Marjolein Katsma's dashboard

Biz actually updated his original post with a reference to the Confession account: "Note that this same popular blogger issued another complaint more recently about a different account which was some sort of "confession" mashup that allowed multiple people to post to one Twitter account. Our support staff investigated as soon as we received the complaint but the account had already been deleted by it's creator."
What bothers me the most about this, is that this is a LIE. Numerous people have actually commented on the contents of the confession account, and in fact it still existed until *yesterday* - it was not actually closed/emptied until I pointed out in my (moderated-away) comment yesterday that it was still there for all to read.
Sorry, Twitter, that's not "investigation" - that's sloppy misinterpretation from kosso's comment that the confess /service/ was closed, if you even paid any notice at all.
So Biz is either quite naively, or quite deliberately lying. In either case, that's an excellent way to win trust of your users back, isn't it? No, it isn't. – Marjolein Katsma, on May 26, 2008 20:58

Well, sure enough my comment wasn't let through - but confessions now is empty... so maybe it did some good after all.

Maybe next time Twitter can actually check before saying "account was closed"? ;) – Marjolein Katsma, on May 26, 2008 14:53

...where biz comments: "Folks, I may have to reject comments referencing the confession account. That account was deleted and it was not the account that inspired the accusation."
Meanwhile the confessions account DOES STILL EXIST and all the disgusting content is still there for everyone to see. I posted a comment to that effect on the blog, but it will probably be moderated out... – Marjolein Katsma, on May 25, 2008 19:10

Then how come the *account* still exists? It looks like you closed the *service*, not the Twitter account. – Marjolein Katsma, on May 25, 2008 18:21

If there was an account that was closed in March, it's not the same account as this: http://twitter.com/confession - the anonymous posting service has been closed, but this Twitter account still exists, including all its content, for everyone to see. – Marjolein Katsma, on May 25, 2008 18:10

Possibly there was "an account" that was closed in March. If so, there clearly are (were) two sources of harrassment and slander: the @confession /service/ has been closed, but the @confession Twitter account still exists, including clear evidence what has been happening. The slander is still there.
Why on earth Twitter do you not close that account? Stop dragging your feet and uphold your *current* TOS (never mind what any future one would be) - you have a TOS now and have a moral obligation to remove this account and all its content. – Marjolein Katsma, on May 25, 2008 18:05

"Essentially, Twitter is a communication utility, not a mediator of content."
Clearly, this is simply not true. It would be true only if everyone's tweets would be private, and be able to be seen only by those who are explicitly allowed. The mere fact that by default tweets are public /does/ make Twitter a publication platform, not merely a communication platform. If someone is slandered /publicly/ then all the tools Twitter gives to a person to block at person's content from being seen, will only apply to the person who's slandered - everyone else will continue to see it. Twitter hiding behind "we're only a communication platform" is disingenuous. While I agree we're talking about content that can no longer be seen /here/ there are caches and search engines out there - and that's not a reason *not* do do anything about it, it's a reason to take swift action to send the message that this is not tolerated. If other services can do it, so can Twitter.

When I signed up, I agreed to the TOS - guess why? Because I actually *agreed* with it. As long as it is in operation, you should stand by it and enforce it. But if you change it, I, and a lot of other people, will likely no longer agree. Providing a "communication" platform may have been the original intention - but that is clearly not what Twitter is now. Accept it and take your responsibility - don't go and hide by wishful thinking about what you want the service to be: as long as most people's tweets are public, it's a *publication* platform - you made it, you should accept the consequences of that. Changing your TOS is entirely the wrong thing to do, and sending a very negative message. – Marjolein Katsma, on May 24, 2008 09:17

I'm one of those 'all @ replies' people because that - like track - is an incredibly efficient way of finding new people with the same interests. – Marjolein Katsma, on May 13, 2008 13:56

I really don't care who sees my content - everyone's welcome to read that. Even spammers.

"Block" I use only to avoid seeing *other* people's content - and I apply that only to spammers. And be cause it applies only to spammers, I need it to work in track as well. "Unfollow" doesn't even come in to it because I wasn't following the spammers in the first place.

If I block a spammer and that prevents /them/ from seeing my content, then that means to me block is broken. "Block" should apply only to whose content you don't want to see "passively": no replies, no direct messages, no track. (By "passively" I mean receiving their content without making the choice to go to their page.) "Block" is for those cases where "unfollow" just does not apply.

If I wanted my tweets to be private I'd set it to private - but I don't. "Block" is not for hiding your own content (or should not be).

But Twitter needs to be protected from spammers though, and the sooner the better. The sooner that happens effectively, the less need there will be for blocking.

*sigh* onlineschool is still spewing the same tweets (text apparently scraped from RSS feeds!).

And no word from squidoo yet.

1. and != or
2. freedom of speech != right to be heard
3. on someone else's service, their TOS is "law", and freedom of speech does not apply; if you don't want to be bound by their rules, run your own server. – Marjolein Katsma, on May 12, 2008 11:14

Yes, spammers do do that, there are even systems that actually OCR CAPTCHA's (which the more "sophisticated" they become only manage to keep out more real humans more efficiently - CAPTCHA has always been a dead-end keeping out humans quite efficiently, too). Looking at "multiple accounts from the same IP address" will of course also not work - spammers so far are only testing the waters here, it seems, but they'll use all the devices already used in other contexts, including botnets of zombied machines to have an endlessly varying pool of IP addresses to post from.
Dealing with spammer is always an arms race - it has begun, you can't stop that. - All you can try to do is anticipate by looking at their techniques in other contexts, as well as successful anti-spam (but not anti-human) techniques. – Marjolein Katsma, on May 11, 2008 04:56

What will the spammers do in response? My point about having two ways to vote actually addresses that (think of spammers taking "revenge" for having their accounts marked as spam). It is a concern, but it goes both ways, also giving real humans a way to counteract spammers' revenge. – Marjolein Katsma, on May 11, 2008 04:47

Like others, I'm worried about a completely automated mechanism where an account would be removed when a "sufficient" number of "spam" votes has been received. A few thoughts:

* In the web interface provide TWO buttons: one "spam" and one "not spam" so people can "vote" either way. When the "spam" vote is higher by a certain amount or percentage than "not spam", FLAG that member's updates so others can jump in to either confirm or vote "not spam". One person one vote, you could even change your mind by voting the other way, but you still get one vote. Then human review by Twitter after some period, if the "spam" vote is still (considerably) higher than the "not spam" vote - otherwise just "expire" the voting and flagging. This would go some way towards preventing someone being removed as spammer merely because a bunch of people dislike the member, and might also reduce the amount of effort needed for review by/on behalf of Twitter.

* For the IM/phone interface provide two commands that are equivalent in action to the "spam"/"not spam" buttons on the web interface.

* The IM/Phone interface would also need a command to block-in-track so that a "keyword spammer" won't get through anymore (regardless of whether the blocked account is/will be seen officially as a spammer)

I realise that "block in track" as an option is a separate discussion, but it ties in closely, so I'm mentioning it here merely as a reference.

By allowing voting both ways with some checks and balances in trigger levels and elapsed time, followed up with evaluation, I think all-too-hasty banning of what looks like a spammer (or someone who is merely disliked by a group) would be prevented. It wouldn't be fast, but I think that's a good thing. With many "not spam" votes, a review might not be necessary; but for a review a suspected spammer might be given a chance to defend their case (unless it's too blatantly obvious, maybe).

Ha! Squidoo have an anti-spam policy and a report form. Needless to say I used that form.