14

There was a recent case on SO where a user that was highly active in a very specialized tag asked on Meta.SO why their comment was deleted. The comment was linking to a site (non-profit, no ads) created by that user and to one of their SO answers. A few meta users essentially called this user a spammer, because any time you link to your own site and don't disclose it that is spam.

This kind of thing happens regularly in different variants. Users tend to stumble over the self-promotion rules, either because they simply didn't know them and are well-intentioned or because they are actually trying to self-promote. And I think far too many users are far too harsh and aggressive in these cases. I find some of the comments in the meta post very hostile and insulting, and those comments are heavily upvoted. Of course there are some users that self-promote excessively, but the distinguishing feature there is also that it is a pattern of behaviour. In this case it is about a single comment.

In this particular case, I'm familiar with the topic and the website the user linked to, and this only makes it much worse. For this particular, topic the website is one of the best resources on the internet, it is very far from some random blog spam.

There is one aspect I want to exclude for this discussion here, and that is the comment deletion itself because the issue there is not the self-promotion rule alone but also the discrepancy between how comments were designed and intended to use and how people actually use them, which is a giant topic on its own.

I want to focus on what happened on meta here, because while this case is a particularly bad case, it does represent a larger issue in my opinion. Some parts of the community react unnecessarily hostile when enforcing or explaining this rule. Calling anything that contains a link with undisclosed affiliation spam is hostile, insulting and also just plain wrong. Spam gets nuked and the accounts destroyed, in cases of self-promotion usually the rules are simply explained to the user and the links are fixed. Self-promotion is a somewhat related topic to spam, but it is also fundamentally different. Conflating the two issues leads to bad moderator actions and to overly hostile interactions with the often well-intentioned users.

I think we should give users that don't disclose affiliation to non-profit sites far more benefit of the doubt than they get right now. There is not that much harm to this unless it is truly excessive or untargeted. And we certainly shouldn't call users spammers (or their content spam) for breaking this rule. Pointing it out to users nicely should be enough in most cases, and I'd go as far as ignoring this entirely in many cases when the linked resources are well-known and respected in the particular topic.

I'm not sure if any specific rule change is necessary, though it might be helpful. The bigger issue is the mentality around this that has developed among some parts of the community, which is much harder to change than the rules themselves. But I do think that the overly rigid and harsh application of this rule is making Stack Exchange sites a less welcoming place.

2
  • 4
    This certainly brings back memories. I can't say I've ever cared about not disclosing affiliation and don't think I've ever deleted something for that reason myself. I delete obvious spam and then get stuck explaining these "self-promotion" rules to others via support. I understand wanting to see a bias but I'd be curious to see some real data about how many people actually care about that bias or alter their decision to look because of it. I still ask myself: does it really matter?
    – animuson StaffMod
    Feb 17 at 20:04
  • 3
    @animuson there's a version of this that is annoying, a user posting an answer to their own blog post on a bunch of questions where it is just barely related to the actual question. But that is something that can easily be handled by a diamond mod in a targeted manner, and it only becomes an issue with certain patterns or in excess. Feb 17 at 20:06
19

Some parts of the community react unnecessarily hostile when enforcing or explaining this rule.

Because they're trying to explain a rule, not improve the site (neither by improving the utility of self-promotion nor by driving out spam). I believe this is mostly done in good faith, that the folks doing it believe that by doing so the site will be better than if they didn't...

but...

When you become fixated on rules you inevitably forget why those rules exist. And so we get debates over "what qualifies as spam?" instead of "is this good?", arguments over "what does link-only mean?", grousing about "is this VERY low-quality though?"

But hey. At least we're not debating whether bears will eat you if you don't enclose your conditionals in braces. There ARE stupider arguments!

We're over a decade into this discussion now. I felt like most of the discussion then was a red herring, and... I kinda feel like we just keep on following that herring and losing the rabbit.

This time around, the discussion isn't even about answers. I know you don't want to discuss the inherent brokenness of comments, but... To me, they're inseparable, because my preferred approach here has always been to judge links in answers first and foremost based on whether or not they're part of an honest effort to answer the question!

But, Gregg wasn't trying to answer the question. And he wasn't trying to sneak non-answers or promotional answers past anyone either. He was trying to annotate posts with links to relevant resources, and comments were the obvious way to do that. They've always been an obvious way to do that, and folks have always used them for that purpose. Even though editing would be better, there's very little encouragement for that, much less the plentiful extant examples of it that exist for comments. The longer you spend on SO, the more "annotation" comments you'll see, and the more it starts to just seem like a natural thing to do. What you won't see is all of the deleted ones, much less the rationale for their deletion. I'll respect your wishes to not rant further about how utterly broken the comment system is here, but... It is a major factor in this dust-up, just like it has been in countless others.

So that's how we got what we saw the other day:

  • Gregg annotates posts with links to relevant information
  • Those annotations get deleted, because - strictly-speaking - comments aren't for annotating posts
  • Gregg asks on meta, and someone tries to find a rule about it. They fail to do so, but they do find that guide to not spamming in answers that I wrote years ago, which mentions answers 21 times and comments zero times and also doesn't discuss annotation or editing.
  • Gregg interprets being given a link to a guide to "not being a spammer" as the same as being called a spammer, and things go downhill from there.

Note that BoltClock did try to provide some useful guidance there, but... I tend to think even that is a bit unnecessary, since Gregg has listed those sites as "things he worked on" in his profile for years and I highly doubt anyone ever considered this "promotion" to be even the least bit underhanded. In short, there was no actual problem to be solved here beyond that caused by our embarrassingly broken comment system.

Oops. I said I wouldn't rant more about that. Sorry.

It's true though.

But ok. As far as behaviors go... This was all sorta doomed from the start because at no point did anyone sit down and think, "what problem are these specific comments causing, and how can we fix them?" That never came up. Nobody in that entire thread ever pointed out an actual problem. Despite all the speculation, we don't even know that those comments were deleted for self-promotion - we do, however, know that there is no spam flag for comments, because this just isn't a big problem in general. Since Gregg didn't mention being contacted by the moderator team about his spammy ways, I... Kinda suspect they weren't flagged or deleted for that reason.

Just another red herring, another rabbit got away.

18
  • 1
    Which leads to the question of how better do we onboard, and encourage appropriate use of comments. There was this lovely experiment a few years back of re-labelling them that didn't quite get traction but... what else?
    – Journeyman Geek Mod
    Feb 18 at 2:52
  • 1
    That experiment worked shockingly well. But honestly... Unless there's a will to rethink comments as a concept, "onboarding" isn't gonna help here. The stuff comments get used for is... Useful. Folks will still do it.
    – Shog9
    Feb 18 at 3:00
  • Well - Maybe but... well, I think you know better than me the end result was,... those changes never got adopted. And even if we change things substantially, we'll still need to get people using them appropriately, and there's going to be a ton of inertia from both here and how they are used elsewhere
    – Journeyman Geek Mod
    Feb 18 at 3:12
  • Here's the thing: there is a need for almost everything comments are used for. Comments just aren't a particularly good tool for any of them. So we can keep trying to make it work, with a lot of care and patience... Or, y'know. It won't.
    – Shog9
    Feb 18 at 3:16
  • 2
    For how secondary comments are to the Q&A system the community takes them far too seriously. That's my take. As for the guidance I provided, it was mostly in the spirit of letting the reader know what they're getting into by clicking the provided link, which is why I specifically said no need to be formal about it. Feb 18 at 3:35
  • I suppose given that the site is a definitive WebGL resource, saying "this article on the subject" would've been fine too without the need for disclosure of affiliation. Just something would've helped... and FWIW the comment was custom-flagged as self-promotion, whether the moderator deleted the comment on that basis or used it as an excuse to delete it on other (canned) grounds. Feb 18 at 3:52
  • 1
    I also think it's not exclusively a tech problem and there's the social or at least human aspect to the brokenness. I would know, I used to fixate on the letter of the law just as much, until I realized the importance of the spirit of the law. And I think that falls on each one of us who enforces it, be it with tools or with words. Feb 18 at 3:56
  • 1
    I didn't see anything wrong with your guidance, @BoltClock'saUnicorn. I just... Can't shake the feeling the whole thing was kinda unnecessary. Comments got deleted; it's a day that ends in "y". And yeah; DEFINITELY not limited to tech. Anyone who's ever lived in a HOA knows the pain of rule mongering neighbors.
    – Shog9
    Feb 18 at 3:56
  • This is going to sound OOC for me (as far as you remember me anyway) but I think Gregg kicking up a fuss was necessary, since I think (and you agree) the deletion itself was unnecessary in the first place. But I say that only because I'm known for disdaining the "comments are ephemeral and you should expect them to be deleted eventually" mindset. But yeah, me seeing the MSO post as necessary is why I reopened the question minutes after it was closed just to post an earnest answer. Feb 18 at 4:00
  • 2
    See, I love the idea of ephemeral comments; I love the idea of ephemeral everything. Nothing should last forever, that's horrible. But the tragedy with comments is that cheap storage means in digital systems we can have our ephemeral cake even after eating it - I can have every branch, stash, temporary change somewhere in my git repo without tripping over it in my working directory. Pekka & I worked out a similar scheme for comments ages ago, but it was rejected as too complicated. Now, a decade+ later... 🙄
    – Shog9
    Feb 18 at 4:13
  • Anyway, thanks for being cool with that question. Was good to see.
    – Shog9
    Feb 18 at 4:14
  • 1
    I don't think this is limited to meta/power users, or even that it is predominant there, @MadScientist. Heck, I see it a lot more in newer users, and in the behaviors of the company itself. That's why I believe it is so important to focus on the reason for rules - and these rules do have a good reason to exist! I'm reminded of it every time I visit Quora, which... Nominally has the same rules, but... Not the same goals, and certainly not the same enforcement. That site is littered with self-promotion of the most useless and shameless sort, and suffers for it; we should try to avoid that!
    – Shog9
    Feb 18 at 15:14
  • 1
    TBF, it's awful hard to prove anything WRT comments, @Nick - there's no visible history of deletions, so most attempts suffer from either too few samples or some form of survivorship bias. It's one of the most opaque parts of the site, even for avid users - and the massive variance in perceptions reflects that. We all tend to let our perception of both how comments are used and how they are moderated become skewed by what we actually see, ignoring the (often massive) volume of comments that are deleted before we have a chance to see them.
    – Shog9
    Feb 18 at 16:14
  • 2
    This is what I mean by sampling bias, @Nick - you see your comments deleted, you see similar comments remain, you don't see similar comments that were also deleted. Flagging doesn't help, because that's also opaque: you don't see flags on your own comments nor can you flag them, and you can't see others' flags on other comments (much less effective flags on already-deleted comments) so trying to black-box the system by flagging other comments just gives you a different but still skewed sample - you're almost certainly flagging "survivors".
    – Shog9
    Feb 18 at 16:21
  • 2
    Shog, I missed those answers of yours, and really happy you keep posting them even after everything that you've been through. Double kudos!! Feb 22 at 6:54
8

There appear to be several aspects to your question, and while they're definitely related to each other, I'll try to address them separately so as to keep it a little organized.

1. Is not disclosing affiliation a problem in general?

I would say, yes, it's a problem. While undisclosed affiliation with a for-profit link is definitely problematic, even if it's non-profit there are issues. There are non-monetary benefits that one gets from these resources being widely shared, such as increasing one's visibility in the community.

When I see an answer, or even a comment recommending that I look at a resource, it matters very much to me whether the recommendation comes from the person involved in producing that content. Even if there is nothing to be gained, a person is always going to be biased towards content that they have produced; it's just human nature and there's no getting around that. Understanding the biases of the person presenting the content is an important part of judging its merit, and disclosing affiliation is crucial to this.

2. How much undisclosed affiliation is too much?

I think the previous section covers my views on this quite clearly. Any amount of undisclosed affiliation is too much. There's no need for there to be more than a single case of violating this, whether in an answer or in a comment, for it to be a problem that needs to be addressed. Which leads to ...

3. How should this problem be addressed?

This is a much harder topic to deal with, since there is some amount of subjectivity involved. It seems clear that if a user's only contributions are for the purposes of promoting their own content, that should be treated as Spam. If the posts (or comments) are relevant to the question, but a vast majority of them are promoting their content, I would consider that Spam as well.

If only some of the user's contributions are self-promoting but relevant, then ideally they would be informed of the self-promotion policy in a comment, and perhaps contacted by a site moderator (if this is brought to their attention, maybe via a flag) to explain the rules regarding affiliation. As you've mentioned, in the vast majority of cases, simply being told that affiliation is required should be sufficient for the user to start disclosing that in their contributions.

If the user continues to flout this policy, then I think it's appropriate to warn them more directly. Perhaps discussing this in a private chat room to understand their objections would be a good idea, but slightly more heavy-handed approaches such as suspensions may be in order as well. If they still persist in this behavior, then treating them as a Spammer is the final, but I think unavoidable, resort.

4. How should this policy be communicated to users in public forums?

You're right, we as a community can work a lot on improving our communication styles to be more productive than hostile. The Meta post that you link to is an example of how that conversation should not happen (I believe a common term for how that ended up is "dumpster fire"). I watched that Meta unfold and I think I even left a comment at one point, but then deleted it right away, as I didn't feel I could contribute to that discussion in any profitable way.

I would like to point out that the OP of that question admitted to posting multiple comments of this nature, i.e not disclosing affiliation with links they shared. Also, the OP made several comments along the lines of "huh, this is spam? Then all my contributions are spam. Delete them all", which only added to the fire. But you're right, the root cause of how that conversation ended up is the use of the word "Spam", and categorically describing the OP as a "spammer", which understandably raised the OP's hackles.

While I agree that this word has negative connotations, and how we convey this should be improved, I don't disagree with the usage of the term. I personally try to use "You could be considered a spammer if you keep doing this", and let either moderators, or Spam flags, do the job of saying "You are a spammer". I don't know if this is what you have in mind when you refer to making the conversations less hostile in this regard.

5. What can we do about veteran users who are unaware of this policy?

A large aspect of your question seems to be focused on the user who was affected by the fallout from that Meta. If the result of that Meta is that the user chooses not to contribute to the site any longer, that would be a real shame. I strongly suspect that if they were made aware of the rules either earlier, or in a more gentle fashion, we would not have reached this stage. If I were a site moderator I would reach out to them to engage in discussions, and I sincerely hope that this is already underway.

However, and I want to make this abundantly clear, I am very opposed to the idea of relaxing the requirements on affiliation, or what we consider Spam, based on how it might impact some veteran users. I do not want users in general getting the impression that a little bit of undisclosed affiliation is ok, just because the link is to a non-profit website, or the user has a lot of reputation, or that there's "no harm done really". This is a very slippery slope, and I'm uncomfortable with loosening the requirements. Spam is Spam, regardless of whether we sugarcoat it when describing it. I still want users to flag the cases they think are problematic. This has the advantage of bringing potential problems to moderators' attention even faster, and mitigates the potential for veteran users to discover these rules in an unpleasant fashion (as happened in the Meta).

4
  • 3
    I really don't see why we are supposed to enforce this rigorously on every single instance. Not all links are equal, it makes a huge difference if this is a recommendation for a tool, a link to a personal or commercial blog, or a link to a widely used resource in a particular area. This is really something I'd judge on a case-by-case basis, and I would not consider individual cases any issue (unless we're getting into the more blatant commercial stuff). Feb 17 at 21:40
  • 6
    @MadScientist I'm not sure what you mean by "enforce rigorously on every single instance". As I mentioned in the 2nd para of #3, in the case of a single example of undisclosed affiliation, I think the user should be told about the policy in a comment, that's all. If you're suggesting that it shouldn't even be mentioned, then I have to disagree. It's much better to nip it in the bud, before it becomes a more serious problem that needs to be dealt with.
    – cigien
    Feb 17 at 21:44
  • 1
    Are you saying that an editor on a canonical reference, like wikipedia, should mention that affiliation every time they post a link to a relevant article that they happen to have touched some time in the past? Feb 18 at 14:23
  • @BartvanIngenSchenau No, I don't think that's necessary, since wikipedia is a collaborative effort. Similarly, if a user has contributed to a github project at some point, then they don't need to disclose affiliation. However, if a user is the primary contributor, or author, or founder, of a github project, or any blog, etc, then I think affiliation should be disclosed. Deciding where to draw this line is hard of course, and requires judgement. My personal preference is to err on the side of disclosure when in doubt. Also, I don't think whether the reference is canonical matters too much.
    – cigien
    Feb 18 at 14:30
8

I totally agree that the Meta post wasn't handled adequately and I contributed to the perceived hostility. I'm sorry I didn't leave any comments to turn the debate into something a tad more constructive.

We have turned a large bunch of our curators and moderators into link checking slaves with only a true or false outcome. Some seem to think this rule abiding leads to mathematical accuracy, an ideal world. That is not how it works.

I do think that a chat message, a comment, a post and your about me are different and as such require different interpretation for when a link is posted and when that link becomes a problem. Yes, it does matter if a fresh account posts links to vague external sources or if a user with some track record post links. That is a bit of a bummer for the flaggers because now they have to use their judgment instead of their rulebook. It is probably not what they want to happen, too much uncertainty.

This is not an incident where we go extreme on the handling of flags. A quick search returned

and I'm pretty certain in 2019 we have had a ban all GitHub links saga, because "spammers" spam their code on a coding Q&A platform. The meta posts must have been deleted since as I can't find those anymore.

To conclude:

There is no need to overly hostile enforce self-promotion. No human will die if you first try to, I don't know, maybe talk to a user that obviously tries to be helpful? You can @-mention them and politely ask or warn them that their link might be considered a tad too much on the promotional side. Yes, that is a bit scary if they start to talk back, but then at least you know you're not dealing with a spam robot or a poor underpaid clueless soul. Let's be nice to users that share and link to their free knowledge.

As some might still think that the help article about self-promotion is the most important rule to follow, let me remind you of No name-calling or personal attacks. in the code of conduct. If you call out someone in a comment as a spammer, you're in direct violation of the code of conduct and that is ground for suspension. And if you're a moderator your diamond should go.

For now, until both guidance and we get ourselves sorted, let me add some advice if you're a blog owner, Medium author or GitHub exploitant. Create a sockpuppet/alternate identity on those external sources. They have no idea it is you. You can link as much as you want here on Stack Overflow without affiliation because you're not affiliated. Pr0blam solved.

As for the general idea that seems to float that spam is spam: Happy hunting

0
2

That comment thread derailed quickly. Some moderator intervention might have been useful in the heat of the moment, to let things cool down. Things escalated back and forth, both sides being convinced about themselves being right. Now the dust has settled there is not much to be done there.

Besides that I see two perfectly fine answers to the question. They both explain the guidelines that we have on this network, and they do it in a respectful and educational way.

Regarding the tone of the comments, you are right, they are not very friendly. But then again I personally don't find them overly offensive either. They are short, direct and to the point. That is a style of communication you can either appreciate, or take offense in. And to be honest I can understand why people write these short and direct comments in response to self promotion, as it is abundantly present in the network. If you follow reports from SmokeDetector on a regular basis you see plenty of cases of it in posts. That mean there most be a multitude of occurrences in comments.

In case the person posting self-promotion is a new user I tend to write a somewhat longer comment, otherwise I'll comment something along the lines: "Not disclosing your affiliation with the linked content is considered posting spam.", and flag the comment as such. Comments are second class citizens just for these reasons.

9
  • 6
    I find "your comments are spam" very offensive. But ignoring that part of the comments, a large part of what I find insulting is the connection users make between this behaviour and spam, because that is simply wrong and a bad association. And while I didn't want to make this a focus of my post, the damage this incident has caused to the webgl tag on SO is enormous as this user represents an unusually large part of the answers there and is now gone. Feb 17 at 20:39
  • 1
    I read the blogpost this user wrote on their personal website, and I can understand their point of view and why they are upset about this matter. What I did find disturbing about the post however was that in being an very active member of SO for over 9 years, they only now realised that comments are ephemeral. If for such a veteran user something so fundamental about how comments work isn't known to them tells me we do something wrong with regards to expectation management around comments.
    – Luuklag
    Feb 17 at 20:56
  • 3
    I think that is one aspect where users like us that are very active on meta dramatically underestimate how familiar users that only post questions or answers might be with various rules. You can heavily participate on an SE site without ever enaging with many of the moderation aspects, not to speak of even meta. And I don't see anything wrong with this, I'd rather have a subject-matter expert spend their time on answering that on cleanup Feb 17 at 20:59
  • 3
    I don't disagree on that, hence my part: "If for such a veteran user something so fundamental about how comments work isn't known to them tells me we do something wrong with regards to expectation management around comments."
    – Luuklag
    Feb 17 at 21:02
  • 1
    The comments part is much easier to diagnose, they are very often used in a way that is relevant and useful, but still not according to the actual rules. The rules are inconsistently enforced and enforcement is silent unless you know the comment existed. So you can't actually figure out the comment rules from simply observing the site, because a very large fraction of comments doesn't follow the rules. Feb 17 at 21:05
  • The "hostility" in the question title is not hostility in meta comments to the link poster but hostility of the interpretation of the link postings as self-promoting spam.
    – philipxy
    Feb 17 at 21:23
  • 5
    @philipxy I don't follow. Promoting one's own content via links without disclosing affiliation is spam. The network policy is clear on that. Why is it hostile to interpret it as such? Note that the fact that the OP seemed unaware of what they were doing counts as spam doesn't change that.
    – cigien
    Feb 17 at 22:12
  • 3
    @cigien this is one of the cases where the user's own content is also the kind of reference content anyone else would use for that topic. To me there's a very big difference in linking to your own blog that nobody knows about, or linking to an established reference where you either contributed or created it entirely. Feb 17 at 22:54
  • 5
    @MadScientist Yes, there is a difference, but I'm not sure why it's different enough that disclosing affiliation would no longer be required. I would understand if disclosing were a difficult, tricky or tedious thing to do, but given that all it essentially requires is phrasing "see this content", as "see my content", I don't see the issue with simply requiring disclosure in all cases.
    – cigien
    Feb 17 at 23:02
-3

It suffices to say the user posted a question that was mostly between him and a moderator. You can go to the thread and count the number of false accusations against the OP - most of them against the CoC in terms of tone and content.

[Correction, the most reckless accusations in the comments have since been deleted. What's left is a string of defenses from the OP.]

You must log in to answer this question.

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