A little over a week ago Dalton Caldwell updated the App.Net blog for the first time in almost three years to let people know that the headless project was going to be shut down on the Ides of March due to diminishing revenues. This is the first time Dalton has communicated in any fashion to the people who continued to use the platform his company created since his State of the Union post in 2014 where he announced mass layoffs and a halting of further development. Usually when posts of this kind are made, it's often more interesting to read between the lines to glean what's not said than to take the words at face value. Interestingly, this post doesn't have anything written between the lines at all. It's a straight up statement of reality wrapped in a blanket of disingenuous words.
From the blog post:
Ample time to bake? The tech press and a lot of the most active people on the network saw the 2014 blog post as a sign of failure. People left in droves, and developers nervously continued limited development of existing applications that used the platform. Very few new applications were made after the State of the Union blog post. The phrase "ample time to bake" is a very polite way of saying "we wanted someone else to make the platform great for us".
He goes on to say:
Ultimately, we failed to overcome the chicken-and-egg issue between application developers and user adoption of those applications. We envisioned a pool of differentiated, fast-growing third-party applications would sustain the numbers needed to make the business work. Our initial developer adoption exceeded expectations, but that initial excitement didn’t ultimately translate into a big enough pool of customers for those developers. This was a foreseeable risk, but one we felt was worth taking.
Seriously? There were hundreds of people working with dozens of developers to create lots of interesting tools on the platform. The problem — if it could be called such a thing at all — was that most of the paying subscribers on the service weren't there for the platform. They were there for the social network. The proof-of-concept project that become the core use of the system. Ultimately, ADN was marketed as a Twitter alternative during the KickStarter campaign and, when people who paid good money were completely happy with that alternative, the creators of the project were disappointed and quit.
Since then, the system has carried on as best as it could. The ever-growing database of automated posts has resulted in an increasingly slower response time from the APIs. And the ever-shrinking community has resulted in an atrophy of actual usage. ADN is a shell of its former self, and any developer who might want to develop on the API would undoubtedly first look at the existing uses of the platform before investing countless hours and dollars into creating new software and services that relied on a system they'd have no direct control over. Who in their right mind would say "Oh, here's a pretty decent API1. This used to be popular, and the creators have pretty much disappeared and are unreachable, but let's see what I can do with this!"?
Nobody would say such a thing. Anybody with enough ambition to build a new tool on an a service that was abandoned by the very people who made it would grab an open API project from GitHub and host their own back-end, or use one of the many commercial APIs that are actively maintained and enhanced by people who give a darn.
Dalton & Co. did not fail to overcome the chicken-and-egg problem at all. They failed to overcome their inability to communicate their goals and ambitions. They failed to interact with the vibrant community of 27,890 active humans using the network in May of 2014. Ultimately, they failed to listen to what people wanted and work within the realities they faced.
Not All Bad
All this aside, I did enjoy a lot of my time on ADN. I met some really incredible people. I learned a lot about gender and sexuality. I learned a lot about unfair bias. I learned how important transparency is when providing a service to people who give you money. Heck, I even learned a lot about software. The current version of 10Centuries has an API that was heavily influenced by what I found while working with App.Net's API, and I've made efforts to build and improve upon what I found so that others hopefully won't shout at their screens like I occasionally did when expected datasets were not returned. And while I may not have been around on that network very much in the last year, the community will always have a place in my heart.
Nothing can last forever … though some of us certainly try.
P.S. For posterity's sake, here's a link to an image of the last three blog posts on the ADN Blog.
ADN did have a pretty decent API despite the various inconsistencies one would find when working with it.