-9

I just tried to answer a question, but sadly the editor didn't let me submit my answer. After searching here for the issue I found out that there were some other people who had the same problem. Sadly this didn't result in correcting the feature but in correcting the people...

If changing or improving the detection algorithm to prevent triggering the message in cases like these is not an option, why not change to message to a type which the user can ignore if he is sure that there is no code in his answer?

8
  • 6
    Any chance you could show us what you tried to submit that caused the message to appear?
    – Bart
    May 6, 2013 at 18:42
  • It was the answer to this question stackoverflow.com/questions/16310552 and thx to this meta-post I seem to have gained some minus votes... (the question is some days old but didn't receive a minus vote before...)
    – JepZ
    May 6, 2013 at 19:17
  • I was hoping you could share a "before" version. If anything it might have helped identify what the filter choked on and perhaps that could be addressed.
    – Bart
    May 6, 2013 at 19:21
  • the before version didn't have the bullets in front of the links
    – JepZ
    May 6, 2013 at 19:22
  • @Bart btw. when I posted this I thought the links provided enough examples to find examples to improve the algorithm
    – JepZ
    May 6, 2013 at 19:28
  • 1
    Listen, I'm trying to constructively address the issue you faced by trying to figure out what your issue was. Especially when you yourself say that all of your the links in this question had other solutions than the one you applied, it's good to know that the situation was.
    – Bart
    May 6, 2013 at 19:30
  • @Bart thx, sorry I'm just a little frustrated. This was the original message: pastebin.com/DQ1FUttg
    – JepZ
    May 6, 2013 at 19:34
  • I'd say it simply shouldn't have choked on that. Perhaps your "arrow" was interpreted to be code-like. In any case, if your feature request doesn't make it, this case might well serve as some good information for a little tweak.
    – Bart
    May 6, 2013 at 19:37

1 Answer 1

12

If changing or improving the detection algorithm to prevent triggering the message in cases like these is not an option, why not change to message to a type which the user can ignore if he is sure that there is no code in his answer?

Because then there would be thousands of people posting code without properly formatting it. Many people won't even read such an error message without clicking through, and of those that do, many will still click through rather than taking the time to fix the problem (because in most cases they really will have code not properly formatted).

You've found half a dozen people having problems over the course of two years, yet there are dozens of people posting code without formatting it every day. That's a pretty decent margin of error in my mind. I think that is a small enough false positive rate that it can be handled as it is currently. If this becomes a more widespread problem and the filter itself cannot be improved then the issue can be re-addressed.

21
  • and why not changing the message to somthing meaningful? I finally fixed it by putting bulletpoints in front of my links but all the other had problems with whitespaces (missing linebreak, too many linebreaks)...
    – JepZ
    May 6, 2013 at 19:00
  • I mostly agree with this, but if this is not worth a fix, a dev should at least explain why it happens and how to avoid or work around the issue.
    – user200500
    May 6, 2013 at 19:00
  • 1
    @JepZ What do you propose changing the wording too? Keep in mind the idea is that your post managed to hit a filter indicating there was code not properly formatted as code. It's possible that it was indeed a false positive and you didn't have unformatted code, in which case a small change that seemingly wouldn't be relevant could change the post enough to not hit the filter, but I don't see how the error message should be changed as a result of this. If it was known to be a false positive it just wouldn't filter it at all.
    – Servy
    May 6, 2013 at 19:03
  • btw. I'm not sure if your statistic is right, because you don't know how many people failed to answer a question and didn't know about Meta or weren't alllowed to post here
    – JepZ
    May 6, 2013 at 19:04
  • @JepZ While that's true, I still don't see this as a serious issue as there really are so few people who have brought it up here. I of course realize that it doesn't represent everyone with this problem.
    – Servy
    May 6, 2013 at 19:06
  • Btw. nobody came back on fixing the orginal issue... funny. Instead we discuss about statistics which are wrong or don't exist. Most web-developers fight for every single new customer. Not so at SO. Here are new people not so rare, since there are enough experienced ones... I know that it is important to have high-quality answers but wasn't that exactly the reason to introduce the upvoting system? to upvote the good answers an let the bad ones down?
    – JepZ
    May 6, 2013 at 19:15
  • @JepZ The model that SO follows, and the reason it has been so successful, is that SO does everything it can to get high quality answerers, and it does what it needs to in order to keep them happy and to allow them to post quality answers to questions that they are likely to find interesting. It does not go out of its way to assist people asking questions at the expense of the experts who answer them. It is under the assumption that people who have a problem they want solved for free by an expert will jump through the hoops they have to to get their answer. This has proven correct.
    – Servy
    May 6, 2013 at 19:18
  • @JepZ You'll notice the sites that don't apply restrictions or expectations on those asking questions, instead leaving those who answer them to deal with all of the problems that they create, have been virtually starved to death by SO. The true experts don't go to those places anymore, they go to SO, so anyone who wants a solution to their problem also goes to SO, because that's where the experts are.
    – Servy
    May 6, 2013 at 19:20
  • So many experts but the newbe who points out a weakness in the system and makes a (wrong) suggestion to improve the system gets away with -4 point (until now)...
    – JepZ
    May 6, 2013 at 19:29
  • @JepZ meta.stackoverflow.com/faq#vote-differences
    – Bart
    May 6, 2013 at 19:31
  • @JepZ Those people seem to feel that the current filter adds more benefit than the problems it causes, and voted accordingly. Note that votes are a little different on meta.
    – Servy
    May 6, 2013 at 19:31
  • @Servy So many experts... they downvoted my original question and didn't leave a comment why. So I dont even know what to improve. The poor timing was that they downvoted (twice) right after I posted my suggestion here...
    – JepZ
    May 6, 2013 at 19:52
  • 1
    @JepZ, I think it was probably just a down-vote spree sort of mentality. A lot of times someone will post something on meta, and if their meta post gets downvoted, people will sometimes without thinking downvote the example question as well (and similarly, some people that upvote meta questions will non-thinkingly upvote the linked example). Personally, I thought your question and answer were both good, so I upvoted them both. And as voting on meta works differently, I downvoted this meta question, because I disagree with the proposal, not because I think it was a bad question.
    – Ben Lee
    May 7, 2013 at 6:49
  • 1
    @trlkly You've just proved my point for me. You had a poorly formatted post and your immediate reaction was, "how do I subvert this message and submit my post anyway" rather than, "how do I fix the problem with my post so that it no longer has this problem". Clearly you'd have just submitted an improperly formatted post anyway if you were given the option to do so, and only contemplated fixing it when you didn't have such an option.
    – Servy
    May 3, 2017 at 13:12
  • 1
    @trlkly Assuming it's this post that's just not true. You're trying to format your code as not code, rather than formatting your code as code. If you format your code as code, as I've done, everything works perfectly fine. So indeed you had an improperly formatted post in which you had code that wasn't formatted as code, and all you did was search for ways to subvert the check, rather than actually formatting your code as code.
    – Servy
    May 9, 2017 at 13:25

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