To Twitter or not to Twitter? That is the question!


By now every developer must have read through Michael Sippey’s blog post at Twitter on upcoming changes to the Twitter API. And yes, it will affect us at EmpireAvenue.com and GamrCred.com as it will affect everyone in the Twitter ecosystem. The furore though is interesting and certainly shows that there is a need and a want for an independent, not-for-profit microblogging platform managed at scale, the question is can it gain the traction enough to ever supplant Twitter.

First, let’s examine the changes and how they affect our services at Empire Avenue and GamrCred. Authenticated requests are already the norm for both Empire Avenue and GamrCred when it comes to Twitter, so there’s not much in the way of change. However, this does seem that if we wanted to look at public data, we are now going to have to pay for tweets through one of Twitter’s providers, Gnip or Datasift. I’m on the fence on whether this is a bad thing. I would say that this change is not evil and indeed is a great revenue stream for Twitter (and its partners).

Per-Endpoint limits are also not that evil. We know from Empire Avenue and GamrCred that judicious use of endpoints, good data caching mechanisms and other techniques can actually reduce your need to go the Twitter API all that much. However, I am sceptical on the announced number of calls/hour to endpoints for the top tier users on Twitter. They will likely be too little and certainly we might not be able to analyze someone’s traffic on Twitter because of it if they are a top user. My hope is that Twitter releases some sensible per-endpoint limits such as getting your followers with a few less calls than it takes today.

The now-infamous 100,000 limit is very interesting. Personally, I have no problem in the sense that it is after all Twitter’s service. Twitter as a corporation must make money and running Twitter cannot be cheap. Here are my paranoid fears of the 100k limit requiring approval by Twitter to go higher. As a startup, we don’t know when a product may hit that 100k limit. Should we apply now for GamrCred? We obviously need to apply for Empire Avenue. What happens if Twitter sees an application that they would like to build themselves, is it then anticompetitive for Twitter to halt the growth of a startup so that they can catch up and do it themselves? I know Twitter promises to do no evil, every company promises that, but a company must first look at itself not the promises it has made in the past.

Furthermore, the 100k token limit pretty much kills any hope of Twitter connect becoming an ubiquitous login standard. But then again, Facebook Connect has pretty much won the war there.

The Display Guidelines which will be Requirements next year will probably spell the end of creating consumer services that show tweets. My biggest beef with it is the inability for GamrCred.com or EmpireAvenue.com to show tweets grouped together with updates from other networks. We currently, for example, on our latest product GamrCred, show you the things that someone has talked about about a specific game. From March onwards, we will likely need to find another way to do that for Twitter content. This actually hurts Twitter in the ecosystem even while it helps adoption of their clients. If we tell people that they will no longer be able to get as wide a reach with the tweets as we will not be able to show it along with others, means that Twitter is not as effective a method of getting your message out. Ultimately, remember that much of Twitter’s content is being produced by comparatively few accounts and those accounts are creating content to be seen, not to have private conversations.

Here’s what I see: had these limits been given to me 4 years ago, I would have thought twice about developing anything in the Twitter ecosystem, in fact I probably would never have joined as a developer even. Now, we are in a situation of Twitter having a large number of users as a network and hence, as a business, I would be stupid to not tap into that ecosystem. However, be careful, Twitter, a few wrong turns and a good competitor can kill a network faster (ahem, Myspace) than one can imagine.

The problem comes down to the following question: is Twitter a protocol or a social network? If Twitter is a social network then the connections you have made on Twitter matter and hence my developing an application to tie into that is important. The problem is that very few of my connections on Twitter aren’t also on other connections such as Empire Avenue, Facebook, Google+ or LinkedIn.

If Twitter is a protocol then it is like E-mail, we need to find someone to host a distributed microblogging service that Twitter as a company and any other service can use… like E-mail. In this situation, an entire industry will spring up supporting the idea of creating networks around a central microblogging service.

So who wants to start one? Anyone drawn up an RFC? How can it gain the adoption that Twitter has? Can it? I don’t know the answers but I am interested in hearing people’s thoughts!

(Edit: I was reminded today that identi.ca does exist and could provide this service, though my thinking goes along the lines of PubSubHubbub maybe which is supported by identi.ca and others, just needs to reach masses!)

Advertisements

2 thoughts on “To Twitter or not to Twitter? That is the question!

  1. This is a topic that Dave Winer has been writing about for months Dups …. and has been predicting the problems of Twitter dependancy for longer – he has been working on parallel ideas – is trying to launch – to get a rapid idea of some of his thinking – try here to start.

    http://scripting.com

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s