Scripting News - Latest Comments in How spam will likely enter the Twitter community (Scripting News)

Re: How spam will likely enter the Twitter community (Scripting News)

.LAG — Wed, 13 May 2009 12:01:46 -0000

...your blog is a time machine man!

Re: How spam will likely enter the Twitter community (Scripting News)

RacerRick — Wed, 19 Dec 2007 18:57:23 -0000

You should only be able to receive @ replies from folks that you're following. That would make it a little more like instant messaging.

Re: How spam will likely enter the Twitter community (Scripting News)

yndy — Mon, 17 Dec 2007 23:43:43 -0000

Except that Twitter has already given users the option to only see @username entries from people who are listed on their own follows list.

They anticipated it and already fixed it.

Re: How spam will likely enter the Twitter community (Scripting News)

Dr.Mani — Mon, 17 Dec 2007 22:24:51 -0000

Dave, Twitter does have the feature to turn off @ replies.

Settings > @Replies > no @ replies

Poof! G'bye spam.

Dr.Mani
(drmani on Twitter)

Re: How spam will likely enter the Twitter community (Scripting News)

mdy — Mon, 17 Dec 2007 20:42:05 -0000

Apparently, if an account is blocked by a large number of users, it gets flagged and may be suspended.

See Why Twitter Accounts Get Suspended

So community blocking seems to be one of the ways we can help keep Twitter spam free.

Re: How spam will likely enter the Twitter community (Scripting News)

mdoeff — Mon, 17 Dec 2007 20:31:48 -0000

Sssshhhhh! Don't give the spammers any ideas!

Re: How spam will likely enter the Twitter community (Scripting News)

Jeremy Franklin — Mon, 17 Dec 2007 19:18:55 -0000

You can't receive Twitter spam when Twitter isn't working which seems like all the time.

Re: How spam will likely enter the Twitter community (Scripting News)

heavyboots — Mon, 17 Dec 2007 17:40:49 -0000

Captcha. Plus requiring an account to reply. Plus a slowdown if xx number of your replies are reported as spam, so that perhaps only one reply every yy minutes after that is allowed from the "possible spam" account.

Re: How spam will likely enter the Twitter community (Scripting News)

pwfenton — Mon, 17 Dec 2007 17:38:18 -0000

You say "You can direct a message with a url to anyone as long as you know their username." I can direct a message to anyone... however, that doesn't mean they will receive it. Send me a message from someone I am not following... my ID is "P_Dub". I don't think you can.

Re: How spam will likely enter the Twitter community (Scripting News)

pwfenton — Mon, 17 Dec 2007 17:22:09 -0000

I don't understand what you are suggesting. I just created a new Twitter account and tried to send both a "reply" and a "direct message" to my normal Twitter ID... neither one went through because I am not "following" the new account I just created. So how does someone that I am not following send me SPAM?

Re: How spam will likely enter the Twitter community (Scripting News)

JG — Mon, 17 Dec 2007 16:56:18 -0000

This is another reason why they should have rolled their own URL shortener. It would be a lot easier to kill off spam links if they were all going through your own system. MySpace started doing that a while back and would have avoided a lot of headache if they did it from the start.

Re: How spam will likely enter the Twitter community (Scripting News)

Cameron Watters — Mon, 17 Dec 2007 16:46:11 -0000

Whitelist for replies: check! -- I can currently limit replies to only those people I'm following. Also, I can go w/ private updates.

Blacklist for people I don't like: sortof check. -- I can prevent other users from following me and/or seeing my updates on a per-person basis. This is a race, though, that I will always lose.

What's needed? Community gestapo: the ability to report user accounts and or IP addresses. If we can collectively blacklist accounts, we may have a fighting chance against the spam.

Also, rate/message size limiting helps. Email is designed to make spam trivial. You address a single message to infinite recipients. In twitter, you're basically limited to one (or a few maybe) per message since the addressing mechanism is a part of the message, which is size constrained.

In order to effectively spam twitter (remember, spammers make money on volume), you'd have to send tens of thousands of replies per day.

The final option is to create a some kind of filtering based on rules. Message lengths are super short, so bayesian techniques may or may not be effective. However, you could create rules to filter/flag responses likely to be spam (e.g. flag messages whose entire content is a URL).

Re: How spam will likely enter the Twitter community (Scripting News)

kosso — Mon, 17 Dec 2007 16:35:16 -0000

You're absolutely right - I think it has already started. There appears to be many accounts out there which are 'harvesting' people to follow. No doubt using these 140 character sentences to build a keyword profile on you. (Rather like the non-official Twitter search engines work)

It would be very easy to then spam this through a 'TwitterBot' account.

There are couple of things you can do:

They recently added a new setting in your notices section of your setting page on Tiwtter to control the @replies feature : see http://twitter.com/account/... (while logged in to twitter)

Also, on receiving a 'twitter spam' post - or what I like to call a 'SPITTER' - you can 'block' that user account using the existing feature on Twitter.

I hope that we'll soon have a feature in Twitter to report 'Spitters' :)

Cheers!
Kosso

ps: check out some seesmic rss action @ http://imadethis.tv :D

Re: How spam will likely enter the Twitter community (Scripting News)

Bill Seaver — Mon, 17 Dec 2007 16:35:11 -0000

Great thought and I could certainly see this happen. I have also noticed that sometimes people have long running conversations in Twitter. Why can't they just take it to a medium that everyone else isn't watching. It's like conversation spam because they dominate my Twitter stream. After two or three replies it's time to take it off-Twitter and pick up the phone or email.

Re: How spam will likely enter the Twitter community (Scripting News)

Robert 'Groby' Blum — Mon, 17 Dec 2007 16:26:40 -0000

Spam has already entered twitter. Currently, it's in the form of random "users" subscribing to your twitter feed. (Yes, I consider this spam. If a political candidate subscribes to me, I'm fairly certain she's not interested in my tweets)

Re: How spam will likely enter the Twitter community (Scripting News)

Michael Markman (Mickeleh) — Mon, 17 Dec 2007 16:26:19 -0000

per your example "tinyurl" is ironic? taunting? descriptive?

Re: How spam will likely enter the Twitter community (Scripting News)

Lukas — Mon, 17 Dec 2007 16:24:28 -0000

Simple: elminate direct messages from anyone I'm not following.