tStrike Twitterformat Proposal


/strike "word or phrase" from last
/strike "word or phrase" from -tweetOffset
/strike "word or phrase" from tweetID

Note: the "word or phrase" is actually enclosed within quotation marks. Single quotes may also be used. When the phrase itself contains quotes, standard escaping applies (\", \'). Single quotes within double quotes and double quotes within single quotes do not require escaping.


The tStrike Twitterformat (e.g., /strike "impossible" from last) informs your followers that you want to remove (strike out) a word or phrase from the referenced tweet (in this case, your last one). Twitter clients that implement tDelete can alert users in realtime that a tweet has been edited to have a word or phrase struck (displayed with a strikethrough).

Raison d'être

You send a tweet and need to remove a word or phrase after the fact. Perhaps there is a redundant word (a simple typo) or perhaps you want to revise what you've said in light of new information.

Your followers will be notified, in realtime, that you have edited your tweet by removing a word or phrase. Twitter clients that implement tStrike will update their timelines and display the word or phrase in the referenced tweet in strikethrough. This will impact future conversations as your followers may respond differently based on this new piece of information (e.g., by not replying to your original tweet but to the updated version).

Manual usage

The tStrike Twitterformat is simple enough to be used manually by users. e.g.,

/strike "some phrase" from last

The above communicates the user's intent to replace a word or phrase from her previous tweet with her current tweet.

/strike "word" from -2

The above communicates the user's intent to strike the word "word" from her second-to-last tweet (the tweet before her previous tweet).

/strike "another phrase" from 10242048

It is harder for the user to find the tweetID unless the client provides a mechanism that reveals it.

Implementation guide for Twitter client authors

A. User wants to strike a word or phrase from an existing tweet

1. Limited support (a)

  1. Allow the user to get the tweetID of tweets easily
  2. User writes out the tStrike command manually, as outlined in the Manual Usage section, above.

2. Limited support (b)

  1. Create an option to strike a word or phrase from a tweet (e.g., via a context menu option on the tweet)
  2. When user selects this option, create a new tweet and populate it with a semi-complete tStrike tweet. e.g., /strike "" from 10242048
  3. The user fills in the word or phrase to strike within the quotation marks and sends the tweet.

3. Full support:

Allow the user to select the word or phrase that she wants to strike and automatically create the tStrike tweet for her and tweet it.

One possible flow (for desktop apps):

  1. User selects a word or phrase in an existing tweet
  2. User presses the Strikethrough button
  3. The client creates a tStrike tweet (e.g., /strike "the selected phrase" from 10242048) and tweets it.

B. Client receives a tStrike tweet

  1. Extract the tweet reference ("last", -tweetOffset, or tweetID), and the word or phrase to strike from the tStrike tweet.
  2. If the tweet reference is "last" or in relative offset format (-tweetOffset), find the corresponding tweetID.
  3. Is the tweet with the referenced tweetID being shown in the user's timeline (i.e., is it currently cached?) If not, don't do anything and ignore the tStrike tweet.
  4. Display the word or phrase in strikethrough.
  5. (Optional) Bring this to the user's attention via a Growl notification or some other ambient alert mechanism.
  6. Don't display the tStrike tweet itself in the user's own timeline as it doesn't add any further value as a separate tweet.


This Twitterformat Proposal is released under a Creative Commons Attribution License.