Jackson Miller

There is a huge debate going on in all corners of twitter-land. This is not the debate about The Business Model or Rails Is Slow. This is not even the debate about Decentralize Twitter. It is a debate that doesn’t make the front page of TechMeme and you won’t find much about it in the Google index. You see, the biggest debate about Twitter is an internal one.

I am sure there are a huge number of people like me who have been questioning each day whether or not that is their last day using Twitter. The downtime is killing it for me. The value of investing time and energy in Twitter is quickly dwindling (whether that value is derived from networking, promotions, or personal expression). I want Twitter to succeed, but much more importantly I want the 140 character medium to survive, and my experience has been that the most important thing about 140 characters is the connected community.

Part of me wants to wait for people to congregate at another service, but I would much rather there be a spec for multiple services to interoperate. It is going to happen, eventually.

The best possible scenario I can think of is for Twitter to go ahead and spearhead the effort. An open spec could be Twitter’s best chance at survival. Otherwise I am going to wake up one morning and the debate will be over; I will quit using Twitter.

Update: After writing this I have seen several conversations about this very thing on Twitter (during the time it has been working).

I just read Charlie O’donnell’s prediction that Twitter SMS ads are on their way. Hopefully you are having the same reaction I am. “Oh shit!”

Twitter is not just another Shiny Object. For me, and many other people using Twitter, microblogging is a new and vital communication method. It is a near-real-time, asynchronous, many-to-many medium where dynamic groups of people are able to self form or even stay un-formed and only track certain topics. Microblogging is building and deepening relationships. The potential for this communication channel isn’t even close to realized. I can think of huge uses for e-learning, community organizing, and the most common un-tapped example of emergency communication.

Microblogging is too important to let a single company control it. Currently, Twitter is a walled in garden like CompuServ or early AOL. Twitter and Pownce are like having an arpanet mail address and a darpanet mail address. Just like in the past, time will prove that these disconnected messaging platforms will not work. What is ironic here is that Twitter is all about joining disconnected communication platforms (SMS, IM, web, and email through the API and twittermail).

We need for microblogging to be based on a protocol, not a service. I don’t have any experience drafting a standard or protocol. Maybe someone with more experience can jump in and help. Basically I am thinking a microblogging protocol would look something like this:

• A domain based microblogging username. i.e. like email: jaxn@twitter.com or even more like a namespace: jaxn.twitter.com (since subdomain account are popular with many webservices)
• A DNS record-type for identifying the microblogging server(s) for a domain. This could be like MX or SRV records including priority for backup spools and load balanced servers.
• A protocol for transmitting and receiving microblog messages over TCP / UDP.
• An open source software package that provides a daemon implements the microblogging protocol.

In addition to send and receive methods, the protocol would also need methods for “follow” and “leave” at the bare minimum. The “follow” method should receive a complete address (i.e. jaxn@twitter.com) as well as a nickname (i.e. jaxn). This would preserve the shorthand while allowing for direct messaging to non-followed addresses.

The microblogging daemon would be responsible for bridgins whatever external communication protocols it wanted (for delivering messages to the accounts on that server. All server-server communications would happen with the microblogging protocol). For instance, if I wrote one it might support IM, SMS, IMAP, and speech-to-text / text-to-speech.

This would also allow for the creation of microblogging servers that allowed for “private” communications. For instance, IBM might write a microblogging daemon that integrated with the Notes IM and required approval for all “follow” requests. Universities may create microblogging daemons that delivered emergency messages to every available device regardless of user preference.

Last, but not least, a microblogging protocol would allow us to ensure that the 140 messages stay a valuable communication medium regardless of any company’s need to make a service profitable.

I would really love comments and feedback on this idea. If others think it is useful I would love to be a part of a group to create a standard protocol for this stuff.

I signed up for Twitter on March 7th, 2007. In the six months since then I have 1159 “tweets”. I have looked at other microblogging tools, but haven’t had a compelling reason to use them (aside from a handful of Facebook status updates). While I am a big fan of microblogging in general and Twitter’s distributed messaging platform in particular, I have one giant reservation that I really want to find a way around.

When I was a kid, my dad had a business card that had 3 or 4 different “electronic mail” addresses. There was the ARPANET address, the internet address, the DARPA net address. This was a necessity because not everyone had access to the same networks. Microblogging has re-created that problem. This is a problem that is bugging me because people either join Twitter or we are disconnected from a microblogging perspective. I don’t like that. Each day I think I like it a little less.

Much of the disconnected microblogging problem can be solved with the same solutions that solved disconnected email. A message delivery protocol like SMTP can easily be applied to microblogging. Addressing can easily be applied with DNS and something like MX records. The network of relationships needs a protocol that would allow it to be distributed. Maybe something like XFN or OpenID can be used to store distributed databases of relationships. There could be a RX record in DNS that said something to the effect of “here is the server(s) that handle relationship information for this domain”. Then whenever someone “subscribes” the notification can be sent the appropriate relationship manager.

I have been talking for a long time about the need for a service that ONLY provides social network management. Not a social network and photos, not a social network and bookmarks, not a social network an “walls” or “message boards”; just a social network. Now I am thinking that it may be more important that there is a protocol than there is a service. All we need is an API and a little willingness for people to play along. I would think this is something that Marc Canter would be all over since it seems to be the crux of what he wanted to achieve with People Aggregator.

Microblogging needs this relationship protocol. Many other services and products can benefit from it, but microblogging NEEDS it.

Microblogging is bigger than Twitter. Twitter has shown us a vital new communication method, the benefits of which are only just beginning to become apparent. Unfortunately, Twitter can not provide what the world needs. Emergency management services don’t need to rely on Twitter. Internal corporate communications can’t rely on Twitter. We need a way for microblogging to outgrow Twitter. I think that way is a set of specs and standards. I don’t think they are difficult specs, but they are necessary.