9 comments

• 

  Steve Ivy

  10 months ago

  Hi Jesse,
  
  One thing you don't touch on is that @replies and #hashtags are still useful to the SMS user. A lot of Twitter's popularity early on was due to the clever integration of SMS. Any time we start talking about adding meta, non-textual content and context to microblogging, we lose the (broken, but still terribly useful) SMS integration.
  reply
• 

  jessestay

  10 months ago

  Steve,
  
  I covered that in the second to last paragraph. SMS just needs to be
  treated as another client, just like XMPP. The hashtags and intended
  recipients shouldn't be parsed as part of the original message, and
  the SMS client should parse those out and apply them as meta
  information via the API.
  reply
• 

  Marjolein Katsma

  10 months ago

  SMS is a protocol - there's no such thing as an "SMS client" that could do something like "parsing". all there is is a field to enter a message in and a way to send it on its way.
  
  All semantics needs to be inside that message - and remain there. And the same applies to any context where a message can be posted that is not a "microblogging client".
  reply
• 

  jessestay

  10 months ago

  I recognize that, and I know SMS is a protocol - it's a great protocol
  in fact that has some great uses. However, at the end of every SMS
  request (on the server) is a client that can parse SMS requests. Just
  as I mention it could easily parse out all @addresses and #hashtags,
  and leave the body of the message so only pure content is left and the
  remainder is sent as meta information back to the microblog service.
  SMS should be treated like any other client, just as XMPP would be
  treated as well.
  reply
• 

  jessestay

  3 months ago

  I view SMS as just another UI. With a proper meta backend, Twitter could
  simply interpret the SMS text, parse out the @'s and the #'s and insert them
  into the appropriate meta tag. This way focus could remain on the content,
  and 140 characters wouldn't be affected by the person in the Tweet or what
  the Tweet is being tagged with.
  reply
• 

  Marjolein Katsma

  10 months ago

  So sorry to find you still don't get it, Jesse.
  
  Twitter and identi.ca are not broken at all for embracing a useful convention: a @ to directly address a person (multiple ones in a single message) is not "clutter" at all, it's useful markup that allows any reader of a message to see at a glance who a message is addressed to - in any medium the message is rendered. Even if their interface doesn't support linking, the addressing remains clearly visible by means of this markup.
  
  That's not broken; that's useful and very powerful. Auto-linking enhances that, but remember not all contexts where a message may be read may support linking: the @ stands on its own as useful markup for the reader, including all readers who are not addressees. Replacing @ by tagging addressees loses this advantage because it doesn't "travel" with the message itself and is application-dependent. At the poster's end, having the addressing part separated from the message makes it much harder to quickly fire off a message to one or more poeple - nothing beats to type one a few phrases and hit Enter.
  
  No matter what "micro blogging clients" may be able to do - don't forget that not everyone uses a microblogging client. For instance, a Jabber client is not that - all it does is transport the *messages*. Same for SMS. And how on earth woudl I use SMS or Jabber to "tag" addresses? All I have is a field to enter a *message*, and a button or Enter to send it on its way.
  
  The basic semantics needs to be in the message itself, not outside it - and addressees are very much part of the basic semantics. For this reason, tagging also doesn't fix the ambiguity problem that is /created/ by doing away with @ to directly address someone - @ already does that perfectly and no workaround does it as well as the original and widely-used markup convention (widely-used outside microblogging for years already, don't forget that - it was simply adopted, for a good reason).
  
  I'm on the fence about hashtags though they are useful inside a message; if a message is about something already, just mention that and the meta layer won't be missed all that much in contexts that don't support this layer and have only the message content.
  
  But giving up on @ is simply very, very shortsighted: it breaks what is not broken in the first place. You need to think wider than "microblogging client" - there are way more contexts in which our messages are created and more contexts even in which they appear (don't forget search engines!); and for some of us our major UI is not a microblogging client at all.
  reply
• 

  jessestay

  10 months ago

  You and I will simply have to agree to disagree on this one. I've
  stated my points.
  reply
• 

  Stephen Paul Weber

  10 months ago

  I don't think this in broken - but by design. If you want to be sure someone gets a message (and you don't know their reading habits) txt/dm them or something. @ is good for unambiguous mentions of names, and easy replying (which can be invisibly turned into data, but typing the @ is still a nice UI)
  reply
• 

  mclaughj

  3 months ago

  I have been with you on not including meta information in the 140 character limit since the very beginning and I think the points you make in this post are valid and very applicable.
  
  If I could 'spend' my entire 140 characters on a comment I had on a blog post, and then include the "reading:http://url.com" meta tag for example, I think it would be much more powerful.
  
  But, as you mentioned on Twitter, it's not likely that Twitter will implement anything like this soon.
  reply