It's been two days since unveiling Twitter Data to a wide audience, and I have to say that I'm feeling happy at the overall reactions, both positive and negative.

The truth is that something like in-band data encoding on a platform like Twitter is controversial at best, and so to hear as much positive feedback as we are, in conjunction with intellectually honest critique, is very encouraging. This is exactly why we wanted everyone to get involved, take a stand, and help bang out something interesting and useful.

Here's a roundup of the main feedback that I've heard so far:

• Lots of light bulbs going off in people's heads about how in-band data could be useful for a lot of things. I think our simple example of voting is already pretty powerful, but we're hearing suggestions also about RSVPs, subscriptions, opt-in/out, and lots of other cool but simple ways of leveraging Twitter Data.
• Many questions about our choices of character delimiters, both "$" and ">". The questions are valid, and I'm open to comments. However, we picked these characters after quite a bit of trial and error with Twitter's search engine. If Twitter gets it into their minds that they want to assist, then there could be more options available.
• A mix of indifference, confusion, or horror at the suggestion of using namespaces in Twitter Data. It's not clear to me that namespaces will be necessary, but I'm looking at this from a long-term perspective of in-band data encoding across multiple communication channels, not just Twitter. My response at the moment is that namespaces are optional, but they are there if someone finds a way to use them effectively. Let's wait and see; I'm perfectly fine if community best practice is to avoid them.
• Twitter Data seems to making lots of connections with s(S)emantic w(W)eb people. That's perfect, since one of our interests is making Twitter a more semantically rich platform. I invite more discussion on this particular topic, and this is Jirka's particular area of interest.
• I've heard a number of comments that Twitter Data is complicated. Rereading the proposal after pushing it out, I can see how it could be interpreted that way. However, the limited amount of rigor in the proposal now shouldn't matter to most people, so I think I need to beef up the descriptions for lay people and show how it's useful and easy for everyone. I want Twitter Data to be simple for the simple things, but able to grow a little bit with platforms on which its used.
• Questions about non-English languages and how Twitter Data may or may not fit with them. I have to admit that this isn't something we put much thought toward in the rush to get it out, though Jirka is a non-native English speaker. We need to hear more feedback from non-English users of Twitter to see what would work well across a spectrum of languages.

Thanks everyone for the feedback--it's all extremely useful. Keep it coming, and please spread the word about Twitter Data by tweetting comments with the tag #twitterdata.

opening the kimono

I tweeted about the site tonight. Am watching the retweets and counting the #twitterdata votes. Let's see what happens...

version 0.3 posted

Version 0.3 is up. Changes:

• Fixed typos
• Added Why and How it works sections
• Clarified use of multiple subjects, dominant subjects, and how they relate to tuples

Version 0.2 posted

We've made some significant updates to the proposal and posted version 0.2. Changes:

• Significant adjustment to proposal around topics, now called subjects
• Added clarifying information (for example, retweeting)
• Structural changes to the document to make it easier to read

Hello world

We've posted the nascent 0.1 version of the Twitter Data proposal. Feel free to read and comment; we expect significant updates soon.