Charlie Warzel reports on hints and statements of intent at Twitter about moving the ‘arcane’ language of microsyntax (@, #, in particular) down into the infrastructure. He mentions a recent talk by Vivian Schiller:
What will Twitter look like in a year? Two years? A lot less like itself.
At least that’s the impression Vivian Schiller, head of news at Twitter, gave addressing the crowd two days ago at the Newspaper Association of America’s mediaXchange conference in Denver. During her talk, Schiller called at-replies and hashtags “arcane” and hinted that Twitter might soon move them into the background of the service.
When asked about the comments, Twitter replied that Schiller was echoing a similar sentiment that the company’s CEO, Dick Costolo, addressed in a recent earnings call:
By bringing the content of Twitter forward and pushing the scaffolding of the language of Twitter to the background, we can increase high-quality interactions and make it more likely that new or casual users will find this service as indispensable as our existing core users do. And we took initial steps in that direction with the introduction of media forward timelines and in-line social actions in October, and we’re already starting to see early signs that those initiatives are working well.
Unlike Schiller’s, Costolo’s statement makes no specific mention of hashtags and at-replies, suggesting that Schiller may have accidentally hinted at specific targets for upgrade. While it’s not immediately clear how this disappearance would work, it’s possible that at-replies will be auto-replaced by formal Twitter names, like they are on Facebook.
Hiding the # and @ characters would be like a city burying all the electric wires and TV cables so that people can see things better.
Twitter has already done away with explicit retweets (the old RT), and streamlined the way that URLs are handled, so why wouldn’t they want to clear out the unintuitive #hashtags and @mentions?
Hiding the # and @ characters would be like a city burying all the electric wires and TV cables so that people can see things better.
And this is coming from the guy that coined the term hashtag. I would be happy to drop the hash and just have real tags. But we still need to be able to tag tweets, even if we won’t be wasting characters with them. And by making tags real metadata, Twitter may finally get around to treating them as something more than just a # and a string of characters.
Here’s a picture of the explicit @mention being suppressed in an Android experimental app:
Benjamin Zimmer and Charles Carson of the American Dialect Society have released Among The New Words (American Speech, Vol. 88, No. 1, Spring 2013), and which lays down the lineaments of the invention and naming of the now mainstream hashtag.
Benjamin Zimmer, Charles Carson, Among The New Words
Hashtag represented the third time in the last four years that the WoTY winner has come from the world of technology: tweet was selected for 2009 and app for 2010. The convention of the hashtag (a word or phrase pre- ceded by the hash symbol #) has, since its inception on Twitter in 2007, grown into a powerful organizer for online talk. but beyond simply keeping track of the flow of conversation, the hashtag has been applied to a multitude of other discursive functions, including self-mockery (Ben Zimmer, “#languagenerdalert: A New Tool for Self-Deprecation,” Boston Globe, Sept. 25, 2011, K2).
Befitting such a flexible conversational tool, the term hashtag itself has flourished with multiple meanings. Whereas it has typically referred to a string of characters prefixed by the hash symbol, now it often gets used for the symbol itself. And it has crept into oral use, in a spoken equivalent of the Twitter convention, as in “hashtag fail” or “hashtag Yolo.” Such oral hashtagging, while linguistically notable, does not appear destined for success if it remains a self-conscious callback to the online form. overuse for marketing purposes does little to help the term’s long-term fortunes, either: while hosting the 2013 Grammys, ll Cool j took the trend to absurd levels by using hashtag six times in 20 seconds (“i’ve been backstage reading all your tweets about hashtag Grammys… we’re going to see hashtag Carrie Underwood, hashtag jack White, hashtag Kelly Clarkson, hashtag bruno mars, and hashtag Sting”).
Hashtag also illustrates how even very new terms are prey to faulty or incomplete memories about their origins. While Chris Messina is recognized as the “hashtag godfather” for first proposing the convention in online discussions among early adopters of Twitter in August 2007, he did not come up with the term hashtag himself. Messina made the suggestion of using the hash symbol in a tweet on August 23; two days later, he followed up with a post on his FactoryCity blog in which he dubbed his creation “channel tags” or “tag channels.” Another participant in these early discussions, Stowe Boyd, tweeted his approval on August 25: “I support the hash tag convention.” After that first attributive usage, Boyd continued using the term hash tag (written as two words) on his blog on August 26. Messina and Boyd did not fully recall this sequence of events when asked about it on Twitter. Fortunately, it is possible to reconstruct such cyberhistory thanks to archived tweets and blog posts.
It’s worthwhile to note that others — like Anil Dash — did recall the sequence of events involved in the reuse of IRC chat tags in Twitter by Chris Messina, and my calling it a hashtag.
It’s also worth noting that Messina and I were involved at the time in a Twitter and blog-based discussion about membership in groups versus what I call ‘groupings’. In a post at the time, Messina specifically referenced my writings on groupings. A grouping is a collection of people that share attributes in common, like the use of a certain tag on their blog posts, or frequent a certain bar, or study a particular martial art. So the set of people that have written about ‘Hemingway’ are a grouping, or those that have tagged a photo ‘Golden Gate Bridge’. So these people form a sort of a group, but one they weren’t invited to: their own actions make them a member of the grouping.
My argument then (see Hash Tags = Twitter Groupings) was that tags in Twitter would wind up being used in the same way: an indication of the nature of the contents of the tweet. But in aggregate, everyone that uses the hashtag ‘TEDx’ are a grouping, and that self-selected identification could be an important cultural marker. And that’s pretty much what happened.
Chris was trying to suggest that tags could be used to direct tweets to specific defined groups of people (‘channels’), more like Google+ Circles than the way that hashtags are generally used. But he is the guy that started the convention, and I completely forgot that I named it, until Ben Zimmer’s research turned it up.
I also coined the term ‘microsyntax’ to represent the use of special characters, acronyms and keywords in Twitter and elsewhere as syntactic markers (like ‘#’, ‘@’, and ‘RT’), but that term hasn’t become widely used.
John Brownlee, The Unlikely Evolution Of The @ Symbol
I have been using a convention on Twitter for the past few months, a bit of microsyntax, and I guess I should spell out my intent.
When I am writing a tweet that is principally focused on me, as a person, I enclose the tweets in parentheses, like this:
(moves desk to standing orientation)
This is intended as information of a personal-but-not-private nature, the sort of thing that might be interesting for those following me as an individual, as opposed to me as a public figure. And yes, I am a public figure, and you are too.
My point in this post isn’t a rehash of the privacy/publicy argumentarium, but just a mild advocacy for the use of the parens for these asides. And the hope that Twitter clients in the near future would allow people to dial in/out the asides intended for the inner circle.
Behind this is the groupings concept: you don’t have to be invited to be a member of my inner circle. It’s not — and really shouldn’t be — controlled by me. It’s a decision that others make: how much does this other person matter to you? If you like someone enough to know when they raise or lower their standup/sitdown desk, or what the weather is from their window, or what kind of sandwich they ordered today, then tune in to their asides.
And those who want to filter out that stuff: please do. I am creating enough social exhaust without it, I am sure.
So: please start using asides, and maybe we can get Twitter clients — maybe even Twitter — to support them, just like they did @mentions, retweets, and #hashtags.
I was involved in a twitter thread today with Ben Zimmer, who is a well-known lexicographer, and chair of the New Words Committee of the American Dialect Society. He has been researching the Twitter hashtag, which was recently selected as Word Of The Year:
In its 23rd annual words of the year vote, the American Dialect Society voted “hashtag” as the word of the year for 2012. Hashtag refers to the practice used on Twitter for marking topics or making commentary by means of a hash symbol (#) followed by a word or phrase.
Presiding at the Jan. 4 voting session were ADS Executive Secretary Allan Metcalf of MacMurray College, and Ben Zimmer, chair of the New Words Committee of the American Dialect Society and executive producer of the Visual Thesaurus and Vocabulary.com. Zimmer is also a language columnist for the Boston Globe.
“This was the year when the hashtag became a ubiquitous phenomenon in online talk,” Zimmer said. “In the Twittersphere and elsewhere, hashtags have created instant social trends, spreading bite-sized viral messages on topics ranging from politics to pop culture.”
Word of the Year is interpreted in its broader sense as “vocabulary item” — not just words but phrases. The words or phrases do not have to be brand-new, but they have to be newly prominent or notable in the past year. The vote is the longest-running such vote anywhere, the only one not tied to commercial interests, and the word-of-the-year event up to which all others lead. It is fully informed by the members’ expertise in the study of words, but it is far from a solemn occasion. Members in the 124-year-old organization include linguists, lexicographers, etymologists, grammarians, historians, researchers, writers, editors, students, and independent scholars. In conducting the vote, they act in fun and do not pretend to be officially inducting words into the English language. Instead they are highlighting that language change is normal, ongoing, and entertaining.
One interesting wrinkle is that Zimmer contends that I was the first to use the term ‘hashtag’ back in a post on 26 August 2007. My use was a response to Chris Messina’s proposal for so-called Twitter ‘channels’, which had the form of hashtags today (like ‘#hashtag’), but apparently I was the first to use the term hashtag to denote them. I also coined the term ‘microsyntax’ to represent the developing use of symbols — like ‘@mentions’, ‘#hashtags’, ‘RT”, ‘$ticker’ — in Twitter and related apps. (I still haven’t been successful in getting '/geotags' implemented.)
I didn’t even have that post up on my blog. I moved my blog several times since 2007, from Typepad (where it was called /Message), to Squarespace, and then to Tumblr. And I hadn’t reposted all the older posts, since it has to be done manually. I reposted that piece today, copying the text from the Wayback Machine.
No one should be surprised that Twitter has decided to colonize the microsyntactic space that stock tickers ($AAPL) have been playing on Twitter. Howard Lindzon may be expressing displeasure since it steps on the toes of Stocktwits, but it shouldn’t be surprising.
Twitter has at long last tried to make some money out of hashtags, which they basically ignored for years. And they reworked retweets to simplify their internal architectural problems. And of course, long before that, they decided that the @mention was a good thing, and pulled in down into the infrastructure.
Every exploitable bit of microsyntax we, the users, dream up they will take and run with. More power to them.
Here’s a few from the microsyntax archive that I’d like them to start using:
Place tags —
@stoweboyd: I just landed in /Montreal/ and I am looking forward to some smoked meat
@stoweboyd: Standing outside /St Paul Hotel, Montreal/ waiting for friends headed to dinner
@stoweboyd: In October, I’m speaking in /Amsterdam/ and /Brighton UK/
— intended to structure the discussion about locale, especially helpful in the last example because GPS or other location sensing doesn’t help.
If Twitter’s becoming a media company, they could route appropriate tweets or other ads/offerings my way in all three of the cases above.
Or ‘twhich’ tags, or twiches, where people can ask questions, or poll people, as @f does here with his friends @a @b and @c:
Hey @a @b @c chinese food at 7? yes no maybe
to which the recipients could respond
@a: Hey @f @b @c chinese food at 7? yes[x]
@b: Hey @f chinese food at 7? maybe[perhaps 8pm?]
Or a twhich could be more time oriented or choice oriented
@a: Chinese food tonight? @b @c 7pm 8pm joes wongs
with these kind of answers:
@b: Chinese food tonight? 8pm[x] joes[1st] wongs[2nd]
@c: I’ll eat anywhere: Chinese food tonight? 7pm[x] 8pm[x]
Or more open-ended questions thrown out to your followers:
@stoweboyd: How do you like the idea of twhiches? cool dumb I’d never do it
And all sorts of support could be provided by appropriately aggregating the thread of the discussion with a summary, like how many responded, are coming, alternative suggestions, etc.
Obviously, there are media and recommendation options available in twhiches and place tags. So we should just expect that if services started to support them — the way that Stocktwits supports cashtags — then sooner or later twitter will come rolling in too.
And I have a few other bit of microsyntax they could make money from… like geomessages:
@stoweboyd: Dear @/Amsterdam/ I’m speaking there in Oct at @rtecheurope and looking forward to it!
Perhaps to send a message to all my current followers that are based in Amsterdam, or a sponsored message to non-followers in Amsterdam? (In the latter case, I would have to have a credit card on file, they’d have to send me a direct message with the price, etc.)
So, if the folks at Twitter want to use these ideas, fine. I willing to talk, too.
Update: 4:12pm — Jamie Holzhuter offers this:
@holzhuter: @stoweboyd /location/ microsyntax? yes[x] (yes, please)
I have proposed a microsyntax for sending and receiving structured Twitter messages during and relating to disasters. See the emergency+codes tag for all discussion.
Why Not Hashtags?
One of the problems with microsyntax based on hashtags is that hashtags are words in specific languages, so there is an immediate divergence in this case with English and French, and perhaps Creole, as well?
This is countered by the creation of a second glossary of hashtags in French, but the equivalence is not immediately obvious.
The second problem is that people aren’t using the templates as defined. For example, “#name American & UF Alumni Lee Strickland is stuck there alive’ does have a name in it, but it’s buried. To use a simple metric, a stupid program wouldn’t be able to extract ‘Lee Strickland’ from that.
I think that a few other approaches could work better even given the requirements that a disaster imposes:
The Bang Microsyntax
My recommendations at this point for Disaster microsyntax are these:
Here’s an example, for a hypothetical disaster, a hurricane called ‘Bette’ that has hit the eastern seaboard of the US:
!bette !@john jones: alive /wellfleet hospital/
This is an emergency message stating that John Jones is alive and is located at Wellfleet Hospital. Alternatively, the hospital could have been identified as an emergency-related organization or business, with ‘!@@wellfleet hospital’ instead of being treated as a location.
!bette @carlabreck !?@sam ying: with you?
This is directed to @carlabreck using her twitter ID, asking the status of Sam Ying, specifically whether he is with her.
!bette /usps, provincetown MA/ !!food blankets: 20 people stranded here !!medevac: 1 compound fracture
This indicates a request ‘!!’ for food and blankets for 20 people stranded at the post office in Provincetown, and a request for a medevac for someone with a compound fracture.
Note that this message could be jumbled in different ways — !bette !!medevac: 1 compound fracture /usps, provincetown MA/ !!food blankets: 20 people stranded here — and it would still have the same meaning.
!bette /usps, provincetown MA/ !@hassan haque: compound fracture of the lower right leg
This is an accompanying message to the previous, indicating the name of the person with the compound fracture.
!bette /home depot, hyannisport/: roof has blown off the main building and is blocking Main Street www.sto.ly/8797gd
This is an informational post, identifying a hazard so that authorities monitoring might do something.
Getting Into Circulation
I am open to working with other groups interested in implementing tools and techniques to circulate this microsyntax for emergency messaging, or something like it.
My initial reaction to the announcement that Twitter was going to implement the retweet (RT) microsyntax as a basic function of the platform was shock (see Project Retweet: When Ultrastructure Becomes Infrastructure). Retweets are created in a variety of ways — most importantly in a vague and imprecise way, with comments added — that I thought that nearly any attempt to pin the semantics of retweeting down would fail.
I am among the folks fooling with the new implementation of RT, and it works as advertised. Which means that I don’t think it adds much to the user experience. And it specifically does not allow me to add a comment to a tweet that I am retweeting, which is bad. Clicking the retweet never leads to a text experience with anything editable, so the richness of the RT experience is being drastically curtailed.
I guess there is some performance rationale for implementing RTs this way, since the initial tweet can’t be altered, and 100 retweets to the same initial tweet don’t create a hundred copies, but just a hundred pointers. This may be an implementation that is inherently better for tracking references, for example, but is also inherently worse with regard to communication between people.
I would like to see Twitter implement the RE microsyntax (see A Useful Bit Of Microsyntax: RE), as part of the RT overhaul. A RE — as implemented currently in Tweetdeck — is like a RT, but doesn’t copy the text of the message. Instead, it has a pointer to the original message (in Tweetdeck this is encoded as a short URL), and then the remaining characters are available for making a comment:
RE http://bit.ly/892d6 I disagree with @gregarious about daylight savings time
Of course, within Twitter, the URL would not have to be exposed and could be part of an internal implementation. But the idea of a RE could be paired with the new retweet as a second tweet, associated with the retweeted message, and then displayed as a pair, like this:
gregarious I really like when the clocks change
about 2 hours ago from Twitter
Retweeted by you
stoweboyd: I disagree with @gregarious about daylight savings time
Perhaps more interesting would be if a number of other folks retweeted @gregarious’ post as well:
gregarious I really like when the clocks change
about 2 hours ago from Twitter
Retweeted by 3
stoweboyd: I disagree with @gregarious about daylight savings time
themaria: he sleeps all day anyway
brianthatcher: I never come out in the light of day
I don’t hold out much hope that Twitter will be stop, and to take ideas like these and incorporate into the current plans for retweet. They have gone a long way down a cul de sac, I fear, that will in the long run decrease the richness of our interactions on Twitter. Perhaps the howling from the user community when this is rolled out will get them to reconsider it.
My prediction is that people will revert to manually copying and pasting messages when they want to do something more than just a blind retweet. So we will have two contending sorts of RTs — classic RT v new RT — until the Twitter folks get around to implementing something like RE.
The Twitter retweet convention — where a user copies the text of another’s tweet, prefixes an ‘RT @another’ on the front, and then posts this amended copy — has become commonplace, widely supported by Twitter clients as a single mouseclick. However, Twitter has resisted making this convention part of the Twitter platform as a core primitive.
Note that ‘@username’ was at one time a convention, used to draw a specific user’s attention to a tweet, but the Twitter folks quickly saw the utility of that as a means of communication, so they rapidly incorporated that into the platform’s design. And retweets are becoming a metric used to determine the relative influence of twitterers, along with follower count, more or less playing the role that links play in the blogosphere.
Now, they have decided it is time for the microsyntactic convention of ‘RT’ to join ‘@’ in the infrastructure:
Project Retweet: Phase One - Biz Stone
Some of Twitter’s best features are emergent: people inventing simple but creative ways to share, discover, and communicate. One such convention is retweeting. When you want to call more attention to a particular tweet, you copy/paste it as your own, reference the original author with an @mention, and finally, indicate that it’s a retweet. The process works although it’s a bit cumbersome and not everyone knows about it.
Retweeting is a great example of Twitter teaching us what it wants to be. The open exchange of information can have a positive global impact and the more efficient dissemination of information across the entire Twitter ecosystem is something we very much want to support. That’s why we’re planning to formalize retweeting by officially adding it to our platform and Twitter.com.
Biz goes on to state that they will begin tooling an API with RT capabilities included, so that developers of Twitter applications can start to think about integration with this new capability.
Most important, Biz spells out how the user experience is supposed to work, and clarifying that RT will continue to work much as it has in the past. In other words, if you are following me and I retweet something from my friend @gregarious then you would see the retweet even if you aren’t following @gregarious yourself.
Why is this important? Because it supports the convention as it has emerged. It would be easy for Twitter to coopt the convention and make it work in a different fashion. In particular, they could have adopted the same model of visibility that they implemented in their last reworking of ‘@’ replies (or mentions), which is widely referred to as #fixreplies. In that case, the Twitter team decided that the way that ‘@’ had been working was annoying. In the older approach, if you were following me and I replied to my friend @gregarious then you would see the comment I made to him even if you weren’t following him. They changed these semantics so that you would only see the comment to @gregarious if you were following him too.
The logic of this is that some part of the Twitter community expressed the view that this was creating a noisy and confusing stream, because a lot of tweets would be directed to people they weren’t following, so they would only see half of these conversations.
Others — me included — suggested that the convention had emerged in the way that it had, organically, and it has many interesting properties, not the least of which is that users would see part of a conversation and then might visit the other half’s Twitter page to see the rest, and then perhaps start following that other person. When the half conversations are taken away, you might never know that the other half was there at all.
Twitter might have made this some sort of opt-in or opt-out feature, but they argued they could not think of a way to do so efficiently. They have waffled, suggesting they might revisit this in the future, but it is all very opaque.
It’s worth commenting that various counter-conventions have emerged to circumvent the restricted semantics of “@’, which is why you see ‘.@’ used widely now. Twitter does not recognize this as a @reply, so if I post ‘.@gregarious wake up!’ all my followers get to see it, and know that @gregarious has overslept.
However, retweets won’t look the same: the ‘RT’ and the ‘@username’ will not appear in the text portion of the retweet. These will become metadata, displayed in some other fashion. On one hand, this helps a great deal with the text squish problem, where adding ‘RT @gregarious’ at the front of a message may lead to the original message no longer fitting. On the other, though, the retweets might not stand out as they do presently. Designers of Twitter clients will simply have this as another design challenge, and I am sure that they will rise to the occasion.
And the last, and most serious issue in this reimplementation of retweets: the comments that people put on retweets won’t be supported. This is a serious shift away from the everyday convention:
RT @gregarious My head is killing me. | Yeah, take an aspirin.
where the text after the ‘|’ is a comment left by me. Other ways of offsetting comments are in use, like ‘<—’, ‘«’, and so on.
Twitter’s aim is architectural in an obvious way: they want to simply point at the original tweet from every retweet instead of creating 5000 slightly mangled copies.
But they are going to have to extend their implementation to allow users to leave an (optional) comment. Perhaps a second 140 character tweet twinned with the retweet? The outcome of this would be interesting since the comments added to the original tweet could be aggregated, like the chat in Friendfeed, although participation in the chat would still be distributed.
At any rate, this is the rub: if Twitter takes away our ability to comment on the retweets, people will start running around outside the implementation to get back the capabilities that have been taken away.
When the company behind a platform decides to take a user convention, like ‘RT’ and implement it as part of the infrastructure it is pulled from the ultrastucture, where the users live and invent. If the implementation doesn’t fit the contours of use that have become convention then there is a misfit. This misfit could be a gap, where less than the convention has been implemented. And a gap like that leads to a thunderclap, just like the vacuum caused by a lightning bolt creates a vacuum in the sky, and the surrounding air moves in to fill it.
I am hoping that the feedback from the community and the client developers will lead them toward a solution to the retweet comment problem before it actually surfaces. Biz stated that this would be phase one of Project Retweet, but I hope they aren’t planning to defer the comment issue to phase two.
[Other analysis: Jennifer Van Grove does a good job laying out a wish list for various sorts of filtered timelines based on a systemic RT, as well as the downsides of the new RT.]
The new release of Tweedeck was released yesterday, and Iain Dodsworth and crew implemented (along with a long list of other new features and improvements) a small bit of microsyntax that he and I discussed a month or so ago: RE.
RE is based on the everyday notion of a RE — ‘in reference to’, or ‘with regard to’ — that business folks have used for decades in memos and email.
In the Twitter context, RE is similar to and complements RT. RT, as we all know, makes a copy of a tweet, and creates a ‘@username’ reference to the author of the original tweet, and prepends a ‘RT’ indicator at the start. RE, on the other hand, creates a URL that points to the original tweet, and does not copy any of the original tweet, and prepends ‘RE’ at the start.
The idea is that the user creates a RE with the intention of commenting on whatever the original tweet contained. Imagine that my pal @gregarious has tweeted something like this:
I think Jeff Pulver’s #140conf sounds great. Wish I were there.
and I might RE in this way
RE http://bit.ly @gregarious is one of many who wishes they could have attended the #140conf this week