-------------------------------------------- Protocols vs. Apps 2019.08.09 17:52:14 -03 -------------------------------------------- In a Telegram discussion group, people were chatting about alternatives to Slack - the chat application. I didn't much of the discussion going on, because as I explained there I'm not a Slack user and have not understood the appeal of Slack at all. The Slack users in the group presented the main advantages as being: ease of integration with other tools and a nice GUI (app) on different platforms. I come from the IRC world, I can see the value in the threaded-view in Slack, but what's left? Ease of integration? How many IRC bots and botnets are there? This is very easy to implement. So I started thinking about the second point: the app. I like IRC because it doesn't force me into any single app. It gives me the protocol and I can choose how to implement my client. But the feeling I'm getting from the Slack users is that they don't want that. They want *one* app, one *official* app, preferably available on *most common* platforms (including mobile ones). In the information overload world we live in, people don't want choices. Do I have to pick an app? No, too complicated. Now I'm let wondering if my talks about Site Reliability Engineering could be improved with this in mind. I present SRE as a protocol, as an API. It's up to you to find the tools to implement it. What if what people want is not an SRE protocol, but just one SRE app? Just a single choice: Yes or No. Maybe I need to rethink my arguments.