21

I recently came across this answer (to a question on implementing echo cancellation in Java) in the low quality review queue:

The only way to supress echo is to use "C" libary of AEC and call it from Java. AEC is a very computationally extensive algorithm and only "C" code allows optimization under your platform. Commercial optimized AEC can be found for example here IntegrIT. Btw, AEC will work properly provided you have done proper integration with audio drivers...

Without forming any opinion on the technical content of the answer, I viewed it as a valid answer (albeit perhaps lacking thorough detail) and clicked "Looks Good". In response, I saw the following banner:

This post has severe quality issues. 

It is abusive nonsense, noise, spam, blatantly off-topic or otherwise irredeemable - readers will find it offensive or repulsive rather than helpful. You should delete or recommend deletion when reviewing such posts.

Don't worry, we've already handled this post appropriately - but please take a minute to look it over closely, keeping in mind the guidance shown above.

In what way is the post "abusive nonsense", "noise", "spam", "blatantly off-topic" or "otherwise irredeemable"? In what way will readers find it "offensive" or "repulsive"?

Should one really "recommend deletion when viewing such posts"? I'm not entirely sure how to differentiate between "such posts" and 99% of the posts on SO. Grateful for some guidance!

Improve this question
17
  • 15
    "Without forming any opinion on the technical content..." => "Learn to love that Skip button."
    – gnat
    Dec 29, 2012 at 13:53
  • 2
    why didn't you skip? I do this almost always when I can't form "any opinion on the technical content" - expecting more competent reviewers to take care of it. This of course slows down my bagde progress, but I'm reviewing not for badges. And, well, past experience taught me when there's link in the answer in LQ queue, do check it very thoroughly when I am not inclined to skip
    – gnat
    Dec 29, 2012 at 13:57
  • 1
    @gnat: I can assure you that I'm not reviewing for badges either; I've previously been told, here on Meta, that a short two-line answer (on which I did have an opinion of the technical merits) was, contrary to my belief, valid because the technical merits (or lack thereof) are not relevant. AIUI, flags/reviews are not intended to assess technical merit but rather the overall quality of the post and therefore should not require any knowledge of the subject matter.
    – eggyal
    Dec 29, 2012 at 14:00
  • when you see link in LQ queue answer that looks otherwise innocent - it's an indication to check for spam (note this is one of the officially stated reasons for posts to get in this queue "for content"); if you don't feel like checking, then skip should be your next option. And, well, spam has nothing to do with "technical inaccuracies" that's why having spam in LQ queue is consistent with guidance you referred to
    – gnat
    Dec 29, 2012 at 14:04
  • 2
    @eggyal, every time you click "looks good" without thinking, you're destroying chance to review for people who could understand the content. I don't think that's good.
    – Oleg V. Volkov
    Dec 30, 2012 at 13:27
  • @OlegV.Volkov: Hang on... when did I say that I reviewed "without thinking"? I gave the review considerable thought, trying to apply advice that I have previously received here. I frequently skip reviews about which I am unsure, but in this case I was confident of my decision and remain confused by the conflicting advice received in different cases.
    – eggyal
    Dec 30, 2012 at 15:12
  • @gnat: I also object to the interpretation that the review was "outside my area of expertise". I have never said that. I merely assert that I did not form a view on the technical merits of the answer, as it was my understanding that reviewers should NOT assess technical merits.
    – eggyal
    Dec 30, 2012 at 15:22
  • well the very fact that you pressed looks good at the spam shows it was out of your area of expertise - at the moment of pressing it - and to me, it looks unreasonable of an audit failure message to ignore pointing to skip option here. Whether you were capable to recognize it or not through thorough study is a different matter... or not too different - I for one wouldn't participate in review that would bring me difficult items like the one you stumbled upon without skip option - I would find it simply unfair if the system would force me to spend a lot of time deciding about potential crap
    – gnat
    Dec 30, 2012 at 16:17
  • 2
    @gnat: I fail to see how you reach the conclusion that it was "outside my area of expertise" from the fact that I thought the quality of the post was okay irrespective of its technical merit, which remains my understanding of what reviewers are supposed to do. There is no suggestion that it would have taken "thorough study" to assess technical merit.
    – eggyal
    Dec 31, 2012 at 9:02
  • you mean, recognizing spam turned out harder than you expected of items in this queue?
    – gnat
    Jan 1, 2013 at 7:51
  • @gnat: No. As I've said in comments beneath, I don't think it's correct to classify this particular answer as pure spam; but even if one felt the link was inappropriate, surely a better response would have been to edit and remove it than delete the answer altogether?
    – eggyal
    Jan 1, 2013 at 11:09
  • then why did you accept an answer that classifies it as spam? as for editing, you might be interested in checking this MSO discussion: Should Spam posts be edited?
    – gnat
    Jan 1, 2013 at 12:42
  • @gnat: I accepted an answer that explains why the banner appeared and answered my question "In what way is the post "abusive nonsense", "noise", "spam", "blatantly off-topic" or "otherwise irredeemable"? In what way will readers find it "offensive" or "repulsive"?".
    – eggyal
    Jan 3, 2013 at 12:59
  • @gnat: I just read You're doing it wrong: A plea for sanity in the Low Quality Posts queue and it reminded me of our discussion here—thought you might like to contribute over there?
    – eggyal
    Mar 9, 2015 at 21:02
  • 2
    Nice I just got another one after reviewing like 100 where someone was asking something about vertical alignment how to do something with vertical alignment and at the end he posted his blog. I marked it as Should be improved and I got a ban for a month cause it was 'spam'. I'm done with reviewing. Worst system I've ever seen tbh.
    – Loko
    Mar 27, 2015 at 11:20

3 Answers 3

Reset to default
19

If you look carefully at the post, it's in fact well crafted SPAM. It begins to vaguely answer the question, then does nothing but promote a commercial product instead. This is actually rather common, make it look 'just enough' like an actual human being trying to help in hopes of avoiding detection.

It was deleted by the community after accruing enough SPAM flags. While moderators are being stricter in quality related flag handling to ensure that audits you see in review aren't confusing, this one was pretty cut and dry.

Let's take that answer apart sentence by sentence:

The only way to supress echo is to use "C" libary of AEC and call it from Java. AEC is a very computationally extensive algorithm and only "C" code allows optimization under your platform.

... that's a start. Vague, but a start. After reading that, you'd expect to see something supporting that sentence. But that didn't happen, what follows is this, which is pretty much just a link to a commercial product to alleviate all the pain:

Commercial optimized AEC can be found for example here IntegrIT. Btw, AEC will work properly provided you have done proper integration with audio drivers...

... and the answer ends. So to recap we have:

  • A vague explanation of what might be going wrong, which is never expanded further in the answer
  • A link to a commercial product, without even an example of how to use it
  • EOF

SPAM aside, the answer has other quality issues. It really doesn't answer the question fully and just provides a link that may or may not work in a few months. If you edit out the whole link part, you end up with at best a vague comment. This is because you just removed detail from the answer on how to do what it recommends, which was a link to a commercial solution and nothing else - not even an example of how to use the recommended solution.

You don't need to have an in depth domain knowledge to judge quality, which is all you're judging. Just look at the post on the merits of how well written and seemingly helpful it might be in the context of the question. Also, don't forget - you can skip to the next item if you're really unsure.

Don't worry about not passing an occasional audit. They're designed to help you hone your moderation skills. If one catches you, just take the post apart and you'll usually figure out why. If after that you think the audit was completely off base, it's definitely worth reporting.

Improve this answer
8
  • 1
    I can see that the post might be viewed as spam, but (personally) I think it's pretty harsh to assume that was the user's intention. Surely editing out the link to IntergrIT (or adding alternatives, including non-commercial/FOSS ones) would have been a better response than deletion? Or, at very least, commenting to advise the (new) user of why their post is not up to scratch? In any event, I think this review is sufficiently subjective/border-line to be a bad case for an audit.
    – eggyal
    Dec 29, 2012 at 13:44
  • 3
    Well, again, the community deleted this without moderator intervention. Looking at it as a mod it trips my 'spam-o-meter' just by its characteristics (brand new user, short vague blurb, doesn't really answer the question, link). It's a pattern we see often. That aside, if you remove the link bit, what are you left with? At best, a comment really.
    – Tim Post
    Dec 29, 2012 at 13:46
  • 6
    @eggyal I disagree with the bad case. Audits are designed to teach users how to moderate just as much as they're designed to catch people that are obviously not paying a bit of attention to what they're doing in order to get a badge. As this is a very common case, I think it's well suited as an audit.. but my opinion only.
    – Tim Post
    Dec 29, 2012 at 13:49
  • 1
    When reviewing the post, I was particularly mindful of this answer to a previous question of mine. In this case, with the link removed, one is left with (paraphrasing) "for performance, you should call a C-implementation of an Acoustic Echo Cancellation algorithm from within Java; commercial products exist." - it's not a great answer, but I feel it is as much an answer as the other that was deemed valid?
    – eggyal
    Dec 29, 2012 at 13:53
  • You seemed to be more focused on the technical merits of the other answer, however the answer itself is at least trying to be useful in the context of the question. It also wasn't vague and linking to a third party solution, which is why this one was zapped as quickly as it was without comments from other users calling for it to be improved, or users just editing it. Context is really important, so it's difficult to compare answers on totally different questions.
    – Tim Post
    Dec 29, 2012 at 14:27
  • 1
    I think this just reinforces my point that the decision is sufficiently subjective as to be a bad case for audit. I would argue that this answer was "trying to be useful in the context of the question". I have read elsewhere on Meta and in the FAQ that valid answers include posts that do not answer the question, but still provide helpful advice. I still fail to see how this post (minus its link to a commercial product) fails that test. Anyway, I've said my piece - thank you for your thoughts. :)
    – eggyal
    Dec 30, 2012 at 15:30
  • 1
    @eggyal I'm going to bring this up internally once the holidays come to an end. There is more than a bit of merit in what you have to say. A broader issue is consistency when it comes to what you find on meta.
    – Tim Post
    Dec 30, 2012 at 15:53
  • 1
    I recently stumbled upon a user repeatedly promoting his own product. Checking out his other posts, I found yet another example, which struck me as a clear case of spam according to your explanation of why I failed this review audit; dutifully, I reported it as such but the flag has been "declined - a moderator reviewed your flag, but found no evidence to support it". If a moderator perceives such a post to be passable, how can an almost identical one be suitable for audit?
    – eggyal
    Dec 18, 2013 at 18:34
3

Some might consider the answer spam because it mentions a particular c implementation. But that link could be edited out (maybe just a reference to the fact that several competing commercial products exist). I think the point of the answer was implement it in java by calling a c program - which I agree might not have been the OP's intent but is still a somewhat valid answer. It might have been better worded as 'if performance is an issue, you may want to use a language more optimized for bitbanging such as C - although others exist as well such as erlang or haskell that might be better suited for a distributed environment (wasn't clear from question if the VoIP code was server/cloud resident".

Improve this answer
3

I think the issue you discovered can be attributed to a problem in an "educational" message posted in the case of audit failure. Particularly troublesome is the following requirement:

"You should delete or recommend deletion"

For items that are in Low Quality queue because of content (like spam), above would be the right thing to do only for reviewers who are sufficiently competent to evaluate this content.

Thing is though, this is not always the case, as it happened in the audit you stumbled upon:

  1. you are not sufficiently competent to form an "opinion on the technical content"
    - which is by the way 200% legitimate since no one can know everything

  2. spam post has been "well crafted" - as explained in another answer, it even began to "vaguely answer the question"
    - which makes it really highly possible to pass under the radar of even a careful reviewer who just happened to be not sufficiently skilled to judge content

In cases like above, skipping the review, expecting more competent reviewers to take care of it would be a reasonable option (and by the way it wouldn't fail the audit).

The problem is, the audit failure message does not even mention skip option - which makes it rather misleading than educative.

Some ways to address this issue that come to mind are either to 1) edit the audit failure message to explain about skip being allowed in case of doubt or to 2) avoid having audit items where user is expected to evaluate technical content ("well crafted spam").

Improve this answer

You must log in to answer this question.

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