I was going through my latest ‘followers’ and this popped up :

Hackers embed attack in Twitter profile image

I assume it was in one of the Britney type I blocked but what disturbs me is that a) it ever got through and b) it’s flagged at a known attack

If one got through you can be very sure that hundreds or probably thousands more bot accounts are similarly affected.

If you want an in depth look at this form of attack watch with 40 minute video DefCon 15 – T312 – The Executable image Exploit

You can read of the older GDI hack on this Cisco page : Microsoft GDI+ GIF Image Parsing Memory Corruption Vulnerability

There are other similar exploits and hacks. Scary eh. What’s irritating the heck out of me is Twitter let this through!

It was up around 10,040 when I got up this morning. Several hundred will be me spammers and bots, but that’s always the way.

Near as I can tell, @lettercult was 10,000th.

I’ll add graphs and more later. Need to sort new phones this morning!

For some time, the follower and following counts we display have been incorrect for some folks. We’re soon to push a change that will address this issue. This means that the count you see in your sidebar should match what you see on your follower and following pages.

However, a consequence of this change is that follower counts will drop for some people. In particular, those with large followings may see significant changes as we correct for spam accounts and data inconsistencies. No legitimate followings should be affected—we’re just cleaning up artifacts in the system.

@Mashable took at look at this here too: Twitter Correcting Follower Counts: 1000s of Spammers Perish. Pretty much he repeats Twitters own message, as above. What’s of more interest on his blog entry are all the replies – over 500 when I looked, including my own, copied below:

I have three accounts, one for close friends, a main bulk one and a new one I’ve been specifally adding spammers too…

My friends one… No losses at all

On my Whale account… I lost 370 bots out of over 10,000
(For this one I but lost few if any ‘following’ – I vet everyone I add, yes all!)

However, for my “spam” account* which I just started to look for bots – it went UP 400 for ‘friend’ and UP 700 for following (even after the purge)
*(This is largely manual but automated overnight to overnight)

Seems to me the bots are winning !

*(while no doubt ignoring the untold millions of spammers!)

OYE! @Twitter, what you playing at at all – so far at least two of my friends have been suspended now! Give us back @trinawright !

rt @AndrewTF @piginthepoke There are a few accounts that are being suspended, including the social networking correspondant for the Times!!

OYE! @Twitter. Why have you gone and suspended @shandreen Give her back to us!

Accounts gone include:

@() @milliontwits Hi @trinawright here nicked daughters acc as they’ve suspended me as well, whats happening?

…
(new account)
@TrinaLWright Right am logging off for a mo as v pissed off to see 2 porn bots following whilst my real account is suspended.. Facebook here I come!

Yes – in the few minutes after she created a new account a new account she acquired two p0rn spammers, including the ever present ‘Sexy ABEGALE’

(@Trinawright’s is back now) (20.45pm)

[@pembsdave] RT @Ladybug8321: Hi Dave my original a/c (ladybug8320)has been suspended, will you plse add me and RT. Thank you

… They deleted the new one too. She is fuming!

[@pembsdave] RT @MontyStillRules: @PembsDave Original account @MontyRules, now @MontyStillRules Please RT Dave

@Treagus@milliontwits oh I see! I wondered whether it was connected. @johnweston said it takes up to 30 days to investigate!

mummbles01 …@shandreen is back as @Applecored temporarily

… @Shandreen (is back now too)

@paul_steele @milliontwits hi can we add @gibbzer ? seems these people are getting jailed for 30 days?

@pennynash @milliontwits @beachhut81 has been suspended too. I’m reading tweets and telling her via fb.!

@agnieszkasshoes @milliontwits @violamaths suspended as well & has filed a request

(Soon after)
@violetmaths Wow, SURREAL evening. Getting suspended, now back. Really is time for supper methinks.

@ ViolaMaths @milliontwits Thanks 4 your tweets on the #twitterfail business. They’ve now restored me. Phew!! Checking out some of the others.

@Treagus @milliontwits that I know of so far @girl_foo and @johnweston
…

@Nuff55 @milliontwits: RT @weeelady Twitter is restoring accidentally suspended accounts tinyurl.com/qvbfte … Actually that was a week ago :S

Sun 5th July 2009
Restoring accidentally suspended accounts 1 week ago
Earlier today, we accidentally suspended a number of accounts.

We regret the human error that led to these mistaken suspensions and we are working to restore the affected accounts—we expect this to be completed in the next several hours.

One additional note: some the accounts suspended were using the third-party site Tweetlater. However, Tweetlater is not to blame for these suspensions nor is it in violation of our Terms.

@paul_steele @milliontwits Hey @felicityfuller came back from suspension, maybe she knows some more info?

@paul_steele hiya they said that “It looks like you got caught up in some sort of spam cloud (through no fault of your own)” @milliontwits

Twitchuck's Friend Chuck in action

Today I am taking an unusual step – using automated software – namely Tweet Spinner to find followers. If should however be pointed out that each time it adds a few I manually check them!

I will only be using the free version but the paid one has a few nice features to filter out a lot of the junk. Primary reason for this is to help @muudboard develop the anti-spam features for his Twitchuck web site. I do this not because I get anything out of it, but because I genuinely think the guys doing a great job – one that I very firmly believe Twitter should be doing themselves, but clearly are not!

This is also an acid test for me for Tweet Spinner and not a reflection that I think it will be full of spammers. It’s merely a cheap, fast way to bulk up on followers for a few days.

*(Minutes later) Meh. Paying for the Pro version as £10 a month as it lets you block out numbers.

One thing I’ve discovered, though hardly unexpected as it has been equally true for years in email, forums and blogs, it that is you filter user out users with alphanumber names (like anna4623) you cut spam dramatically as it kills all the randomly generated names.

If you have some here after being added today (7pm gmt 7/7/09) then yes, you have been “harvested” by automated software! Meeep! That said, you will only have been added because you match things I’m genuinely interest in chatting about. I’m NOT selling and if you are you can take your bag and begger off now, hmm!

I’ve used fairly tight parameter too, like over 100 folllowers, no large margin and so forth. Not as tight as I’d like though. Can’t find the setting to ignore anyone that used numbers or certain words and verbs for filtering. (Found it!)

Keywords to be used:

• Discworld (40 added)
• Terry Pratchett (40 added)
• Rincewind (only 3 ! For Rincewind. Ptt!)
• AD&D (14 added. I wonder how many are false result for that. Ah well)
• Warhammer (Only 15? My!)
• GuildWars (5)
• Warcraft (Zero? No way!. Trying again! 21, better (No point searching for ‘wow’ though!))
• Starcraft (17)
• thesis theme (18)
• wordpress theme (16)
• hosting (25)
• autistic (26)
• Asperger’s (27)
• (thought these would have been spam clean – forgot about gluten free quacks!)

On the ones I’ve trusted Tweet Spnner to add, several have already been removed after a manual check, most for not even bothering with a profile image, 2 or 3 spammers and the rest languages I couldn’t read. (I can struggle with several, but German, Russian, Greek and Asian languages, nope).

Interestingly, it shows I already share 2.5% of both Chris Garret and Jason Bradburys followers. Presumably because of the overlap in interests.

More on this in a few days.

(After adding a little under a hundred, then remembering…)

Current grade with Hubspot’s Twitter Grader is 99.97 (Has been 100%) Does depend on a balance of plain entries, most of mine at @replies
Current rank with Twitter Grader is 1,099 out of 2,619,434
Followers 6,904
Following 6,799
Updates 3,831

Twitter Counter currently rank me 6,583 in a world. Not bad for 45 days of manually adding and vetting everyone until now!

I have my account profile protected and accidentally blocked @someusername. Can you please unblocked it for me (and accept).

On that point:
Why is the ‘block’ right under under the ‘Accept’ instead of to the right of the ‘decline’. Also, why is there no confirm?

New request message for protected accounts - as is

Not being funny but it’s a really bad interface design you have there. So easy to misclick, no confirm and impossible to undo without raising a ticket (like this one).

~ Paul

Below I’ve added an editted version showing how it should have been coded. It still bothers me that with all their staff and venture capital millions in the bank Twitter are unable to get something as simple as a web interface right! Personally I’ve have looked at working something to report spam in there as well, but it seems Twitter are happy to turn a blind eye to that problem!

New request message for protected accounts - as it should be

(It should be noted that @justamoochin is on my friends list. Which emphasises how easily it is to accidently block (and plus flag!) a friend.)

Anyone else noticed that Twitter has revised follower/following for browsers ? Still no ‘report spam’ button *POKE* but it’s a start.

I particularly like that Twitter offer the expanded / contracted list. Nice touch. You can see the last post as a glance, see the sort of thing they are Tweeting and act, if necessary. OK, nothing amazing, but for sorting your lists, helpful.

Twitter follower update expanded

HOWEVER:
I am REALLY, REALLY miffed that they removed the Bio information supplied when you hover the mouse over a persons name. Now, when you do that, it just gives their name…
I ALREADY KNOW WHAT THE NAME IS, I JUST MOVED MY PIGGING MOUSE OVER IT!!!

That information helps us easily decide whether we want to follow or remove a person – or look at at their profile and timeline in more detail. In removing this information Twitter have further alienated genuine users, again, favouring the Spammers!

Twitter browser update - stating the obvious

Also, why have they moved the ‘direct message’ links to the drop down menu and not replaced it with anything to see if your followers have recipricated? Before I could see at a glance how many followed me back!

Why should this matter, you may ask. If you don’t already know, Twitter set a limit of 2,000 followers. After this point, they enforce a 1.1 follower ratio. Unless you are some ‘celebrity’ that racks up a one way following, you are screwed. The only way to get more following you is to keep it balanced. For instance, if you are following 4,440 and only have 4,000 following you, you cannot add anyone. The only way to sort that is to remove anyone that hasn’t followed you back. Now – thanks to this new change – the only practical way to do this is to trust third party applications.

SO, Twitter, I have this to ask you. Why do all the changes appear to favour spammers, penalise regular users (someone mention #twitterjail?) and actively drive users into the hands of third party developers – a majority of whom, it has to be pointed out, are snake oil merchants, outright scams and even phishing sites.

WHY Twitter, would you push users this way?

How the updated browser looks when unprotected

How the updated browser looks when protected. Note the remove option

No, really, have to shout at them. I have no tolerance for stupid stuff like this. Twitter has hundreds of programmers and technical staff, tens of millions of dollars in Venture Capital sitting in the bank. And they can’t even design a simple interface. Even with the api limits people are moving to the likes of TweetDeck by the million because Twitter clearly haven’t a ruddy clue how to do it! Not only that – that are copying the ideas that others have – only badly. Can’t they think for themselves?

Twitter update vs Tweetdeck

Hint: Note the lack of a ‘report spam’ button that Tweetdeck was rightly praised for when they added it. AGAIN, I am forced to ask why Twitter themselves are reticent to do this. What’s in it for them that they seem to want the millinos of spammers flooding in. Ah, perhaps there we have it. The more accounts they can claim, the greater the hype and greater their preceived value, should they sell up. Cynical? Of course, but this is the stuff conspiracy theories are made from, hmmm. Naturally I am not accusing Twitter of this, merely pointing out that, to some, this is how it looks. *cough*

If they could just include the followering/follower count I’m sure it would reduce Spam further. Apart from those autofollowing back (a crazy practice), who really want to follow someone that’s just added 1,000 names and got 3 back. For all the genuine businesses obsessing over getting your follower count up, I offer you this thought:
If you have 20,000 followers and only a few actually see your posts, what value have they?

@Twitter really do need to get on with adding a simple ‘report & block & remove spammer’ action button

Twitter follower update contracted

Twitter following update expanded

Twitter following - contracted

Having the devil’s own job with adding followers, been like this for the past 8 days, (noting the issues on your status page). Firefox gives no error message, just sits there with the icons thing spinning at the top.
MS Explorer tosses up an error message like this one:

Webpage error details

User Agent: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1; Trident/4.0; SU 3.23; .NET CLR 1.1.4322; .NET CLR 2.0.50727; .NET CLR 3.0.4506.2152; .NET CLR 3.5.30729; InfoPath.2)
Timestamp: Mon, 29 Jun 2009 11:44:59 UTC

Message: ‘$flash’ is undefined
Line: 1
Char: 721
Code: 0
URI: http://assets3.twitter.com/static/1245990937/follow.js

What’s particularly irking is I never add/remove more that a few hundred at most a day – and that manually – and always keep the numbers balanced. Yet I’m seeing spammers (using automated software) harvest 20,000+ followers with no problem whatsoever. You see spammers like ‘johngoogle’ with literally hundreds if not thousands of accounts at a time harvesting away, yet using Twitter properly with just a web browser seems to get you penalised.

While using Twitter, I had a problem/bug report

My request is in regard to the Web::Followers/followings

I feel: very frustrated

Fixed!

What I forgot to add – and I stupidly forgot to take a screen grab – was that one one point I had a load of spammers in my ‘following’ list that I’d never added (nor could I remove). Possibly a database listing error?

Initial Stats (rounded):

• 80% failed to provide a homepage URL
• 75% of users have not entered a bio in their profile
• 69% have not specified a location
• 55% are not following anyone
• 54% have never tweeted
• 53% have no followers

Personally I am very reluctant to follow anyone without a profile.

I also work on the principal that 95% of all accounts are in the hands of spammers and self-interested affiliate marketers etc aggressively pushing one way links. As such I also don’t automatically believe or trust any profile unless everything else bears out.

However you look at it, half the population of Twitter is dormant and most of the other half is spam!

Hubspot are more forgiving and use the following criteria to assume only 9% is dormant:

• Fewer than 10 followers AND
• Fewer than 10 friends AND
• Fewer than 10 updates

Reversing the first set of figures, we get:

• 24% of users have a bio in their profile
• 31% of users have a location in their profile
• 20% of users have a homepage URL in their profile

The document also give graphs for things like distribution by tweets by hour – but neglects to give the timezone basis, rather a critical omisssion, really. It does however show that tweeters are most prolific on a thursday.

However: the report glosses over a few home truths, namely:

Of the measly quarter that bother with a profile many will either lie (especially the spammers) or simply put whatever comes to mind in, making any demographic data unreliable beyond a vague guide.

Of the third that enter a location many will either lie (again, usually spammers) or be obscure. So we have Americans claiming to live in London, England and Londoners living in places like ‘Neverneverland’ and ‘Twitterverse

And the rest? A huge percentage of the ‘homepages’ I see tend to be affiliate link. Meh!

Jaded? Cynical? Maybe so, but it doesn’t mean I’m entirely wrong!

One figure we don’t have if that number of accounts – the percentage of accounts that are banned.

Compare these 2009 figures for Twitterholic on Twitter

• 1st Mar 281,904
• 2nd Apr 628,349
• 1st May 990,563
• 1st Jun 1,257,912
• 26th Jun 1,575,535

As show below in the Twittercounter graph below, @twitter’s account has grown by over a million in a few short months:

Twitter counters stats for Twitter June 2009

I have doubts about the accuracy of the following two graphs from Twitteranalyzer, other than them backing up the feeling that the continued exponential growth is not with active users. I could be wrong, and again am deeply cynical, but I suspect a vast number of accounts are actually dormant ones prepared by spammers for ongoing campaigns. In the same way email has been over-run they are laying down literally millions of accounts and bringing them into action as others are blocked.

Twitters followers broken down by join date

Below we can see that a massive 92% of followers, new accounts, have less that 20 followers.

Graph of @twitter's followers showing their own follower base