Open strategy for a new kind of business
By Jeffrey M. Barber
I’m nearing a launch of something which got me thinking about pricing. Pricing should ultimately reflect the values of the company which are in alignment with some kind of strategy. As freedom is a core company value, the best way to spread freedom is to lift all boats with rock bottom low prices. Unfortunately, low margins make for tough business. Fortunately, the infrastructure is only one aspect of the business strategy.
Deep Thoughts on Developing a Distributed Stream Query Language from the Ground Up
By Jeffrey M. Barber
As I type up my notes on the problems facing Adama, I’m thinking of how to integrate Adama within existing environments. A key problem for Adama’s future is that data held within Adama is currently very siloed. For board games and collaborative applications, this is perfectly fine. However, this can become problematic for many reasons. This is especially important for tactically migrating anything to Adama since developers need some degree of moving applications via a half-state.
Quick update on the march towards production
By Jeffrey M. Barber
Progress is being made, and things are looking good. Wanted to post an update on the march towards production
How I grew to hate PubSub; this one weird trick for building PubSub
By Jeffrey M. Barber
I’ve mentioned, in passing, that PubSub (i.e. the publisher subscriber pattern) over-commits, so I intend to dive into what I mean by this and talk about how I think about real-time services like PubSub. Having spent over half a decade as a technical leader for a very large (soft) real-time distributed system, I think it all boils down to this toy model.
A SaaS marches towards production; defining an engineering culture.
By Jeffrey M. Barber
Good news!
I’m now testing the integration of the components that I’ve been talking about by building a staging/integ environment. This new environment looks and feels like the future production site except I play fast and loose. As I move fast, I reveal problems which will be problematic, and I’m taking notes. Fortunately, most things are working as intended. My gossip is gossiping. My routing tables are routing. Deploy kind of works.
Gossip as a Failure Detector
By Jeffrey M. Barber
The other day, I sat down and wrote a gossip failure detector for the Adama SaaS that I’m building. What I love about gossip is how resilient it is to failure. Currently, the gossip failure is working in unit tests and a few small scale tests. I’m running out of innovation tokens, but I feel confident implementing a gossip detector from scratch.