A Shell For Twitter: Shhh

I wrote this idea up a few years ago in a Posterous blog I no longer use. I got email today about the final shutdown of Posterous, so I am posting here to a/ not lose it and b/ because maybe the folks at Twitter might implement all or part of it.

A Shell For Twitter: Shhh

This is a sketch of some ideas surrounding the premise of a shell — a command line interface with some Unix shell-like capabilities — for Twitter.

The Case For A Command Line

There are a lot of things within Twitter that are difficult to do via the Twitter UI, or even through the somewhat better interface that clients support.

For example, updating the location field in Twitter is too annoying, so no one does it much. And the provisioning of lists takes a bunch of clicks to add one person to a list. And the opportunities of doing things dynamically are completely blocked in a very static Twitter world, which I will touch on last.

First, imagine that the Shhh would be based on a command line premise of ‘!application {argument}’, using ‘!’ as microsyntax to indicate the tweet is a call to one of the defined apps:

!follow @username @username2

This would generate whatever series of Twitter api calls would lead to the two accounts being followed.

!list @username @username2 /listname /listname2

This would lead each of the two accounts being placed in both the named lists.

!location San Franciso, CA

Would set the user’s location appropriately.

Group DMs

It would be nice to be able to send the same direct message to more than one person:

!dm @user @user2 @user3 when do you want to have the conference call?

which would create multiple messages, one to each of the users listed

!dm /listname when do you want to have the conference call?

which would use the members of a list as the users to send direct messages to.

Dynamics: Shaped Lists

Users don’t change their configurations in a dramatic fashion in Twitter, but if tools did the work, they might. If I could easily manipulate lists, and follow and unfollow all the people on lists, that might change.

!follow @w2e/twitterers

This would lead to following all the people in the @w2e/twitterers list, perhaps a group of the most interesting commenters at the conference. This would also allow the unfollowing at the end of the conference, too.

!unfollow @w2e/twitterers

Now, I might have been following some of those people from other lists, so Shhh can see if that’s the case. We could also do this like so:

!unfollow @w2e/twitterers -follow

Which can be read as ‘unfollow all the people in the @w2etwitters list, except those that I am following independently of that list. To perform this sort of thing, Shhh has to manage a history of who is being followed independently or by lists.

One last sort of dynamics: dynamic lists. Imagine I would like to define a list that is totally controlled by Shhh, and where those in the list can change based on other factors. For example, imagine a list associated with the news around some breaking event, like the killings at Fort Hood, or a manhattan earthquake (/NYCeq). Such a list could be manually curated, but we could look for retweetings, hashtags, and geolocational information, too. For this we would use the ‘shape’ command:

!shape @stoweboyd/NYCeq -open -location ‘New York, NY’ ‘10 miles’ -watch #nyceq #quake #manhattanquake -size 50

This command reads ‘shape a dynamic list called @stoweboyd/NYCeq, keep it refreshed based on new information, focus on tweets geolocated within 10 miles of NYC, watch for tags #nyceq #quake #manhattanquake, and select the best 50 to be in the list. Every cycle that this shape command reassesses what it has access to, the list of @usernames would change. The shape command would have some adaptive technique — like following retweets and the best twitterers, to build ratings for what are the best sources.

Sources might be manually added to a shaped list, too.

Obviously, this could be directly implemented by Twitter. Optionally it could be an external service that eavesdrops (benignly) to users’ Twitter streams, and undertakes the commands via the Twitter API. 

I did float these ideas by a VC in a haphazard way, but I never pursued actively. My blessings on anyone who wants to run with it. Keep me informed if you do.

26 April 2013 4 Permalink

#efermr Makes Tweets Ephemeral

#efemr is a service that makes Tweets time-limited.  After signing up for the service — which needs access to your account to remove the time-limited tweets — efermer looks for tweets from you that have a time stamp in a hashtag — #1h (1 hour) or #5m (5 minutes) — and then deletes the tweet at the appropriate time. efermer retains the deleted tweets so that you can report, of archive.

23 April 2013 8 Permalink

I live in Tumblr now, and commute to Twitter

Over the past several years Twitter has shifted from my neighborhood to being a tool I ‘use’. I live in Tumblr now, and commute to Twitter.

— Stowe Boyd (@stoweboyd) April 6, 2013

6 April 2013 14 Permalink

Google Reader Is Deadpooled

Google has announced the deadpooling of Google Reader:

Urs Hölzle, Official Blog: A second spring of cleaning

We launched Google Reader in 2005 in an effort to make it easy for people to discover and keep tabs on their favorite websites. While the product has a loyal following, over the years usage has declined. So, on July 1, 2013, we will retire Google Reader. Users and developers interested in RSS alternatives can export their data, including their subscriptions, with Google Takeout over the course of the next four months.

Google Reader established — along with Bloglines, shut down in 2010 — the paradigm of ‘blog post inbox’.

I never liked being a ‘RSS readerer’, stuck in an always overfilled inbox of posts shouting ‘Read me!’ As a result, although I have a Google Reader account, I haven’t been there in years. 

The biggest alternatives that have pulled me away from any consideration or active use of RSS have been Twitter and Tumblr.

“Twitter ushered in the shift to our social networks as ‘engines of meaning’ and a rejection of the mechanistic, steampunk, media plumbing model.”

Twitter has become my primary source of linkage. The big shift here has been the transition to following people as curators, instead of following blogs. Twitter ushered in the shift to our social networks as ‘engines of meaning’ and a rejection of the mechanistic, steampunk, media plumbing model. And RSS is going, along with it.

For a long time I intentionally limited my Twitter follows to under 500, then under 1000, and I tried to step into the stream frequently to keep my feet wet. I adopted the ‘let it stream’ mindset, which means that I never felt it necessary to read all tweets, anymore than its necessary to look out of all the windows in my house.

Flipboard has shifted my Twitter use drastically. I now rely on a Flipboarded experience of Twitter, and hardly ever directly read my incoming Twitter stream, and so I can let the number of folks that I follow rise past the point of productively ‘following’. I respond to stuff that floats to the surface at Flipboard’s Twitter experience, but not the stuff below that. I also use the meager 20 lists that Twitter allows me for topical searching for things from the inner circle of contacts. And my Twitter client is open on ‘interactions’ all day. 

[Why doesn’t Twitter buy Flipboard, by the way?]

The second — and just as profound — transition away from RSS and from the archaic ‘blogosphere’ for me has been Tumblr. Tumblr’s open follower model creates a user experience that is at least ten times better — maybe 100 times better — than reading blog posts in an RSS reader. While I don’t agree with all the user experience decisions of Tumblr, the overall experience is awesome. The addition in recent years of curated ‘topics’ has ratcheted up the value to me, considerably. I just wish that there was a general mechanism for ‘following’ non-Tumblr information sources in Tumblr. Tumblr had — once upon a time — an RSS import capability, but it just doesn’t work reliably (see Fossilized Tumblr Feature: Importing Via RSS). Still, most media companies have taken it upon themselves to create a Tumblr account and to post abridged ’tumbles’ of stories on their official sites, so an informal, tumbled social fabric covers most everything I care to connect to. 

So, RSS is being phased out through disuse, and the dominance of the open follower motif embedded in social networks. The old RSS steampunk model is going away, and feels as old as the pneumatic tubes in Terry Gilliams’ Brazil.

Source: googleblog.blogspot.com

14 March 2013 15 Permalink

The Naming Of Things, Like #nemo

The Weather Channel is the one behind the naming of Nemo. It’s not a US Weather service provided name.

Brian Stelter, Winter Storm’s Name Means Very Little

Many reporters and weather experts rolled their eyes at the name, just as they did when the channel’s storm-naming plan was announced in October. The common criticism is that it is a marketing ploy. The National Weather Service seems to agree; it has advised its forecasters not to follow the channel’s lead, and a spokesman said it had never named winter storms and had no plans to do so. (The New York Times advises reporters not to use the names in storm coverage.)

But the name game was catching on, as evidenced by the government officials, news media outlets and airlines that published advisories using the name. “We’re ready for Nemo,” the Twitter account for Mayor Michael R. Bloomberg of New York, asserted on Thursday before listing all of the snow-removal tools at the city’s disposal.

Viewers and Web users seem to be playing along, too. Nemo was one of the top nationwide trends on Twitter on Friday.

“The fact is that Twitter needs a hashtag,” said Bryan Norcross, the Weather Channel meteorologist who helped conceive the storm-naming system. The main rationale for naming, he said, is to help raise awareness about the dangers of storms.

The name Nemo in Latin means “no one” or “no man.” Mr. Norcross said that derivation, not “Finding Nemo,” was part of the inspiration for the name, along with the Jules Verne character Captain Nemo from “Twenty Thousand Leagues Under the Sea.”

For the record, the channel’s next names are Orko, Plato and Q. On Friday morning, The Weather Channel declared that Orko had been born: it will affect North Dakota, South Dakota, Nebraska and Minnesota this weekend.

In Genesis, Jehovah gave to Adam the power to name things:

Now the LORD God had formed out of the ground all the beasts of the field and all the birds of the air. He brought them to the man to see what he would name them; and whatever the man called each living creature, that was its name.

which is a useful thing, even today, when it has lost its mystical attributes, perhaps. 

But yes, today we need good short names for every event, happening, storm, battle, genocide, trend: so we can hashtag them.

Interesting that Nemo means no one, isn’t it?

9 February 2013 9 Permalink

My first use of the term ‘hash tag’ on Twitter, 25 August 2007

I support the hash tag convention: http://tinyurl.com/2qttlb #hashtag #factoryjoe #twitter

August 25, 2007

In that tweet I used ‘hash tag’ but in the post referenced I didn’t. Still, I am betting this is the first use of the term ‘hash tag’.

I’ve informed Ben Zimmer of the American Dialect Society whose research suggested that a post of mine (Hash Tags = Twitter Groupings) from 26 Aug 2007 was the first use.

Note that Chris Messina was the guy proposing the use of the hash (‘#’) to precede a term to tag things — in fact these tweets and posts were part of an intense dicussion that he and I and others were involved in that week in August. Chris was exploring a way to create something like Google+ Circles, which he was calling ‘channels’, using the ‘#word’ approach derived from IRC chat. His goal was to use channels as a way to direct where tweets are sent — a group DM, or mailing list, if you will — which didn’t happen.

My thinking at the time was around what I call ‘groupings’: free form communities of people who share the characteristic of using a shared tag, like all the folks who have tagged something ‘social business’ is a fast and loose definition of the ‘social business’ community. My thinking was that the tag should indicate topical metadata about the tweet, and implicit metadata — grouping membership — about the tweet’s author. Which is pretty much how they are used today.

At any rate, interesting support for Ben’s research on ‘#hashtag’. 

Thanks,@DonMacAskill, for the sleuthing.

8 February 2013 7 Permalink