6 ms·

> Historically, 9% of Tweets in English hit the character limit. More like 9% of posted tweets hit the character limit. I bet a much higher number hits above t

> Historically, 9% of Tweets in English hit the character limit. More like 9% of posted tweets hit the character limit. I bet a much higher number hits above the character limit and then are trimmed down to fit. It doesn't sound like they consider that, and so it leads to it seeming like only a small number of tweets are limited by 140 characters when I'm sure a muuuch larger number of tweets could benefit from using "you" instead of "u" and such.

jimktrains2 9y ago
Or just broken into 2+ tweets.
- mortenjorck 9y ago
  Or broken into 10, 20, or more tweets. Which, to me, is the disappointing thing about the new character limit: it does nothing to address this emerging usage pattern, one might even say emerging form of communication, of the tweetstorm. It seems pretty clear that tweetstorms enable a certain mode of marshaling thoughts and communicating them in an off-the-cuff way that blogging services never identified or understood. And yet Twitter's product org doesn’t seem to understand it either.
  - microcolonel 9y ago
    Maybe something intermediate could be done, but I think the tweetstorm format allows for interjections at fixed intervals, which can be helpful.
    - viraptor 9y ago
      Didn't really think of this before, but yes - this is very useful and it would be a shame to lose this feature.
      - mortenjorck 9y ago
        That’s part of the missed opportunity I see: far from losing this ability, a proper tweetstorm feature should embrace it, making it easier to navigate and interact with.
        visarga 9y ago
        > easier to navigate and interact with On the contrary, it's harder to navigate because it is intermingled with all sorts of boxes, whitespace, tags, icons and hidden menus. Everything is hair-triggered by a touch making the whole page unstable. Tweetstorms fork into secondary threads that are hard to follow, unlike HN or reddit, and use 3 screens for content that fits on half a screen.
        aaron-lebo 9y ago
        Yep, I think Tweetstorms are right up there with colds and bad dates for negative experiences. Or, better said: https://www.reddit.com/r/programming/comments/7bekig/an_interesting_rant_about_guis_as_a_programmer_i/dphdiz5/ https://www.reddit.com/r/programming/comments/7bekig/an_inte...
        digi_owl 9y ago
        For some reason Twitter had decided to attempt to make their page behave like a UI rather than a forum or similar. Thus we only get a single linear reading, and it wastes space by being presented as a floating "windows" above the main page. Never mind that it is damn hard to see the full image, as no amount of clicking will actually make it fill the whole screen. Only bringing up the tweet "window" and then right clicking the image gets anywhere close.
        rainbowmverse 9y ago
        At least they stopped requiring 100% of my processor to display them like when they first switched to the floating, easy to accidentally dismiss boxes.
    - cyphar 9y ago
      Sort of like how you can "interject" via quoting in email? ;)
      - theoh 9y ago
        That's, increasingly, apparently not understood by the average correspondent. Anything other than a continuous block of text preceding a quoted reply ("top posting") may cause confusion.
        microcolonel 9y ago
        Don't worry about it, if you are ever confronted, simply explain it. It is a tradition worth defending.
        corobo 9y ago
        I've found top-posting the type of email fixes this, no failures yet: Please see my responses inline > can we make the page do this yes np > Does the flux capacitor have enough jiggawatts to make this work? It does if we don't add anything new > etc
  - prawn 9y ago
    With the right client support (collapsed tweetstorms and stats), readers could delve as deeply into a topic as they were interested, and writers could see how far most readers cared to delve. That's seemed like a no-brainer development in Twitter for some time to me. Entire articles could be written and released in such a format.
    - digi_owl 9y ago
      This is the company that dreamed up lists, only to hide them behind multiple layers of UI rather than provide the ability to display them in columns front and center.
  - digi_owl 9y ago
    In particular as they could just as well have lifted the inline signaling that sms has for stitching multiple messages together to form a long one. After all, the start of Twitter was as a sms relay.
  - cjCamel 9y ago
    Twitter is experimenting with a feature that lets you compose tweetstorms [1] [1]: https://www.theverge.com/2017/9/10/16284526/twitter-experimenting-tweetstorm-feature https://www.theverge.com/2017/9/10/16284526/twitter-experime...
- xellisx 9y ago
  https://github.com/liuggio/sized140 https://github.com/liuggio/sized140 ...
clarkmoody 9y ago
Tweet survivorship bias.
- agumonkey 9y ago
  I'm sure they have (or could have) analysed that.
  - Johnny555 9y ago
    How do they know if someone was going to tweet a longer sentence, like for example: "Donald Trump is the worst president ever this country has ever had. He stole the election with help from the Russians and is running this country into the ground." But knowing that they only get 140 chars, they just decided to tweet "Trump sucks! :-( And he's in bed with the Russians"? You don't have to hit that 140 char limit too many times until you start self-censoring yourself and write shorter messages.
    - davedx 9y ago
      > How do they know if someone was going to tweet a longer sentence Analytics
      - marksomnian 9y ago
        Analytics != mind-reading. Unless they send every single keystroke and character deletion, they can't track that.
      - michaelmior 9y ago
        That works for users using an official Twitter client who actually typed in the tweet and then edited it. It doesn't account for people who decided to enter something shorter because they knew what they wanted to say would hit the limit.
        agumonkey 9y ago
        True, but to an extent you know how people type, start something then stop, hesitate.. an awful lot can be retrieved from an input field.
talaketu 9y ago
Do you really believe twitters product managers did not considered people might be cramming to tweet?
- sametmax 9y ago
  No, we believe that the product manager is voluntarily not talking about it.
  - squeaky-clean 9y ago
    Thank you, someone who gets my point. I'm sure Twitter are smarter than this, but why do they assume their readers lack critical thinking.
- tom_mellior 9y ago
  Twitter only used to show you tweets from people you followed, in chronological order, each tweet exactly once. Those product managers then decided to pepper your timeline with tweets you "may have missed", tweets from people you don't follow but that were liked by others and passed a randomized test, and all this out of order and with duplicates. Of course you cannot turn any of this off (except by using another than the official web client). So no, I wouldn't trust the product managers to make intelligent decisions.
  - alkonaut 9y ago
    Why would someone use any app or page from twitter though? Great alternatives have existed from the very beginning, and they are still much better than the official clients. I haven't noticed any commercial messages, any "suggested" tweets etc. Just nicely rendered chronological timeline (I use tweetbot for iOS)
    - corobo 9y ago
      Honestly for me it's because they (desktop replacements) all seem to split everything into multiple columns. I dislike that design and just can't seem to gel with it. As for mobile I've actually not had any problem with the official iOS app after setting my account to display tweets chronologically
blauditore 9y ago
The very next sentence says: > This reflects the challenge of fitting a thought into a Tweet, often resulting in lots of time spent editing and even at times abandoning Tweets before sending. The whole idea of extending the limit is based on what you describe, and this post shows that it had the desired effect, without introducing other bad things. I'm not sure what your point is.
- squeaky-clean 9y ago
  > I'm not sure what your point is. My point is I bet it is not 9% but closer to 99% of tweets that are negatively affected by 140 characters, and their way of measuring this is very flawed.
  - blauditore 9y ago
    But their test phase with 280 characters has indeed proved the opposite: Median and mode of tweet length are still roughly the same as before, judging from that graph. This number doesn't come from thin air, but was calculated based on the assumption of log-normal distribution if there weren't a limit: https://blog.twitter.com/engineering/en_us/topics/insights/2017/Our-Discovery-of-Cramming.html https://blog.twitter.com/engineering/en_us/topics/insights/2...
    - CamelCaseName 9y ago
      Perhaps the majority of longer form tweeters have already left the platform or altered their tweeting behavior?
  - IshKebab 9y ago
    Look at the graph. It clearly isn't the case.

jimktrains2 9y ago

Or just broken into 2+ tweets.

mortenjorck 9y ago
Or broken into 10, 20, or more tweets. Which, to me, is the disappointing thing about the new character limit: it does nothing to address this emerging usage pattern, one might even say emerging form of communication, of the tweetstorm. It seems pretty clear that tweetstorms enable a certain mode of marshaling thoughts and communicating them in an off-the-cuff way that blogging services never identified or understood. And yet Twitter's product org doesn’t seem to understand it either.
- microcolonel 9y ago
  Maybe something intermediate could be done, but I think the tweetstorm format allows for interjections at fixed intervals, which can be helpful.
  - viraptor 9y ago
    Didn't really think of this before, but yes - this is very useful and it would be a shame to lose this feature.
    - mortenjorck 9y ago
      That’s part of the missed opportunity I see: far from losing this ability, a proper tweetstorm feature should embrace it, making it easier to navigate and interact with.
      - visarga 9y ago
        > easier to navigate and interact with On the contrary, it's harder to navigate because it is intermingled with all sorts of boxes, whitespace, tags, icons and hidden menus. Everything is hair-triggered by a touch making the whole page unstable. Tweetstorms fork into secondary threads that are hard to follow, unlike HN or reddit, and use 3 screens for content that fits on half a screen.
        aaron-lebo 9y ago
        Yep, I think Tweetstorms are right up there with colds and bad dates for negative experiences. Or, better said: https://www.reddit.com/r/programming/comments/7bekig/an_interesting_rant_about_guis_as_a_programmer_i/dphdiz5/ https://www.reddit.com/r/programming/comments/7bekig/an_inte...
        digi_owl 9y ago
        For some reason Twitter had decided to attempt to make their page behave like a UI rather than a forum or similar. Thus we only get a single linear reading, and it wastes space by being presented as a floating "windows" above the main page. Never mind that it is damn hard to see the full image, as no amount of clicking will actually make it fill the whole screen. Only bringing up the tweet "window" and then right clicking the image gets anywhere close.
        rainbowmverse 9y ago
        At least they stopped requiring 100% of my processor to display them like when they first switched to the floating, easy to accidentally dismiss boxes.
  - cyphar 9y ago
    Sort of like how you can "interject" via quoting in email? ;)
    - theoh 9y ago
      That's, increasingly, apparently not understood by the average correspondent. Anything other than a continuous block of text preceding a quoted reply ("top posting") may cause confusion.
      - microcolonel 9y ago
        Don't worry about it, if you are ever confronted, simply explain it. It is a tradition worth defending.
      - corobo 9y ago
        I've found top-posting the type of email fixes this, no failures yet: Please see my responses inline > can we make the page do this yes np > Does the flux capacitor have enough jiggawatts to make this work? It does if we don't add anything new > etc
- prawn 9y ago
  With the right client support (collapsed tweetstorms and stats), readers could delve as deeply into a topic as they were interested, and writers could see how far most readers cared to delve. That's seemed like a no-brainer development in Twitter for some time to me. Entire articles could be written and released in such a format.
  - digi_owl 9y ago
    This is the company that dreamed up lists, only to hide them behind multiple layers of UI rather than provide the ability to display them in columns front and center.
- digi_owl 9y ago
  In particular as they could just as well have lifted the inline signaling that sms has for stitching multiple messages together to form a long one. After all, the start of Twitter was as a sms relay.
- cjCamel 9y ago
  Twitter is experimenting with a feature that lets you compose tweetstorms [1] [1]: https://www.theverge.com/2017/9/10/16284526/twitter-experimenting-tweetstorm-feature https://www.theverge.com/2017/9/10/16284526/twitter-experime...
xellisx 9y ago
https://github.com/liuggio/sized140 https://github.com/liuggio/sized140 ...

clarkmoody 9y ago

Tweet survivorship bias.

agumonkey 9y ago
I'm sure they have (or could have) analysed that.
- Johnny555 9y ago
  How do they know if someone was going to tweet a longer sentence, like for example: "Donald Trump is the worst president ever this country has ever had. He stole the election with help from the Russians and is running this country into the ground." But knowing that they only get 140 chars, they just decided to tweet "Trump sucks! :-( And he's in bed with the Russians"? You don't have to hit that 140 char limit too many times until you start self-censoring yourself and write shorter messages.
  - davedx 9y ago
    > How do they know if someone was going to tweet a longer sentence Analytics
    - marksomnian 9y ago
      Analytics != mind-reading. Unless they send every single keystroke and character deletion, they can't track that.
    - michaelmior 9y ago
      That works for users using an official Twitter client who actually typed in the tweet and then edited it. It doesn't account for people who decided to enter something shorter because they knew what they wanted to say would hit the limit.
      - agumonkey 9y ago
        True, but to an extent you know how people type, start something then stop, hesitate.. an awful lot can be retrieved from an input field.

talaketu 9y ago

Do you really believe twitters product managers did not considered people might be cramming to tweet?

sametmax 9y ago
No, we believe that the product manager is voluntarily not talking about it.
- squeaky-clean 9y ago
  Thank you, someone who gets my point. I'm sure Twitter are smarter than this, but why do they assume their readers lack critical thinking.
tom_mellior 9y ago
Twitter only used to show you tweets from people you followed, in chronological order, each tweet exactly once. Those product managers then decided to pepper your timeline with tweets you "may have missed", tweets from people you don't follow but that were liked by others and passed a randomized test, and all this out of order and with duplicates. Of course you cannot turn any of this off (except by using another than the official web client). So no, I wouldn't trust the product managers to make intelligent decisions.
- alkonaut 9y ago
  Why would someone use any app or page from twitter though? Great alternatives have existed from the very beginning, and they are still much better than the official clients. I haven't noticed any commercial messages, any "suggested" tweets etc. Just nicely rendered chronological timeline (I use tweetbot for iOS)
  - corobo 9y ago
    Honestly for me it's because they (desktop replacements) all seem to split everything into multiple columns. I dislike that design and just can't seem to gel with it. As for mobile I've actually not had any problem with the official iOS app after setting my account to display tweets chronologically

blauditore 9y ago

The very next sentence says: > This reflects the challenge of fitting a thought into a Tweet, often resulting in lots of time spent editing and even at times abandoning Tweets before sending. The whole idea of extending the limit is based on what you describe, and this post shows that it had the desired effect, without introducing other bad things. I'm not sure what your point is.

squeaky-clean 9y ago
> I'm not sure what your point is. My point is I bet it is not 9% but closer to 99% of tweets that are negatively affected by 140 characters, and their way of measuring this is very flawed.
- blauditore 9y ago
  But their test phase with 280 characters has indeed proved the opposite: Median and mode of tweet length are still roughly the same as before, judging from that graph. This number doesn't come from thin air, but was calculated based on the assumption of log-normal distribution if there weren't a limit: https://blog.twitter.com/engineering/en_us/topics/insights/2017/Our-Discovery-of-Cramming.html https://blog.twitter.com/engineering/en_us/topics/insights/2...
  - CamelCaseName 9y ago
    Perhaps the majority of longer form tweeters have already left the platform or altered their tweeting behavior?
- IshKebab 9y ago
  Look at the graph. It clearly isn't the case.