Tweeters: I want a witness tag

A proposal:

It would be terribly useful if there were a separate convention for tweets from witnesses to major events so their reports can be separated from the discussion that follows. How about !jpquake for witnesses vs. #jpquake for discussion?

Moments after the tragic earthquake hit Japan, folks are reporting on TV, people turned immediately to Twitter to tell friends and family and perhaps the world what was happening to them and to use it to get information and services.

But, of course, in only moments, people around the world talking about the event and the hashtag gets overrun with folks who are talking *about* the event than *from* it. That’s all good and wonderful as well. But I want a way to separate the two.

If witnesses used the !tag, it would also be possible to identify and compile a Twitter list of them. This would be helpful in stories where personal security is an issue. Witnesses in Bahrain would be unwise to use geocoding. But the !tag would merely reveal what they are already revealing in their tweets: that they are there. Somewhere.

Note importantly that the !tag would help people in the middle of a major event — people who need information and services — to also filter out the noise of our discussion from outside.

As for reading !tag tweets, I’d want to filter out retweets and just get the originals.

I also would like to run !tag tweets through translation engines. I suggested that to Ubermedia’s Bill Gross and he and his crew had great ideas on that in return.

The challenge in all of this, of course, is inducing millions of people to add this behavior. Thoughts?

: See also: Doc Searls on how the net turns TV news into newspapers (read: stale).

: Somewhat related: I wish TV news would carbon-date its B-roll. In a disaster such as this, I get loop fatigue: The same video shown again and again and again. It’s understandable why they need to do that and fine that they do. But we don’t know the current state of the story if we keep seeing images of the state of things hours ago. So TV — which ought to be in the fresh news business — should show us how fresh its images are. I know they’re likely loathe to do that. But this is news.

: UPDATE: On yesterday’s TWiT at SXSW, Leo Laporte made an excellent suggestion: appending the ! onto a hashtag. So a witness would use #jpquake! while the observer would use #jpquake and one could search for one or both. Brilliant, Leo. Others had suggested #!jpquake but that messes up the search. This appends a claim to be a witness.

Many people have asked in the comments (over and over) why geotagged tweets aren’t enough. Two reasons: First, many (most?) people don’t enable geotagging; I see very, very few geo tweets in my feed. Second, in rebellions against repressive regimes, pinpointing one’s location exactly is perilous. The ! convention claims location as much as writing about an event as a witness or participant does, without giving the authorities one’s exact location.

  • I love the concept. I think twitter themselves us the ! for a twitter based function for former user names so that’s worth checking, but the idea of differentiating between witness and discussion is spot on.

  • I really like the idea. I have to be a little worried about the legitimacy of the “!” tweets though… I know anyone can publish any tweet and just like any other media we have to sift through the garbage to find the truth, but somehow this has a little more stake in it no? If someone is relying on twitter for emergency information, there has be a sense of legitimacy. Right?

    Would making your location known be a form of legitimizing the tweet? I am not sure, in an emergency, disclosing that information could be dangerous…

    What if we just let go of it and have full trust in the people? It would be an interesting experiment.

    I hate playing the devils advocate though… I LOVE the idea!

    Like your idea, mine is also half baked.


    • Nothing will ever be perfect. But it helps.

      • Amen to that.

        There has to be a way to automatically shift through all the tweets and merely show the “!” ones. That could save valuable time for those in an emergency. Depending on who you know and the proximity to your location, there could be a google like “priority tweet” stream for every individual in an emergency. Maybe you would have to tag yourself in order to be included in this stream? Or could it be done automatically?

        The witness tag is still a fantastic idea :)

    • I agree with your sentiments Dan.

      I think the idea is pretty great, but the first thing that bothered me, which I think comes up for any digital or internet issue, is the problem of both the legitimacy of the witness and tweets and the problem of trolling. I would hate to think that there would be people who would completely abuse this, but it is something you have to think about.

      However, this would benefit emergency aide people, from the police and military to the Red Cross and the UN, for example. In fact, I think this would be a better for services then it would be for stories, though I will not completely deny that aspect.

      For people in crisis, you would have at least a sense were one could find the closest if not immediate aide.

      Great idea and I hope it goes ahead in some way.

  • J.C. Derrick

    Outstanding idea.

  • The moment people re-tweet an eye-witness hashtag would become invalid wouldn’t it? Unless people edit the hashtag before they re-tweet..?

    Most of the live-tweeters also curate and re-tweet stuff. In a way maybe that’s part of getting the whole experience of what they see/hear/feel. Certainly, when you’re tweeting something at the same time as your living it, the overall experience and your impressions do sometimes change slightly.

    If you’re looking for voices on the ground In Japan right now, here’s a list of six Global Voices bloggers who are there:!/globalvoices/global-voices-in-japan-8 We have a few more list of tweeps from different countries. I’ve seen other media and blogs do that too. It’s a pretty quick way to get an overview of who to follow when hashtags become too overwhelming.

    • Read what I said: We filter the RTs of the !tags. Easy. Peasy.

  • Awesome idea. The bang(!) might be too difficult to parse though, since it’s a fairly common character, and would be prone to mis-tagging. Sort of like how people post something like “The Astros are #1” and views it as a regular hash. I would propose something less common like the carrot (^). Though, semantically, it might even be better to use an asterisk(*) since it implies a “footnote”, which would make more sense in this usage.

  • That’s a fantastic idea. However, I’d instead like to see a separate hashtag for a witness. Like #witness or #live or #OnTheGround or #OTG. The reason would be that when events happen, nobody really knows the “official” hashtag. The combination with the “official” hashtag would eventually happen, of course. But think of this tweet:

    tsunami waves hitting hawaii now. No damage. #witness

    Even without an official hastag, we could find this report looking for tsunami news from hawaii.

    Wow. Think of searching twitter for just “#witness” to find people reporting live news about anything. That would be powerful.

    (But! Pick something other than #witness — there’s already lots of noise with that hash: )

    • a #witness hashtag won’t work because how do you search on that plus #jpquake when the characters are not contiguous. Also takes up precious space. I say we need another convention.

      • Jen

        I think the #witness hashtag (or something similar, maybe it shouldn’t be called witness) could be used in combination with a hashtag for a particular event. If you review the Twitter advanced search ( there are options (for example: using the “all of these words” field, simply enter both hashtags you wish to search for).

      • Jeff, I’m buying what you’re sellin’ with the new convention proposal. I can dig the ! usage, too. Currently, looks like Twitter does not recognize ‘!’, tho.

        But you CAN currently search for two hashtags w/ Twitter search, even if the hashtags are not one right after the other: ( ex: #jpquake + #fb = people who are updating their facebook account with their earthquake update)

        So, just substitute #fb with #witness.

        And you CAN get rid of Retweets very easily: use ‘-RT’

        But I think the biggest problem with your idea is that people may not know what the official hashtag for an event is. But people could learn what hashtag to use for a first-person account of news. Maybe it is #witness .. or !witness … or maybe !holyshityouwontbelievewhatIsaw

      • I just had a play on twitter with various existing tags. The fewest characters that I could come up with was #8! – the number eight and an exclamation.
        It would need the sort of publicity that the smiley face got. But used in conjunction with the hashtag for the event it could do the job.
        RT are another problem. I always want to edit mine so I would welcome a way to easily do this. The RT mechanism could drop the #8! but if retweeted in full then it serves to point to an eye witness.

  • Pat McDonough

    Indeed the !something pattern would be very useful; and I would love to see this evolve.
    But this lead us, at least in the long run to the semantic vs syntax vs markup problem, which is less of a problem for tech and nerd kind of people, but not as easy for the rest.

    I personally would love to see more semantic context via markup in tweets and also other web content, but I’m doubtful most people could use it.

    Maybe it is easier to filter the tweets according there geographically meta data to get a 80% solution.

  • Mark Cridland

    Human knowledge and communication have never, ever offered the cleanly-sortable distinctions Jeff seeks. Our understanding of the world has always been a blend of our own experiences and descriptions of the experiences of others, including people we don’t trust. There’s no way to demand people demarcate only the things they’ve seen with their own eyes without getting into some violent facist behavior.


    • +1 (except for the violent fascism)

      • Mark Cridland

        Ms. Larson, let me put it to you this way…

        There’s an epic event in New York, and I !tweet! about it from here in Los Angeles. I’m a *liar*!

        Now, like, whatcha gonna do about that? You’ll punish me. Either you’ll fine me (or worse), or you’ll deprecate my ability to communicate on twitter…

        …Which is what you do already by not following me!

        This isn’t a new problem. It’s the oldest problem in literacy, and the reason great new books continue to be written on old topics by young researchers.

        We look at the material and decide who to trust.

        Ain’t no need to be gittin’ all hashtaggy about it.

      • cm

        As you say, trust is built up and if someone keeps feeding you bogus info you will stop following them. Basically you’e talking about building brands. If you want a quality news feed then you’re going to follow CNN, NYT or whatever.

        But that does not help you for unplanned events. If you want to get eyewitness realtime reporting then you have to open up the filter. Those reports are not going to be coming from people you follow but from the random people on the ground.

    • What are you talking about? The hashtag itself is an example of just this sort of distinction: this tweet is about #this.

      • cm

        I think you misread what he is saying.

        #subject just tells you it is about subject.

        It does not distinguish fact from rumour, speculation, analysis and commentary which might all be mashed up into a single tweet.

        You want “!quake Big quake in Japan. Building X fell down” to mean “I was standing across the road and clearly saw building X fall down with my own sober eyes.”

        But it is seldom as clean as that. People will use:

        “!quake Big quake. Lots of people will be hurt.”

        • Life is never clean, people. It’s an added signal. Signals are valuable. Ask Google.

        • Mark Cridland

          Specifically, ask them about their spam filtering. !Signalz!

        • Mark Cridland

          I mean c’mon, Jeff, you just tweeted that it would be on the honor system. OK, fine! The scheme would be spammed to uselessness withing milliseconds, and forgotten within minutes…. No harm no foul.

        • sxa

          @Mark I disagree – it would have value for more local incidents if nothing else what would be less likely to get hit. Can’t slate a system just because of spammers

        • Everything is spammed. So let’s kill email. No more blogs. No more Twitter. That’s it. Shut it down. Silly.

      • Mark Cridland

        CM’s comment (above or below, nearby!) *gets* it. He groks, he’s totally clued in.

        JJ is too eager to believe that things can be mechanized, that everyone can, will, or SHOULD agree what an “eyewitness” is… Or that such a standard can be enforced without ruining the sharing impulse that makes Twitter so glorious.

        I mean, while we’re at it, let’s all agree never to use email to sell low-cost mortgages or sexual stimulants. C’mon you guyz, promise! Cross-your-heart and hope-to-die! !eyewitness!

        • Gawd, man, how many ways can you throw cold water on this. OK, OK, I get it. Now can you make a constructive suggestion?

        • Mark Cridland


          Please understand:

          • That virtue is not a subroutine, with rote algorithms which all will admire. People you think of as distant from events will demand access to the !resource!, whether or not you think they oughta have it

          • That all rational processes are also human processes, requiring discernment, review (not just peer review), and considerations of personal integrity

          • That the p*ckerhe*d self-interested will eventually weasel their way into anything that’s been greased for the efficiency of a noble elite

          Essentially, I’m suggesting that people do their homework as they always have. Twitter (etc) have been a fascinating tool and a thundering blessing in the dissemination of human understanding. They not need also be perfectly convenient. Our individual, time-consuming judgments of the worth of each source are an essential part of progress.

          Listen, go ahead and give it a shot. Why would anyone argue? I’m just saying that when the ess hits the fan, I’m still going to trust the blue-collar voices I’ve always turned to, and the !eyewitness! tuxedo won’t count for much.

  • Matthew Phillips

    Jeff, the problem you’re revealing here is actually how bad Twitters search engine stinks. It’s overrun with “RT”s and links to articles. Seeing people talk ABOUT the earthquake? I’d love that, but it’s hard to find from Twitters terrible search page. is much better.

  • Till Eulenspiegel

    There are a lot of things that Twitter *could* be, but it’s become increasingly clear to me that the “firehose” is worthless. Searching on a tag becomes pointless as soon as there are more than a hundred or so users tweeting about it; it becomes flooded with bullshit and spam.

    I think Twitter is much more interesting if you look at it as a web of trust, people rather than data. You need a few good contacts as a starting point, and you can branch out from there based on who they mention. I used the Al Jazeera journalists for just this purpose during their Egypt reporting.

    The links (who follows whom) are what’s important in Twitter, not the content of a single tweet. You won’t find the anonymous, un-connected person with good information this way, but how would you find them beneath the spam anyway? Any kind of ad-hoc tagging system will inevitably be abused and misused.

    • I’m trying to syphon some stuff off the firehose. Jeesh, you give up so quickly. Damned printing press. Just a firehose of stuff.

      • Mark Cridland

        Eulenspiegel covered this: No one’s offended by the presence of Modern Bride on the newsstand, because they can reach over it for a copy of Car & Driver.

  • Love the idea. To limit fake !tweets, twiter can also somehow apply a geotagging qualifier befoew these tweets are tweeted.

    • that’s an added layer of data but geotagging is a problem because of limited uptake and because of security problems in revolutions.

  • Well done, Jeff!
    All twitter has to do to cement it is offer specific search & publish (similar to #). It’ll catch on quickly.

    % could be used to denote Pictures. Sometimes I just want to see what pictures my folks have published lately.

    – “The Next 10 Amendments”

    • That’s up to the clients more than twitter and one can search on any characters, of course.

  • Mark Cridland

    Friends, to imagine that you can just dash straight to the candy is a twisted, AlGore-style control fantasy.

    In real life, people gossip (see also, Entertainment Weekly). They get information from each other. It’s ridiculous to think you can identify, and readily consult, ONLY those people who’ve had precisely the direct experience for which you want insight. Human experience doesn’t work that way… And it didn’t work that way in the point-to-multipoint TV era, either. Nor inany generation before that….

    And I’m reminded of what someone wrote when George Harrison offered some comments about a Beatles scandal, demonstrating profound obliviousness to the impact his group had on a near-global generation: “The inside of the space capsule isn’t the best place to judge the arc of travel.”

    We’re SUPPOSED to have cites and intermediaries; that’s how our brains are built.

  • Mark Cridland

    I was about to make fun of Till Eulenspiegel for calling twitter a “useless firehose”, until I read the rest of the comment. This part is brilliant-

    > look at it as a web of trust, people rather than data.
    > You need a few good contacts as a starting point,
    > and you can branch out

    That’s perfect. Of course we go to the sources we trust first… That’s the way society is SUPPOSED to work.

    Technology will never supersede our need for personal judgment in communications. Virtue is not mechanical.

  • Love the idea. Obviously, as others point out, there will be abuse and misuse, but journalists will be able to better sort out facts by looking what is being repeated. Certainly, it’s a better system than what we have now.

  • cm

    Ever since the realtime reporting of the “BalloonBoy” scam I’ve been interested in whether the real-time citizen journalist idea is really going to pan out.

    I live in the area of New Zealand that got hammered by earthquakes in the last 6 months or so. I watched the tweeting and don’t think I really got to see any really high quality live reporting tweets.

    There were many of the form “Just got shaken by a huge earthquake” . Hardly tell you anything new.

    There were some of the form: “OMG that was huge. Must be M8 or more”. Perhaps that is what the person thinks they felt, but completely hogwash since the person has no clue of what magnitude means.

    I saw quite a few tweets which were just people itching to be the first to report a rumour. Quite a few tweets from people in the city saying that building X had fallen down when in fact they were just tweeting what someone said (and was completely wrong). These were reported as fact.

    Some news feeds were obviously harvesting twitter for some of their news. I saw reports of building collapses that were later removed.

    Using a different tag for witnesses vs discussion would help separate the threads, but is it enough? Is there really a binary dividing line? Sure a person living far away is not a witness, but is a person in the city a witness, or just a person who actually sees something and would swear in a court of law?

    It would obviously help if people reported professionally. ie If you heard a crashing sound and someone says “That was building X collapsing!” Do you really think your citizen journalist would tweet “Big crashing sound, maybe building X”. No, they’ll say “Building X collapses”.

    I think it is hoping for too much to expect quality reporting from twitter. The people are far too motivated by the idea of being one of the first to tweet on an event and that will always put immediacy ahead of quality.

    • How cynical and snobby. You devalue hearing from the people who are actually going through the experience vs. reporters who arrive hours later? It’s PEOPLE TALKING! Don’t you want to listen?
      Let me guess: a journalist, are we?

      • cm

        No I am not a journalist. What I am is a citizen in a disaster situation in search of quality information. I don’t want to hear rumour, speculation and people saying “OMG that was louder than a bomb!” (a useless thing to say unless they have actually heard a bomb going off).

        What I want to hear is quality information from sources I can trust so I can make informed decisions: Should I evacuate to higher hand? Should I worry about friend X?

        When I got shaken out of bed in Sept 2010 by a M7.1 quake I turned on the radio to find out what was going on. As with most media it was thin on facts and high on drama. Lots of human interest bullshit (phoning stores to hear how stuff fell off shelves). Also lots of rumours etc.

        What it really comes down to is what the consumer of the news is looking for.

        If you are personally caught up in the event then you want accuracy even if it takes a bit longer to get it. That quality is needed because you’re going to actually make serious decisions based on what you hear.

        If you are at arms length then the news is only really infotainment (eg. if you’re in USA or such then a quake in Japan won’t really change your day). High drama is good and fact checking is optional. If not, I’ll flick channels to watch “reality” TV.

        Unfortunately (or fortunately, depending on your viewpoint) the latter is the segment that modern journalism is increasingly serving. Even documentaries – which are supposed to inform – are getting edited for drama and viewability rather than to convey information.

      • sxa

        I agree that I’d trust sources from professional journalists more, but if there is no information, and the reporters won’t get out of bed for hours, then I think there is benefit in being able to identifiy first hand reports, from 2nd hand, 3rd hand, rumour

      • TL

        JJ you got served. Nothing like immature name calling and assumptions (“journalist, are we?”) to make oneself look stupid.

        CM makes very valid points from a real world perspective. Since they don’t agree with your wonderful idea, you bash and dismiss?

        To quote the ‘new media’: you butt hurt?

  • sxa

    I’d propose using something like #! as a sequence instead, for two reasons:

    (a) It resolves the ‘accidental’ use of witness tagging caused by people just not putting spaces after a sentence ending with !
    (b) it will let the syntax be treated and processed by any current tools that already recognise hashtags

    I wrote a blog entry about some issues with determining sources of information from an alleged explosion near me recently, and the heresay and lack of people posting sources was a significant problem – a syntax such as this would likely have helped a lot:

    • I fear it complicates search.

      • cm

        Nope. Searching for #!foo is as easy as searching for !foo. Ask any programmer.

      • sxa

        @cm exactly, not sure if I’m missing something! Although having just tweeted with a test post including #!jpquake, twitter’s search function doesn’t seem to find it, but then it’s not finding !jpquake either … maybe ! wasn’t such a good idea after all unless twitter changes the search function ;-)

      • Steve


        most programmers will tell you that searching for !foo will find anything except foo, as ! is programming shorthand for NOT, so there’s scope for unexpected results

  • Mark Cridland

    Sxa, I hate to be a negative nelly… But this reminds me of something from Jeff’s cuny thing the other night, when the AOL guy said “New features should fail quickly”. I think this will fail instantly.

    Email would be worthless without enormous antispam investment (especially Google’s). If the !eyewitness tag gets abused as I believe it would, you’re not going to care whether actual eyewitnesses use it or not. You’re not going to know whether actual eyewitnesses use it or not.

    (Andy Carvin has been dealing with this issue almost every couple of hours on his firehose [but wonderful!) Middle East stream.)

    • Well, gee, not a single soul has used it and you declare it a failure. That was helpful. Thank you so much for your contribution. Jeesh.

  • Great idea Jeff, although I have a bit of a problem with the exact definition of the bangtag (not sure if that’s the most appropriate name to go with, but…).

    My problem is what defines the event in question – if I am watching your panel at SxSW am I ‘witnessing’ it? Or are you proposing that it applies solely to natural disasters? If not, then does the backchannel at the event have to cover both the hashtags and the bangtags. Thinking about this, it might also come in handy, as questions from the room can be filtered over outside the room.

    And for all those complaining about the ability to game bangtags easily – welcome to Twitter. How many times have you been spammed because of a keyword in your tweet? Despite this hashtags still have value…

    • Communities figure out the norms of usage of conventions over time.

  • Chris

    Great idea Jeff. As for the abuse of such a tag, I can’t see how that changes from what we have now. I’m watching Al Jazeera right now and they’re including tweet content, but there is no guarantee now this is accurate. Right now I can post “Big lie about event #event”, I don’t see how this flag changes that at all. Surely we don’t have to be such a socialogical nanny people, surely most people can discern the amount of trust to give any given source of information?

    If ! could cause a technical issue, how about :jpquake?

  • I’m not sure if you realise that you can search Tweets based on their location. You type: near:Japan in the search field and you can only see tweets posted from that geographic location. Aside from that I agree that !Japan would be excellent. However in this world of people who are now self-organising (see my posts at people may not do that!

    • I”ve said this below; geo has limitations: many don’t use it; it is too precise in places with tyrannical regimes.

  • Fray Close

    I think it is brilliant. With following the revolutions, there is a very real need for something else. I refrain from cluttering up hashtags, but then it does itself a disservice, when people can’t find the conversations. Last night we were both not tagging our posts discussing Japan, while it was happening to keep it more full of news, but at the same time complaining about how long it took to hit trending topics.
    The whole point of Twitter is that it is fluid and changes with how users decide to use it. Adding a different type of tagging, once used en masse would make it a more useful tool. The way we use it doesn’t even have to perfect right now, this is twitter! We go with the flow and adapt.
    Will it be abused, maybe, probably, but it takes pressure off the system we have now.

  • Jeff’s idea sounds great. It’s a simple solution, it will provide substantial value if it works. We don’t have to decide in this forum, whether it will actually work. Time will tell.

    However, I have one suggestion. Why not expand it’s use to include helpful emergency information for those afflicted. Let me explain:

    The !-tag is an attempt to make it easier to find witness-accounts in major events where the firehose-discussion from around the world tends to drown out important information. Jeff’s idea would give eye-witness accounts it’s own channel, which sounds reasonalble, as they are more important than most other tweets on the topic. We’ve seen this during several of the riots/revolutions in the Middle Eastern countries.

    However, Twitter is an important channel for similar – but slightly different types of events – such as earth quakes, plane crashes, flooding, and other major emergencies. (Not a word about conferences).

    In some of these emergencies it would be of great importance if factual, helpful information would also stand out. I’m thinking of Emergency-phone numbers, government support sites (which people may not know about), sites like Google’s people-finder, evacuation orders and or other important facts which may be both urgent and important.

    As journalists we often focus on the big headlines and forget that users/readers are often on the hunt for factual information first. Where’s the emergency-center? When will power be restored? Are major highways open? Where to donate money/blood?

    Afterwards – when we take a look at the web-stats – we realize we should have provided news-you-can-use as well as telling the big story.

    It would be really great if Jeff’s idea could not only help the world locate witness tweets with high credibility, but also find the information which is helpful to those in need.

    So basically my suggestion would be, that the !-tag should be used not only for witness-tweets, but also for factual, emergency-information. It’s still simple. It’s still only one tag.

  • Pingback: The future is here – tweeters to be licenced | International Times of Dominica()

  • Curt Tweedle

    This is an excellent idea.

    It solves a major reason why I don’t use twitter for major events (Politcial unrest in Iran, Egypt.. the whole Middle East and the Disasters in Haiti, New Zealand and Japan for instance). Too much noise and very little signal. A cacophony of retweets and people just talking about the topic but are not witnesses nor actors in the event. The signal I really want to get to is the people that are on the ground or actively involved. Not the people who retweet. All they are is a medium not a message. Useful but not what I want to interact with.

    You could apply the same idea to blog posts or news articles, as well, to help search engines identify a post as the original source. A tag or code could be placed in the post. Preferablely in something that does not appear in the browser window as clear text (such as the head tag or similar) so that the tag/code would not be grabbed in a simple copy/paste of content to help avoid people who do steal… I mean… grab the content.

    Yes, it would take a bit of training but so did hash tags.
    Sure, in both cases people can fake it. Nothing is perfect but it will help cut down on the above problems and makes those who want to imitate an original poster have to do more work. Also, Twitter and search engines could look at which poster is being referred to/retweeted the most as a signal of originality.

    Thank you, Jeff, for a great idea.

  • Curt Tweedle

    Also, I would agree with the timestamp of video footage. Me and my friend were watching the earthquake reporting. We had to make a continual analysis to determine if the footage was new or old since the Live stamp was all over the screen. The Live stamp applying to the reporting and not the video.

    And today brought another issue. Sure all video footage was old. That problem was solved, but now it was hard to tell what time and location the video was from.

  • Love the idea Jeff. Thanks for proposing it. As for adoption, let’s spread the word and see what happens!

  • I think it’s worth trying, someone will have to build a search engine that treats these symbols properly though, as currently none of, or google realtime do so.

    I sent out a test tweet on a spare account with !test %test ~test on it and can’t find it using any of them.

  • Excellent suggestion, I hope it gets heard. At the moment it is so difficult to get information from independant observers. I want would like to also see data on the radioactivity 20-30 km around Fukushima or in Tokyo measured by independent sources if they exist.

  • Pingback: Pleidooi voor “ooggetuigen-tag” op Twitter | Twittermania()

  • Brian Caldwell

    I love the concept.

    We need other types of validated tags also, like voting up or down a newly proposed idea from the crowd.

    Perhaps the overall concept of tagging needs to evolve?

  • Pingback: 7 Thesen zum Erdbeben in Japan: Live-Internet, Crowdsourcing und der Disaster-Capitalism-Complex()

  • jan bar

    What you are suggesing is not a good idea nor necessary. Twitter already has a method for categorization.

    Your observation also proves a fundamental weakness of twitter and why it is only passing fad.

    Please forgive my English…

  • Patrick Fleischer

    as people have pointed out the reliability of this silly idea would be about as good as trying to get spam senders to identify their spam as spam… good luck, jeff!

  • Great idea. Maybe this will spark more creativity around separating out streams of consciousness- new symbols for circles of conversation that need to stand out and bring authenticity to them. Maybe responders to the witness tag can only be groups sending aid- communicating, organizing, reassuring.

  • Interesting proposal, but I don’t see what it gives you over geotagging? Is the concern just the level of detail that geotagging provides? Because if I witness something and tweet about it, you know more or less where I am anyway.

    Twitter does have a way to provide more generic geo-location information. Here in Canada for instance, if I use a geo-enabled web browser, I can specify a neighbourhood to tag along with my tweet, which is different and less specific than lat/long coordinates.

    • Please read the rest of the discussion as this has been addressed often. One more time: Many people do not enable geotagging and in tyrannical regimes exact geotargeting is dangerous.

  • Thank you..really informative!!

  • Jeff, I appreciate the sentiment of your proposal and understand how you’d like to harness people’s existing behavior to provide a means to bubble up information from people *on the ground* versus people talking *about* it. The ability to do that, in real-time, would be very valuable.

    However, I am concerned, as you are, about adoption. As you say, “The challenge in all of this, of course, is inducing millions of people to add this behavior.” Indeed, that is the challenge, especially if prepending this character would only be used for journalistic purposes, rather than mundane, day to day use cases. Moreover, I would imagine that in a moment of crisis, most people tweet for one another, and that producing journalistic information, unless they are journalists themselves, is often a secondary or tertiary goal — especially for people who do not already have an audience.

    So, in order for this behavior to make its way into the vernacular of the long tail, there would need to be some other kind of motivation to use this form more regularly, to signify something globally meaningful. You’d also probably need to get Twitter to develop some feature that would lead to people being aware of the feature, like a real time index of “witnessed” events, or to provide inline linking of these tokens.

    But then we’re sort of back to a second form of hashtag, which seems to work against your original goals (i.e. being gamed, etc).

    Curiously, the solution seems less about the token of expression, but to encourage everyone on Twitter — or any other social media channel — to simply provide more public information about themselves, their behavior, and who they are connected to so that we (or computers) can attempt to do a better job of vetting the sources of information.

    Bottom line I guess is that changing people’s behavior is hard, and getting them to adopt a convention that doesn’t benefit them immediately or in obvious, transparent ways is going to take a long time to get off the ground, if at all. Not that you shouldn’t try, but it’s just that you wouldn’t be the first either:

    • I’ll add one word to your comment that I think clarifies the point:

      “Curiously, the solution seems less about the token of expression, but to encourage everyone on Twitter — or any other social media channel — to simply provide more public [and databanked] information about themselves, their behavior, and who they are connected to so that we (or computers) can attempt to do a better job of vetting the sources of information.”

  • Carlos

    It sounds like a good idea, but how do you suggest it could be implemented?

  • I could see it as a first level of filtering, culling the self-proclaimed eyewitness accounts and putting them in a stream where they can be scrutinized, compared with other accounts, considered for outliers, etc. If you’re going to claim the ! tag you’d better be ready for the spotlight. The existence of the tag would train or market people to value eyewitness accounts as distinct from peanut-galleryism. We’d have to think about how and whether to retweet, unless the eyewitness account can only go as far as the witness’s followers. Maybe there’s a way to point people to the ! tag without messing up the tag.

  • Pingback: Twitter loves natural disasters…In a good way « DigitalDoesn'tHurt()

  • Nanker Phelge

    Jeff, any reaction to the recent layoffs at the Ann Arbor News Web site you were so proud of some time ago?

    • Do you have a link?

    • I haven’t been involved in the site since shortly after its launch, so i don’t know anything current on the state of the business, btw.

      • Nanker Phelge

        Just to be clear, I don’t know that it was a dud – I didn’t write the message below – and I don’t know much about what you did (other than you expressed significant enthusiasm about it). There are certainly reasons for skepticism, though. And, fairly or not, this was seen as a bellwether for the online-only daily model.

        • I found the Ann Arbor Chronicle’s post on the layoffs: here. As I say, I”m out of date with the project.

        • Nanker Phelge

          >>>I’m out of date with the project.

          OK, but what do you think? When the Ann Arbor News made the transition, you said it was “worth watching.” I’m sure you must have paid some attention.

          At the time, you vocally defended crowdsourcing, free news models, and the practice – disturbing to some – of laying off an entire newsroom and making everyone re-apply for their jobs. Do you think this ecosystem is working? if so, what are we missing? If not, why? Would you give them the same advice today?

        • Nanker Phelge

          >>>Do you think this ecosystem is working? if so, what are we missing? If not, why? Would you give them the same advice today?

          Jeff, you’re usually so vocal about your opinions that I’m surprised you haven’t addressed this. Any thoughts?

        • I’m busy

        • Nanker Phelge

          Make some time. You were sure eager to talk about Ann Arbor when you thought it would be a success.

        • Nanker Phelge

          More news from – they put a U of Michigan PR executive on the _editorial board_.

          When are we allowed to say this isn’t going well? When are we allowed to ask your opinion of it?

  • This is too bad…

    Tony Collings is a University of Michigan communications professor and a former journalist. He says a lot of journalists and professional journalism organizations were looking to see whether Ann would be some kind of model for the future:
    “Apparently it isn’t, or at least it doesn’t seem to be succeeding in a business way, and I don’t know whether it is succeeding journalistically either.”

  • Currently ustream is used to share live geiger data from two different places in tokyo.
    In your new book I personally would appreciate to have a chapter on how best to create an own personal “independent news room” – where do go to get critical data from different sources – what to learn from the current crisis. And also what at advise you would give to government organisations and large institutions on using social media in crisis communication. What social media competences countries should give their community members to be more effective in crisis ?

    I started to collect the different sources of radiation monitoring websites worldwide
    My observation:
    By the way twitter searches and having a blog post were the two key tools to identify these sources. Google search did not work as well for realtime critical data sources. Also facebook was another source as people in Tokyo began exchanging there on the crisis.

  • Generally, and the earthquake in Japan is a good example, it soon becomes obvious who is at the source of the news. For instance, there was a spanish twitter, @kirai in Japan, and in a matter of hours he went from 4k followers to 22k. So I’m not sure if the separate hashtags are the way to go.

  • I sent this to Twitter today:!/adspedia/statuses/47808490663714816 … we should all do the same (feel free to RT)

  • And Twitter’s new TOS probably forbids this because it can’t be monetized by them (cynical after two glasses of wine and sorry I didn’t get to talk to you more at Momo’s)

  • I’m all in favour – I suggested adding EW next to the hashtag in the aftermath of the Mumbai attacks in 2008.

    I also wondered at the time whether Twitter could not build some kind of tagging system (like you add tags to blog posts) into its functionality so that it was not necessary to include hashtags in your precious 140 characters of tweeting space…

  • I think your updated idea #! for witnesses is a great place to start. Any journalists worth their salt isn’t blindly retweeting or taking anything at face value without fact checking and considering the source, unless credibility and rapport has been established.

    If the “!” doesn’t work, fine. We can come up with another character that achieves the same thing. Devil is the in the details and gets a much needed conversation off track.

    So I’m in with #! until I hear otherwise.

    Now, about tracking Twitter corrections? That could use a style/convention as well…

  • Twitter should utilize geolocation to prioritize local tweets from disaster zones. They could build the functionality into the main site, but set it to turn-on with a switch flip by the webmaster.

    • After reading more of the comments, I see that this has been addressed somewhat. I expect that this issue might best be addressed with some combination of hashtags with geolocation. Otherwise, it’ll be hard to discern the legit from the spam. Or, in the case of activism (ie: Bahrain or Libya), the counter-posts from the regime seeking to clog the channel.

      The more I reflect on this, the more I think it will take some kind of moderation to compile a list of legitimate tweets. And that process will slow things down.

  • Victorine Quinn


    go away!

    your question off topic
    Jarvis wasn’t that involved
    the curmudgeons ruined it
    It’s hard to judge its success, it’s not a perfect world, give him a break (“jeesh,” he teaches journalism, of course he cares deeply about this)
    he’s busy writing his next book (public parts, be sure to pick up a copy this fall)
    he’s dealing with personal illness, answering your picky questions is not his priority right now

    • I’ve said that I”m not up to date on the project and have no fresh information. The last I spoke with anyone there was about a year ago. So I have nothing new or informed to say.

    • Victorine Quinn

      Jeff Jarvis says, “…have nothing new or informed to say…”

      Well, …that’s never stopped you in the past.

      You must be too busy tweeting “new and informed” things…

  • Pingback: Filter feeding from the Twitter torrent « Brannan Blog()

  • I think that twitter should by default put your country location from where you tweet in brackets next to the method of posting (eg via Bit.Ly (Egypt))

    Those who have location tracking turned on further will have the local town/city alongside the country … it’s not quite a witness tag as you want but its better than the “anyone is a witness if they retweet” rubbish we have now.

    roll on citizen journalism

  • Pingback: This Week in Review: The Times’ pay plan unveiled, a SXSW primer, and a closer look at NPR’s foes » Nieman Journalism Lab » Pushing to the Future of Journalism()

  • Hi Jeff, This is a great suggestion– adding the ! to the hashtag — it does gets really difficult to sift through and determine which information is from from the field, and which is not.

    I’m not sure if you recall me, I was an (inadvertant) citizen journalist in the aftermath of the 2004 tsunami (had been vacationing in Thailand). My question to you–I have the time and willingness to go to Japan, but not the financial resources nor the connections, to interview and write for an online pub, etc.

    I’ve no journalism cred, but have done plenty of blogging for 7 years, tweeting for 4 years, and also a lot of creative writing as well. I have spoken to hundreds of survivors and relief workers post-tsunami (2005) plus post-Katrina (2007-2008) and am extremely sensitive listener (I was appalled by slant of questions that journalists asked me, so learned the hard way).

    Suggestions whom are good organizations top of your head that I could cover Japan from the field? Thanks so much in advance.

  • Ross Bailey

    I like the idea of this. My question is whether you think it would lose value if it was used in general for people to indicate that they were “at something”. Do you think this should only be for newsworthy events or do you think it would be accepted in wider usage e.g. to denote being at a protest rally for example?

  • Professor Jarvis,

    I notice you master degree cannot be found. Does it still exist? Where is it on cuny website. I search but only can find just a few classes.

    • See this page.

      • TL

        Ummmmm. That page you referenced doesn’t mention a Master degree, let alone a degree at all. In our country if you want to teach at an accredited university, you need at least a Masters and preferably a PhD. I guess the standards aren’t as high at CUNY?

      • The person was asking about our new master’s program.

        I graduated with a BSJ from Medill. I teach in a professional school, which does not require a PhD. It values experience.

        TL, whoever you are, what does this have to do with the topic at hand? If you come back to launch another off-topic attack, I shall presume your intentions are only trollish.

      • I still not find master degree. It say “a one-semester course of study”. Is it master degree?

        • Zhi,
          Sorry for the confusion.
          We are not yet approved by the state Department of Education for part of this.
          What we offer now is:
          A masters in journalism over three semesters with an additional one-semester certificate program in entrepreneurial journalism. MA students may continue on for the fourth semester to earn the certificate. We also offer midcareer professionals the opportunity to come and take just the certificate program over one semester.

    • BobJ

      Do you mean CUNY’s Master’s program or Jeff’s masters degree?
      He doesn’t have one, just a BSc which most University’s not only won’t let you be a proffesor with, but you can’t even TA with just a BSc. Don’t worry, Jeff will remove this post shortly.

      • No, I’ll leave it here. I’m enjoying your misspelling of professor. And the misused apostrophe in university. And the unnecessary capitalization there. And the inconsistent capitalization of Master’s. And the missing comma before “which.” And the tortured sentence structure ending the clause with “with.” And the run-on sentence. Oh, the joy.

      • I also am impressed that you’ve used five separate names under the same IP. Fan of Tara, are we?

  • I offered an outline for a more general microsyntax for disaster related communication many many disasters ago, called Bang: see

    The thinking is more general, in that it is not intended for journalists, but it certainly is geared to folks on the scene, asking questions and providing information, like this:

    !bette /usps, provincetown MA/ [email protected] haque: compound fracture of the lower right leg

    The idea of this microsyntax is to structure the messages using ‘punctuation’ instead of natural language tags. The ‘!bette’ at the start states that the message is associated with a disaster called ‘bette’, in my example a hurricane in the NE US.

    These Bang disaster codes could be parsed — or created — by simple software tools, or manually by anyone with a wallet-sized instruction card.

    It was years ago that I attended your Disaster 2.0 session at Winer’s Blogosphere conference, and we still haven’t gotten very far.

  • I’m in. Brilliant.
    I just finished WWGD the audiobook on my ipod. flew through all 9 hours in just 3 days – exciting stuff. I’m an innovator in a field full of regulations and privacy. This gave me some great idea.

  • Pingback: 5 Ways Journalists Can Use Twitter()

  • nikki

    As for legitimacy, that’s a neverending issue. Everyone using the internet needs to pick and choose, since no filter will ever properly do that.
    I think that identifying conversations from witnesses is a great idea in general… although I’m sceptical about live tweeting from a disaster. If you need to be found, then good, do it, but if your safety is going to be comprimised BY tweeting about it, then I’m not for it.
    I think making a website would be a better option for safety reasons(?) .

  • Pingback: ¿Es posible castellanizar las palabra “hashtag”? | Grupo de Nuevas Tecnologías de la SoMaMFyC()

  • Pingback: The blotter: Week ending 13 March 2011 | ARTS & FARCES internet()

  • Ertugal Maxwell

    Wow!!!! what an experience! dr.marnish is a wonderful spell caster, he has made my life complete again by helping me cast a spell to return my girlfriend and also make her to be faithful to me again. I was skeptical at first, but what a believer I am now, his spell really worked! my lover is now faithful to me, if you are also seeking for help to get your lover back? Call +15036626930 or email [email protected]
    Ertugal maxwell

  • Pingback: Surviving The Tsunami That Wasn't()