Warning: Parameter 1 to wp_default_scripts() expected to be a reference, value given in /home/public/wp-includes/plugin.php on line 571

The Future of Microblogging and/on Transit

I don’t listen to the radio (or watch TV or read newspapers much, frankly), so much thanks to John Bollwitt for passing along the news over to Twitter:

JohnBollwitt: Just heard @kenhardie say on @talk1410, 100+ twitter accounts coming to various #translink routes to follow for constant updates.

My first thought? Oh noes, fail whales! Twitter’s architecture has gotten much more robust in the past year, but over a hundred Twitter accounts? Even the MTA in New York has a mere total of 22 Twitter feeds, which are all listed at the 511Ny Twitter website.

There’s a couple of threads to my thinking here, so please bear with me. This first thread is about technical infrastructure. While I’m an enthusiastic user of both Twitter and Facebook, I think there’s a certain line to be drawn when it comes to using (or more accurately relying on) them, for functions that can easily be considered “mission critical applications.” We don’t have to look very far in the past to find instances where this can be disruptive. Not so long ago, Twitter made a controversial change to the way their users received @replies (now called mentions) to their timelines, so that I could no longer see messages mentioning people that I’m not following on Twitter. And let us not forget, that not that long ago, Twitter updates had been discontinued so that they weren’t being sent to cellphones in Canada!

Twitter’s model and community are amazing resources to take advantage of, but they’re more important to me as a proof of concept of the value of mobile and ubiquitous web applications. As I implied in my follow-up to John Bollwitt’s tweet, I think we need to ask some serious questions about whether we want our transit system’s information distribution and notification service reliant on a Silicon Valley-based startup with no readily discernible business plan. There are many smarter and more articulate people who can talk circles around me on how this is both or good bad, so I’m just going to advocate more thinking here.

There’s a second thread here, which is about TransLink’s use of Twitter more broadly. Make no mistake, I unequivocally think TransLink’s use of Twitter is an excellent, innovative and amazing thing. While Twitter is undoubtedly a huge part of our personal, and, increasingly, our professional lives, and I certainly think it’s important to go where the members of our community already are, I also fundamentally believe that the provision of the information that makes our transit system usable should not be contingent upon a rider or (Canadian) citizen’s registration with a third-party, American service.

That said, I think what I really want to ask is, is “what Twitter is for” inclusive of what TransLink wants to use it for? Is it for distributing information (yes)? Is it for sharing our thoughts in real-time with members of both our communities of interest (yes) and communities of location (yes)? How finely should we dice the feeds in order to ensure the technological resources to keep these services running are stable? Where do we cross the line from “nice to have” to “really f*&%ing important”?

If TransLink goes forward with this (which, really, I’m ready to admit I know next to nothing about at this point) I think the very least that TransLink needs to do is to be somewhat transparent about its relationship with Twitter. Are they working closely with Twitter to manage the traffic that may result? Could large amounts of traffic resulting from global Twitter usage (such as a large, social media noisy event like, oh, the Olympics or a hurricane) impact information being delivered to riders? Would our transit information service be vulnerable to DDoS attacks like the one that affected Twitter a couple weeks ago?

I would love to hear from the fellas at Handi Mobility (though they are down in SF right now), since I know they do a lot of thinking on mobile, its underlying infrastructure and end user experience, and have worked with TransLink on their SMS and mobile applications. I don’t think I’m freaking out unnecessarily, but I think it’s important to be wary of the centralization of all our short message information needs into a single application and Twitter particularly. It may save money for application developers like TransLink, and make managing feeds easy for end users, but it personally makes me a little queasy to think that a whole bunch of decisions that TransLink can make about the backend, would be made by Twitter on an ongoing basis.

Another event that speaks to this is the recent spam API incident on the TTCU_Community Twitter account (more info at TTCupdates.com). This account automatically re-tweets messages from Twitter with the hashtag #ttcu, and up until now it seems that the service has been used and maintained by community members without incident for allowing Toronto riders to share their observations about transit service. A couple days ago it all of sudden started sending out spam messages, which the developer is now looking into.

Logically, this is a scaling up of what the CMBCTransit account already does. While the prospect of getting a Twitter account for the #15 bus I ride everyday is exciting, I think I’d be much more interested in having TransLink maintain a separate service rather than relying on Twitter, which SFBART already does to some extent.

As you can see, I’m mostly asking questions in this post. If you have more questions, or any methods of answering the ones I’ve got, I’ve love to hear it in the comments and/or trackbacks.


  1. my blink reaction:

    1. there should be one twitter account per route. Each route should have an RSS feed that is updated only when there is a problem or downtime or a notice of some sort i.e. not very frequently

    2. these rss feeds should then go into twitter using one of the 8 million :-) RSS to twitter tools

    3. which means if twitter dies, we still have the RSS feeds

    Posted August 18, 2009 at 1:36 pm | Permalink
  2. Thanks for the thoughts Karen — I’ve passed them on to the web teams here.

    Posted August 18, 2009 at 3:01 pm | Permalink
  3. Cam Telford

    Excellent ideas and a thoughtful response. Roland is not too far off what is being looked at for RSS/Twitter alerts (it has always been RSS first and Twitter second we just won’t have access to the RSS for a while longer so Twitter by itself is a good stop-gap in the meantime) . The problem with the NY511 model is the amount of information one receives per feed. It is traffic and transit. If you take a look at any of the feeds, it is mostly traffic and over a pretty wide area in places. How do you seperate what is of interest to people, generally Transit or traffic, not both, and give it to them in bite sized chunks that they can do with what they want? Do we need a feed for every route, maybe not, but I think we need more than 22.

    FYI, there will also be an email/SMS subscription component for alerts and other transit related info.

    Posted August 18, 2009 at 3:27 pm | Permalink
  4. My first thought on this was that making a 200+ twitter accounts for each route is a stupid idea, inefficient for both provider and subscriber. Grouping by city or area would be seem more sensible, maybe 30 accounts tops.

    The ideal solution would be an open API of realtime transit service data, then use that to hook that up to Twitter, SMS, web, MSM, everything.

    Posted August 18, 2009 at 3:55 pm | Permalink
  5. IMO, there should be a feed for every route and a summary feed for the entire system and you should be able to subscribe to the feed by RSS, email or twitter

    Posted August 19, 2009 at 11:37 pm | Permalink
  6. Cam,

    Apologies for the late response.

    I think the part I was really interested in is what you covered in your FYI – that there is a separate, non-Twitter, SMS component for transit info. And I think it’s also important that I’m not particularly against using something profoundly Twitter-like (Boris suggested laconica, the open source Twitter, or I’d also add to Jaiku if it makes any headway) as the mechanism for distributing these messages, because I think the interfaces are every bit as important as the ownership issues, and I’d be tickled pink if that were on your radar for the SMS transit alerts service you are considering.

    All the best for you and your web team in coming to a workable and useful solution!


    Posted September 16, 2009 at 10:24 am | Permalink
  7. Thanks for including TTCupdates in your article, Karen. Apologies for being so late with a reply!

    It’s true, Twitter was never meant to support mission-critical applications. Many of the transit solutions you see on Twitter (and, unless Twitter begins to offer paid support, will see in the future) are offered by developers working in their spare time — usually with the understanding that it may break from time to time.

    As an aside, the recent spam issues with TTCu were cleared up a little while ago ;-)

    Posted October 20, 2009 at 7:37 am | Permalink

Post a Comment

Your email is never published nor shared. Required fields are marked *