EmpGrid Post-Mortem (and Resurrection!?)

This post is about my pet project EmpGrid. But first a short story about how I work.

About Promises

I try to be extremely precise with promises. In fact, you could say my Promises work very much like the native JavaScript promises. Let’s look at some code:

We can note some things about that code:

  • The promiseSome(thing) bit should always complete successfully, and always return a real, concrete promise;
  • After that, the then(...) bit will be executed nearly always, i.e. typically a promise resolves successfully;
  • Only in exceptional cases, with a typically well-defined error, will the catch(...) occur.

And that is exactly how real-life promises work for me too. I want people, most notably me, to be able to count on a promise being fulfilled (given normal circumstances).

So, why is this relevant to EmpGrid? Well, I promised myself this for 2018, Q3:

Finalize EmpGrid: finish it XOR do a postmortem

So, this post is that finalization. Because I promised!

EmpGrid Post-Mortem

Since I have not fully finished EmpGrid, the only option I have to fulfill my promise is to write a post-mortem about it.

The idea behind the project remains the same since its inception. It should be a self-contained web application that shows for a group of employees (colleagues) where you can find them online. Typically, you’d be interested in all their “Presences”, ranging from social media (Twitter, LinkedIn) to version control sites (GitHub, GitLab, etc.).

The intended side-effects of the project also remained the same. The repository should be slightly over-engineered, since it should serve as a playground to test technologies useful in larger applications. At least for the server-side bits; client-side a first version could do with a very simple monolithic single page application.

If you clone the repository today, you get a mostly functional thing. You should be able to run it out of the box (given you have a modern .NET Core development setup), and do some CRUD stuff. However, it is not at all “finished”. So, I’m declaring the project dead today, Saturday, September 22nd, 2018.

Here’s what’s still sorely missing:

  • Some kind of user management. Currently, you can only use the built-in seeded admin user with default password. You should at least be able to change the admin’s password. But really you should also be able to create logins linked to Emps, so that people could edit their own details.
  • Persistence. I really wanted to step outside the default EF Core solution that .NET Core pushes you towards. Instead, I would like to see some kind of open source document database integrated. Just haven’t gotten around to it.
  • Cloud-deploy-friendliness. It should be a breeze to deploy the app as a single unit, while still giving a clone-and-run developer experience. Possibly Docker would come in to play.

In addition, there’s a bunch of smaller to do items in the readme left. But the bottom line is: it’s not finished yet.

And that’s okay! Because I learned a lot while building it. Also, the truth is that I spent my time elsewhere past months. And that was a conscious (and wise) choice.

Still, I’m a little sad to declare EmpGrid dead before it got up and running somewhere…


So perhaps I should resurrect the project?

Well, not immediately. Or, more specifically, I will not promise that it’ll be resurrected. For one, I’m due for a long, relaxing vacation to Sri Lanka. And second, I see several other cool ventures coming up (hello Advent of Code!!). So perhaps EmpGrid will stay dead.

Or maybe I need a partner in crime?

Time will tell. For now: good bye, EmpGrid. Hibernation mode: on.