Stories
Slash Boxes
Comments

SoylentNews is people

Meta
posted by janrinok on Sunday July 03 2022, @03:30PM   Printer-friendly
from the lets-see-what-happens-here dept.

I realise that this has been an unpleasant time for many of our anonymous community members, but I can assure you that it has been necessary. I am not yet prepared to go into details but I can at least update you with our findings so far. But first we have to look at some historical data.

Anonymous Cowards (ACs) have always been - and will hopefully continue to be - welcome members of our community. There are many perfectly understandable reasons for wishing to post as AC and how you chose to live your own personal life is of no concern of this site. Equally, you are welcome to use VPNs and other security measures to protect your privacy. We take similar measures to protect all of your data so that you will not be compromised by us. These measures are effective and to the SN administration ACs appear as a single user with the user identity of #1.

We cannot treat some ACs differently from others. While we can manage to sort out your comments etc with the aid of the hashes that we produce, they change so frequently as to be useless for any purpose outside of this site. But the Administration is only concerned with what happens within this site and so this point is moot. We have no interest in the rest of the internet so IP addresses are also of no interest to us. How your comments get from wherever you are to us is irrelevant. The bottom line is that ACs can only be treated as a single account. That account is granted certain permissions or not granted those permissions and they apply to every AC interaction.

Most of our community, both logged in and AC, participate in the discussions in an reasonable manner and discuss the topic that has been outlined and any threads that resulting from it. It is true that, particularly at weekends, there is a slight increase in the number of ACs appearing but on their own they are little more than a minor irritant. There is, however, a 3rd group, consisting of ACs who sole purpose seems to be to derail any sensible discussion. Over recent years they have become more aggressive and often use personal attacks rather than challenging what is being said. Some are more obvious than others and I am sure that you can all think of examples of such people for yourself. A very small number have stated that it is their aim to prevent SoylentNews from continuing.

On 22 Jun of this year we received an implied threat (https://soylentnews.org/comments.pl?noupdate=1&sid=49894&page=1&cid=1254201) suggesting the the person making it had a target date of 6 July for some event or other. It is possible that this is related to another 'prophecy' in which this individual foretold that the site would soon be dead. We believe that we can identify the person making that threat with a reasonable degree of certainty. However, since that time the number of ad-hominem attacks has increased and we have also been subjected to increasing amounts of spam. In small amounts either or both of these things can be shrugged of, but when they come increasingly aggressive and frequent, they can make the entire experience of being in this community very unpleasant. I know that we have lost both staff and numerous community members because of this toxic atmosphere - and not, as some would have you believe, because we administer the site!

Almost all of this behaviour is conducted by a very small number of ACs and occasionally via sock-puppet accounts. As the levels of harassment increased over the last few weeks it was obvious to us that we could remove it by simply preventing AC access. This was not an easy decision to make but we knew that we could protect the majority of the site by this simple action. The result is, as you know, that we reluctantly removed anonymous access by ACs to the front page.

We are now actively looking for more permanent solutions and hopefully to exactly what we had before. I have experimented with providing stories on the front page which are AC friendly, and also in my journal. We are still looking for a better solution but unless we can separate individual ACs then I cannot see what else can be done. I would welcome your feedback and suggestions. The outcome of our decision is also our loss as you can see if you look at the numbers of comments that we are now getting compared to before the ban.

I have spent a lot of time analysing the posts, both current and historical, to try to identify the person or persons responsible for this unwanted content. I am not going to name specific individuals because I believe that you can each reach your own conclusions. By looking at both the spam and comment content, and their meta data, I have established the following.

The person spamming our site is one of our own Anonymous Cowards who is currently blocked because we have removed access for the AC account - and that block affects all ACs. He is also one of the people regularly carrying out ad hominem attacks against other community members. He will be reading everything that we post about this issue.

Unless the abuses cease everywhere on the site including in journals, ACs will remain outside of the main site except for specially released stories until we can devise a better system. For us to currently do anything different would be foolish and irresponsible in the extreme. As soon as the abuse ceases we can readmit all ACs to the main site again.

I know that this will be as much of a disappointment to you as it is to me, and you may also be thinking of leaving. I ask you not to go. Rather I would encourage you all to let the abusers know that they are not fighting for your freedom of speech ("freeze peach") but they are by their actions actively preventing your participation in our site. There is one particular post (https://soylentnews.org/comments.pl?noupdate=1&sid=50204&page=1&cid=1257692) which suggests that this is being done on behalf of all ACs and that you all stand as one behind this action. I don't believe that anybody has the right to claim that if you haven't actually agreed to it.

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 4, Interesting) by Anonymous Coward on Sunday July 03 2022, @04:15PM (27 children)

    by Anonymous Coward on Sunday July 03 2022, @04:15PM (#1257775)

    The spammers definitely do NOT speak for all of us.

    Maybe I'll code up some anti-spammer functionality. I can work in perl (been a few years since any professional work in it, mind you). Some ideas spring to mind.

    If I can spare enough hours, that is. So, middling odds.

    Looking forward to being able to be a full-fledged participant in the community and in the conversations again.

    Signed,
    A Friendly Neighbourhood Well-intentioned AC

    Starting Score:    0  points
    Moderation   +4  
       Interesting=3, Informative=1, Total=4
    Extra 'Interesting' Modifier   0  

    Total Score:   4  
  • (Score: 4, Informative) by janrinok on Sunday July 03 2022, @04:22PM (22 children)

    by janrinok (52) Subscriber Badge on Sunday July 03 2022, @04:22PM (#1257776) Journal

    I am grateful for your offer, but you must remember that we will not be giving anyone access to our hardware systems until we know who they are and their competencies. You can download our source code from Github [github.cosoylentnews]. The code you are looking for is Rehash.

    If you wish to continue with your offer please contact the Admins by email.

    --
    I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
    • (Score: 4, Informative) by janrinok on Sunday July 03 2022, @04:37PM

      by janrinok (52) Subscriber Badge on Sunday July 03 2022, @04:37PM (#1257782) Journal

      link is corrupted - https://github.com/soylentnews [github.com]

      Apologies - more haste and less speed.

      --
      I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
    • (Score: 2) by RS3 on Sunday July 03 2022, @04:52PM (20 children)

      by RS3 (6367) on Sunday July 03 2022, @04:52PM (#1257787)

      As a sysadmin (not here), if I may, I suggest you allow only SN admins access to hardware. I believe there's at least 1 test / development system. Be sure to use proper git checkout, commits, etc., and please please get code reviewers- many coders here- before committing to production system.

      All that said, the beauty of interpreted code like perl is that it's pretty painless to restore working code. Worst-case is database corruption where you may lose some posts, which is a bad thing of course. (hence, test much on development system)

      BTW, If _I_ were sysadmin here, I would just delete the text of the spam comments. Somewhere you or someone mentioned link / brokenness if you try to delete an entire post's database record. I'd get a consensus from some top-tier people here, but using some simple database tools, it's pretty easy to edit the post text field and put something pleasant.

      • (Score: 5, Interesting) by janrinok on Sunday July 03 2022, @05:02PM (5 children)

        by janrinok (52) Subscriber Badge on Sunday July 03 2022, @05:02PM (#1257792) Journal

        We do all that you say and more. Nobody gets access unless they have been accepted by the current sysadmins. I haven't got access. As far as I know the access is kept to only a handful of people.

        Plus, of course, that our hardware is hosted elsewhere and daily backups are automatic.

        I can understand your logic behind deleting Spam content but I am rather glad that we don't do that. It has proven to be very useful over the last few days. People make some unusual spelling mistakes which can be found elsewhere on our site..... The little things can often be enough to give me a pointer to where to look next.

        --
        I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
        • (Score: 2) by RS3 on Sunday July 03 2022, @05:32PM (4 children)

          by RS3 (6367) on Sunday July 03 2022, @05:32PM (#1257801)

          Thanks. Yes, that's why I said I'd get consensus before deleting or editing a post.

          • (Score: 2) by Reziac on Monday July 04 2022, @02:43AM (3 children)

            by Reziac (2489) on Monday July 04 2022, @02:43AM (#1257921) Homepage

            There's also the problem that if spam autodeletes (as it would have to, lest the admins find themselves doing nothing but spam herding) then it becomes easy to target and erase individuals one disagrees with.

            --
            And there is no Alkibiades to come back and save us from ourselves.
            • (Score: 4, Informative) by janrinok on Monday July 04 2022, @05:06AM (2 children)

              by janrinok (52) Subscriber Badge on Monday July 04 2022, @05:06AM (#1257948) Journal

              lest the admins find themselves doing nothing but spam herding

              A situation I found myself in a few days back with 101 spam posts in less than 16 hours if I remember correctly. And they cannot be done at all once - they really should be supressed at the earliest opportunity so I had to respond to each of them as and when they arrived.

              However, I have a few things to help me. I can watch all comments arriving (for all stories/journals simultaneously) in a feed which lets me quickly identify them and then take action. I can also immediately see which posts are being moderated and how. Many of our community do assist by Spam modding as soon as it arrives which draws my attention to the activity. We also have displays for moderation abuse etc. I don't want to have to sit in front of my screen waiting for comments to arrive for 12-16 hours a day! I think that only a handful of people currently have those privileges - but there are not many staff left now either. The toxic community has taken its toll.

              --
              I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
              • (Score: 5, Insightful) by Reziac on Monday July 04 2022, @05:45AM (1 child)

                by Reziac (2489) on Monday July 04 2022, @05:45AM (#1257952) Homepage

                I am thankful for your dedication. Good job.

                --
                And there is no Alkibiades to come back and save us from ourselves.
                • (Score: 4, Insightful) by janrinok on Monday July 04 2022, @06:56AM

                  by janrinok (52) Subscriber Badge on Monday July 04 2022, @06:56AM (#1257968) Journal

                  I appreciate the kind words, but I don't suppose I would still be doing it if it was no longer something pleasurable overall - these times are not much fun but over the 8 years I've certainly enjoyed the task, the community and the comments far more than a few low points along the way.

                  I will nevertheless be glad when we get this matter resolved.

                  --
                  I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
      • (Score: 5, Interesting) by Anonymous Coward on Sunday July 03 2022, @05:13PM (13 children)

        by Anonymous Coward on Sunday July 03 2022, @05:13PM (#1257795)

        BTW, If _I_ were sysadmin here, I would just delete the text of the spam comments.

        I quite like the TechDirt approach where spam and whatnot just shows a single line (no title, no content) saying

        This comment has been flagged by the community. Click here to show it.

        so anyone can click it into view if they really want to. I feel it really makes a point that it's not hidden, but really, Dude, no-one wants to read it. All the spammer can do is to shout louder and louder that no-one is listening to him, which just gets his posts marked down faster and faster.

        Hmmm. I headed off to the TechDirt site to find an example to paste into the quote and there seems to be a shortage of such just recently. Then I scrolled back as far as a post mentioning Trump, which of course came up, well, trumps.

        • (Score: 3, Informative) by janrinok on Sunday July 03 2022, @05:43PM

          by janrinok (52) Subscriber Badge on Sunday July 03 2022, @05:43PM (#1257803) Journal

          I like the suggestion too - but without that elusive programmer it leaves us with a problem, and it is not something that I want to try to do manually with a live database.

          --
          I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
        • (Score: 4, Interesting) by RS3 on Sunday July 03 2022, @05:51PM (11 children)

          by RS3 (6367) on Sunday July 03 2022, @05:51PM (#1257804)

          That's a great idea. Of course that would involve code changes, and afaik SN does not have a full-on perl programmer.

          I like what I thought was the plan- anything marked "spam" has to be true spam or something equally garbage. Anyone abusing the spam mod would lose mod points for a month.

          If that was and is true, then at least one more mod score level might be the fix, and it might be very (very!) easy to expand the range of possible scores. It may be that a -2 for true spam / garbage would solve much of the problem, and only a "spam" mod could achieve a -2 score.

          Then (obviously) there needs to be an appeal / review system where if your post got a -2, you request a review by top-level editors' / admins' consensus, and if the "spam" mod was not deserved, it gets fixed and person doing undeserved mod loses mod points for a month or more.

          • (Score: 4, Interesting) by janrinok on Sunday July 03 2022, @06:03PM (10 children)

            by janrinok (52) Subscriber Badge on Sunday July 03 2022, @06:03PM (#1257807) Journal

            We review spam moderations frequently - although not as often as I would have liked over the last 36-48 hours. We probably revert about half of them but it does vary considerably and that is a wet finger held in the air estimate. If someone repeatedly abuses the spam moderation they receive a private warning and, should they continue, they can receive a moderation ban starting a 1 week for the first offence and so on...

            --
            I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
            • (Score: 3, Interesting) by JoeMerchant on Sunday July 03 2022, @09:07PM (9 children)

              by JoeMerchant (3937) on Sunday July 03 2022, @09:07PM (#1257841)

              One potential tool to slope the field against Trolls would be to magnify the troll moderation effect on AC posts.

              In the 1980s I established the axiom: a message board determined to allow anonymous posts is defenseless against griefers. Community moderation is the best mitigation to come along since then.

              If you want to go high tech you could try an AI lexical analyzer. Feed it a stream of all available data on bad vs good posts and allow it to slow the introduction of material posted by AC which the AI throws in the "smells bad" category.

              --
              🌻🌻 [google.com]
              • (Score: 3, Interesting) by Ken_g6 on Monday July 04 2022, @05:33AM (4 children)

                by Ken_g6 (3706) on Monday July 04 2022, @05:33AM (#1257950)

                Actually, a spam filter is something that could be done by a non-Perl programmer, if you have those to spare. Create a special account with lots of mod points. (Or, just a normal one, for testing.) Then, someone could create a bot, in any programming language, to go through posts, run them through a spam filter, and moderate them if needed. If it's a really good spam filter that ranks posts as "not spam", "kinda spammy", or "clearly spam", those could be mapped to no moderation, "Troll", and "Spam" respectively.

                • (Score: 3, Informative) by janrinok on Monday July 04 2022, @05:58AM (3 children)

                  by janrinok (52) Subscriber Badge on Monday July 04 2022, @05:58AM (#1257955) Journal

                  We can do that part - but we cannot integrate it into the comment processing or allocate it to the appropriate account without interfacing it with the existing Perl..... And any change to the Perl necessitates a new build, a new testing schedule, etc

                  It's chicken and egg I'm afraid, unless I have misunderstood your suggestion. If the latter is the case then feel free to hit me with a clue bat and explain it again please. I am tiring after a long weekend and I am perhaps not firing on all cylinders at the moment.

                  --
                  I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
                  • (Score: 3, Interesting) by coolgopher on Monday July 04 2022, @07:36AM (2 children)

                    by coolgopher (1157) on Monday July 04 2022, @07:36AM (#1257973)

                    I believe the idea (if I understood it correctly), was to have a bot which uses the existing site's interface as it's API, the only difference being that the account it logs in as has a considerable amount of available mod points. The bot could then read the comments and spam-check, and when triggered mod down automatically. So, no changes to the existing code base (assuming there is some way of granting mod points beyond the auto-refresh).

                    • (Score: 5, Funny) by nostyle on Monday July 04 2022, @03:00PM (1 child)

                      by nostyle (11497) on Monday July 04 2022, @03:00PM (#1258038) Journal

                      In short:

                      He is suggesting that we replace some of the moderators with a small shell-script.

                      • (Score: 2) by JoeMerchant on Monday July 04 2022, @09:56PM

                        by JoeMerchant (3937) on Monday July 04 2022, @09:56PM (#1258100)

                        Not replace, augment. And the shell script has the distinct advantage of tirelessly reading every comment posted - as soon as it is posted - and being somewhat unbiased in its mod assignments.

                        --
                        🌻🌻 [google.com]
              • (Score: 2) by janrinok on Monday July 04 2022, @11:04AM (3 children)

                by janrinok (52) Subscriber Badge on Monday July 04 2022, @11:04AM (#1257997) Journal

                You raise an idea worth thinking about. Thank you.

                One problem that we see on this site is that the more a specific moderation type suppresses a comment, the more that moderation type will be abused. If we simply make a troll mod carry more negative points than it currently has, then it will be used more often when it really shouldn't be. If everyone was perfect and their perception and use of moderation was the same we probably could see this working exactly has you stated. But I don't think that any of us actually meet that description.

                The end result is that someone has to moderate the moderations and ensure that they are being applied fairly across all stories and journals. The Admins do that to some extent already. We have displays showing moderation abuse and daily email reports showing who is receiving moderations that appear suspicious, whether they be positive or negative. As someone has to be logged in to moderate it is a simple task to look at who and what they are moderating and, if necessary, look at their moderation history. If it appears accidental or borderline then I usually issue a polite private admin-to-user message. I don't expect a reply from it, the problem resolves itself, and the world continues turning. If the abuse becomes excessive or repetitive, then a more formal warning is issued and/or a ban if thought necessary. The later case is still unusual although it is the one that you hear about because the recipient usually wants to complain about their treatment.

                --
                I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
                • (Score: 2) by Fnord666 on Monday July 04 2022, @09:42PM

                  by Fnord666 (652) on Monday July 04 2022, @09:42PM (#1258095) Homepage

                  The end result is that someone has to moderate the moderations and ensure that they are being applied fairly across all stories and journals. [...]

                  I believe that was what the green site did with meta-moderation. It allowed the community to police itself. Did that code not make it over?

                • (Score: 2) by JoeMerchant on Monday July 04 2022, @10:00PM (1 child)

                  by JoeMerchant (3937) on Monday July 04 2022, @10:00PM (#1258101)

                  For what it's worth, my suggestion was to have the magnified effect only on AC posts... but I suppose I am biased, here in my non-AC persona.

                  --
                  🌻🌻 [google.com]
                  • (Score: 2) by janrinok on Tuesday July 05 2022, @06:11AM

                    by janrinok (52) Subscriber Badge on Tuesday July 05 2022, @06:11AM (#1258167) Journal

                    Looking specifically at AC posts is a reasonable thing to do - particularly as our current woes are all as a result of the actions of a few of our own ACs. As long as the majority are not penalised for the action of a few then it might be a reasonable solution. There are already limits placed upon what ACs can do - they have additional security checks when they post, they cannot have journals, they cannot moderate etc. It becomes a balance between those who are prepared to have accounts and those who wish to remain as AC.

                    --
                    I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
  • (Score: 2) by JoeMerchant on Sunday July 03 2022, @08:50PM (3 children)

    by JoeMerchant (3937) on Sunday July 03 2022, @08:50PM (#1257834)

    You could start with recording a one way hash of the commenters' ip address, enabling blocking of AC access from address hashes matching abusive comments. It doesn't take much effort to vary a source IP address, but any effort is infinitely more than they currently expend.

    --
    🌻🌻 [google.com]
    • (Score: 3, Informative) by janrinok on Sunday July 03 2022, @09:22PM (2 children)

      by janrinok (52) Subscriber Badge on Sunday July 03 2022, @09:22PM (#1257849) Journal

      Er, we do have been doing that for 8 years. That is why people complain about us using hashes. We use hashes because we do not want the iPs. They then accuse us of tracking them. If has been Ari's favourite soap-box topic for several years.

      Many AC's IP addresses change very frequently. Some actually do it between every post.

      --
      I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
      • (Score: 2) by janrinok on Sunday July 03 2022, @09:24PM (1 child)

        by janrinok (52) Subscriber Badge on Sunday July 03 2022, @09:24PM (#1257850) Journal

        Sorry for the typos. It is 23:23 here and i have been on the site since about 08:30. Time for my bed.

        --
        I am not interested in knowing who people are or where they live. My interest starts and stops at our servers.
        • (Score: 2) by JoeMerchant on Sunday July 03 2022, @11:38PM

          by JoeMerchant (3937) on Sunday July 03 2022, @11:38PM (#1257883)

          zzzzzzzzzzz. So, you are tracking them, but not in any way useful to anyone else (as I think I saw you write elsewhere...) Moot, really, unless their ability to run multiple AC personalities on this particular site is of some kind of extreme value to them (in which case they're a very special snowflake, indeed.)

          Thanks for all your efforts, in the end it's just a public forum and nobody can really be held responsible for crackpots entering the conversation with all manner of crazy / deliberately offensive stuff.

          Would be interesting to see how accurately something like this: https://blog.revolutionanalytics.com/2016/12/stylometry.html [revolutionanalytics.com] could pre-screen for troll-text and slow down undesired types of content.

          --
          🌻🌻 [google.com]