I realize now that one of the hardest parts of running a successful startup is “betting” on tech stacks that, 3 years out, will have a groundswell of community support around them.
It’s still shocking to me that when I chose each of the following technologies as a central part of Parse.ly, they were so new/immature as to not even show up on a Google search trends box, but are now very popular technologies.
These include Python’s Pandas library, d3.js, Apache Storm, Apache Kafka, and, recently, Apache Spark.
Each of these also now has several O’Reilly / Manning books and very active open source repositories with lots of outside contributors.
This is what technologists mean when they say they deal in uncertainty every day. I could have just as easily have chosen technologies that would now, in 2015, be totally unknown and/or unmaintained. There are certainly no shortage of alternatives to these technologies.