respond if they are, 3. unlikely to be positive if they do, and 4. adding more comments that will need cleaning up by curators or moderators.

4
  • I don't understand the part at the top about edits. Since when are comments editable by anyone but the author? Jan 27, 2023 at 6:55
  • @KarlKnechtel I'm pretty sure Nij is referring there to the part of my question post that starts like "I also noticed that some of the posts these comments belong to seem to "invite" such comments by the post saying "this worked for me in version X", which tempted my mind to think "... would it be better if I just deleted that?"". Nij, can you confirm? If so, perhaps you can prevent that confusion by quoting that section of the question post in your answer post before you address it.
    – starball
    Jan 28, 2023 at 17:57
  • This post begins by talking about answers. The subject does not change until it is changed, which is after that paragraph.
    – Nij
    Jan 28, 2023 at 23:45
  • Related: I just saw this in the guidelines for reviewing suggested edits: "Common types of edits to Approve: Edits that attempt to add clarification to an answer, like “this doesn’t work in Windows 8”, or addendums to the post"
    – starball
    May 10, 2023 at 1:06
1

It's a clarification and additional information. To an extent, that's what comments are for.

While comments are second class citizens, it's worth considering critically if the comment actually has no value, or if letting folks know an older answer still works for newer iterations of an OS holds value. It might not fit into the original question (and we do encourage folks to be specific in their questions), and 'new' question/answer pairs for other versions might be just as noisy. It might not be neat, but it's a usable way to add additional information for a post. Practically a single comment saying it works for other versions might be less noisy.

If it's a dupe, it's worth cleaning out. Naturally, "me too" comments don't have value.

Another wrinkle here is SO has ... a lot more flags so something like this might be considered less critical to take care of. It might be a bit troublesome if there's a load of these flagged at once.

To reiterate: Is there anything constructive that I can/should be doing about these "flocks" of version-relevancy comments? (Taking into consideration that I don't have the ability or time/motivation to verify their correctness)

Flag the dupes, leave the ones you can't be bothered to check alone. :D

5
  • 1
    "we do encourage folks to be specific in their questions" specific about relevant information. OS version, specially on Linux, is rarely relevant.
    – Braiam
    Dec 26, 2022 at 15:19
  • 1
    I disagree - maybe cause I mainly use Ubuntu and windows, but there often are drastic differences between versions, and defaults in those. Dec 26, 2022 at 19:27
  • Try using a rolling distro, like Arch or Debian testing/unstable. You will discover that versions of the OS doesn't matter, but the specific patches applied/compiler options enabled to the components are.
    – Braiam
    Dec 28, 2022 at 15:30
  • If someone does it for rolling media, that's another story. Some of us run stable OSes Dec 28, 2022 at 22:32
  • Seems like you missed the point. I can have kernels from 5.10 to 6.0 (almost 10 versions) in Debian stable. Kernel devs do not ask you what version your OS is, but what uname -r returns, because of this. Same with every other developer, they care about the version of their thing because their thing is what the user interacts with, not the strings in /etc/os-release. (BTW, if you still insist, there are many packages that have several versions inbetween for a single "stable" version of the OS, so please, be aware of that when responding)
    – Braiam
    Dec 29, 2022 at 17:21
1

I can speak from experience dealing (either from sysadmin or programming) with these, most of the time, they are irrelevant unless something actually breaks. This is to establish known good boundaries, for example "for versions higher than X use foobar, before those you must just bar". In other words, to define the limits of the solution presented, if there are any. Is the case of "no news is good news". When writing for solutions, you rarely see an answer that targets a specific version that didn't have any stay in the real world, so the list of "working in version X" will be longer than the list of versions that it doesn't work.

You must log in to answer this question.

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