86

The announcement of the opinion-based question experiment did not specify the duration or end date of the experiment. If the announcement is to be believed then staff are no longer monitoring that post for feedback about the experiment itself, leaving only the What questions did you like or not like? post for any kind of feedback, and that narrowly-scoped channel seems to have mostly played out (there are still some posts and comments coming in, but they don't appear to be highlighting anything novel).

Apparently, staff are determined to continue the experiment until some endpoint known only to them is reached, despite the fact that it is actively causing harm by misdirecting questions that ought to be traditional Q&A into the opinion-based stream, without even a way to redirect.

So what is the endpoint? Is the nearly three weeks of the experiment so far not already sufficient for an alpha test?

30
  • 21
    i mean... how long did the discussions experiment last? 2 years? Commented Nov 13 at 22:45
  • 5
    yea, 2 years: Discussions experiment launching on NLP Collective Commented Nov 13 at 22:51
  • 26
    I think that what you and I call "active harm" is called "increasing engagement" to the company. I don't know if this has actually increased engagement, but this sorry state of affairs does not seem to be considered a problem by staff. Commented Nov 14 at 0:13
  • 15
    It's decrease my engagement. I mean I still "thumbs down" garbage when I come across it. Which is 90% of these questions seeking my opinion. I can't post an actual answer to the 10% since, it will just be what is essentially a comment, which does not help the community. I can tell you that I have submitted less answers since this experiment started. Commented Nov 14 at 2:47
  • 12
    @SecurityHound but but but the number of answers per week increased by 500 if you count the replies as answers! Commented Nov 14 at 2:56
  • 8
    Yeah, it needs to be done with as soon as possible. I think maybe 90% of the opinion-based questions I've seen come up in tags I watch have absolutely no interactions, the rest get comments basically suggesting the OP delete should the post and create an actual post that can be answered. Commented Nov 14 at 5:38
  • 2
    "Apparently, staff are determined to continue the experiment until some endpoint known only to them is reached" If you read Opinion-based questions alpha experiment on Stack Overflow there's a section "Measurement and Success" that describe the goal of the experiment... Commented Nov 14 at 9:03
  • 1
    If you mean the stated goal is the definition for the endpoint, then I have to say it is not very well defined. Commented Nov 14 at 9:26
  • 1
    My expectation based on other recent experiments is that it will not "finish" as much as simply graduate at some point. If we're lucky we'll get some fixes for the worst issues before then. Commented Nov 14 at 11:09
  • 3
    Is there any way to filter against them? I notice that upvotes on a good reply do not add to reputation so why should anyone bother to engage with this "experiment". Commented Nov 14 at 13:55
  • 2
    i mean... this is just discussions v3, was discussions really revoked? Commented Nov 14 at 15:24
  • 14
    Somehow it's even worse than discussions which I didn't think was possible. Commented Nov 14 at 15:25
  • 5
    I wanted to count how many normal Q&A have been posted since the beginning of this experiment and how many new-experimental Qs have been posted to see how much damage has been done potentially (especially if the number of normal Q&As would go down for the amount that new-experimental Qs go up). Is there a way to differentiate between them in SEDE? Commented Nov 16 at 11:28
  • 6
    despite the fact that it is actively causing harm by misdirecting questions that ought to be traditional Q&A into the opinion-based stream, without even a way to redirect. Yes. And, the reverse should be true. For a traditional Q&A, instead of closing the question as "opinion-based", should that automatically be converted to an "opinion based question" (or an option under the "close reason" dialog)? Commented Nov 16 at 23:21
  • 2
    @Hoid 70% my what measure? non-deleted post count? post count? character count? what content are you comparing it against? just posts? posts and non-deleted comments? posts and all comments? posts and all comments and chat messages? Commented Nov 18 at 8:48

2 Answers 2

Reset to default
-15

Kevin B is correct, this will not be turned off any time soon.

But we are working on an update of what is coming next for the experiment, what we are seeing, what we are working on, additional improvements, etc.

11
  • 12
    Thank you. I am disappointed, but not surprised. This inspires me to observe that SO apparently has two very different kinds of things that they presently call "experiments". "We want to expose a feature preview" is not at all the same as "we want to test a hypothesis" or "we want to compare alternatives". I would much appreciate these being differentiated in the future -- at least in prose, but preferably also in mechanisms for opt out. And of course, any feature needs to have ongoing support, not the "we're only monitoring for a couple of weeks" purported for this experiment. Commented Nov 17 at 22:13
  • 1
    @JohnBollinger I definitely wouldn't call this a feature preview. In fact, there is somewhat of a chance that our next update will change the question type labeling as currently constructed. Commented Nov 17 at 22:22
  • 1
    As for support, we have not stopped fixing bugs, and the opt-out, with a few exceptions, should now be working correctly. Commented Nov 17 at 22:23
  • 7
    With respect to support, I think the issue here is messaging. The announcement of opinion-based questions gave (me) the impression that channels for support for the experiment were limited in both time and location. That would have been tolerable for a limited duration experiment, but of course it's not for a permanent feature. One of the outcomes of this Q&A is that I take SO to be saying that the opinion-based question feature is in fact being supported via the same channels as any other SO feature, Commented Nov 17 at 22:36
  • 4
    As for opt out, I think you misunderstand. The decision process for whether to participate in A/B experiments and similar limited-duration tests is different, and may have a different outcome, than the decision process and outcome for exposure to bleeding edge features, however you want to label that. And for opinion-based questions in particular, I should like to be free of them even after the feature is no longer considered experimental. Commented Nov 17 at 22:41
  • @JohnBollinger my understanding is that will still be possible, through filtering. but you'll never be able to hide them from outside web searches unless they stopped existing (and i doubt you'll be able to get stacks "ai assist" to omit them from it's results. Commented Nov 17 at 22:53
  • 14
    "...this will not be turned off any time soon..." I think it should not be called experiment then. Feature is the better matching term. Commented Nov 18 at 8:01
  • 1
    @Hoid The opt-out improvements, are those shared anywhere? Still seeing those opinion based questions in my feeds. Commented Nov 18 at 11:31
  • 2
    @0stone0 Where did you see one? Was it via search with a tag? That is the only spot they are leaking through to my knowledge. Commented Nov 18 at 14:49
  • @Hoid Yep, saw them on the js tag. So I'll assume thats one last bug to fix ;) Thanks! Commented Nov 18 at 15:07
  • 1
    @Hoid, you mention the opt-out should be working with a few exceptions? What are the situations where it does actually work? Because I am still seeing them every where I was before, despite being opted out. Commented Nov 20 at 11:38
20

Just like past experiments surrounding discussions, there is no specific end date for this experiment. Given it is a place where they are allowing users (both new and old) to create content, it's unlikely it will just be turned off any time soon.

The previous experiment, discussions which was similarly launched in an "alpha" state without many of the features you'd generally expect, remained in that state for 2 years without receiving much of anything in terms of improvements/enhancements before being retired when challenges was released.

I find it unlikely that this will be turned off any time soon; low quality posts with useless replies is the goal, not a bug.

7
  • 14
    You may be exactly right, but I hope (perhaps in vain) to hear an authoritative answer from a staff member. SO seems to be on a crusade to teach me to expect less of them. So far, though, hope springs eternal. Commented Nov 14 at 21:22
  • 9
    I generally feel that, by providing an answer, it puts the onus on stack to come forward and provide a response if what my answer says isn't agreeable with them. If they don't... then we can only assume it is agreeable. :shrug: Commented Nov 14 at 22:28
  • 1
    @KevinB you basing your reasoning on hypothesis that they do care, do long there was no evidence of that. Commented Nov 16 at 12:52
  • 1
    @KevinB How did you know this one weird trick to make me answer. Commented Nov 17 at 20:52
  • 1
    @talex i know they do care, it's just how they've decided to go about addressing the problems this network has is what i disagree with. Commented Nov 17 at 22:56
  • @KevinB Thing is, you aren't only person who disagree with their approach. As far as I can see almost no one support them. I see it as evidence of them not caring about it (if they care they'd at least would explain themselves). Other option is that they don't understand what they are doing. I don't want to thing about third option. Commented Nov 18 at 7:01
  • @talex Third option is that they simply disagree. And fourth option would be that it's a mixture of different options. They probably don't care as much as we would like, maybe also do not understand completely what they are doing and might simply disagree with us and instead focus on those who use that feature. There are some people who use that feature. I guess that in the end we have no more right to SO than anyone else. Commented Nov 20 at 11:14

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.