Summary

Screw it, I’m rebuilding Mudmap’s backend.

Highlights

• 80% feature parity between Mudmap version 1 and now
• Decided to withdraw from MSFT interviews
• Started planning new user interface enhancements

Goal Performance

A review of last months three goals. See October’s Retrospective

Mudmap feature proof-of-concept

• Appraisal: A PoC turned into a rewrite, and it’s working out great
• Rating: A

Firstly, I want to acknowledge two things.

1. Everyone says don’t rewrite, it is a time sink or that it’s chasing waterfalls
2. There is nothing wrong with Django - it’s great

That out of the way, let’s continue. This started as a prototype to see if I could use Go for sending SSH traffic more efficiently. It pretty quickly devolved into a bunch of HTTP handlers with SSH connections inside them.

Somewhere within the first half of the month, I started to feel a lot more productive and in control by re-implementing sections of Mudmap in Go. It was at this point that I realised it’s time to commit and just do it. Despite this, I did have a number of self-conscious thoughts about undertaking this task. But, ultimately they all boiled down to me caring about what others think which is an indicator that you’re doing, or not doing, something for the wrong reasons.

After a month of work, Mudmap V2 is about 80% feature parity with my in production V1 codebase. It also has more tests and better coverage than the current application and some components even have additional features. Something I’m really pleased with is how easy it is to add new features using Go - its type system and interfaces are great for this.

This is a personal thing and not slight on Django, but I’ve found that Go has forced me to think more critically about how the application should be built. Django’s beauty can also be its curse as there is a lot of magic. For a CRUD application (and especially one without DRF), it’s wonderful. For me, some of that magic didn’t work well for what I’m trying to achieve with Mudmap. For instance, my endpoints make calls to another API using SSH. This is not something Django expects and takes away the brevity of Class Based Views. I’m also not a huge fan of the serializers used by DRF (which may as well be a Django core module these days), whereas I am a big champion of Pydantic/FastAPI’s type system. Go feels like a step above both options but it’s a subject comparison.

In hindsight, Go feels like the better option for a couple of other reasons which are mostly related to the developer experience. It has made me think more critically about what I am trying to accomplish at every step along the way. How? Firstly, in Go, a little copying is better than a little dependency. This has made me do a lot more research where in the past I may have just pip install’d something, and it has an added benefit of making me read a lot more source code, a powerful educator. I also love seeing how changing a functions’ signature can ripple across the entire app, which has also made me plan further ahead. Lastly, Go is a concurrent language. Mudmap sends emails and executes things in background workers without Redis or Celery, and it can handle a boatload of requests. I’m still coming to grip with the fact that I don’t need Gunicorn, its dev server is its prod server, and it’s fast.

In all, I think this has been a good move and feel more confident about feature development and maintaining this codebase than I do with the current app.

Study at least 2 pomodoro’s each day

• Appraisal: This or rewrite Mudmap. This lost.
• Rating: C

This spawned from interviewing with Microsoft for a position within Azure. I put in a number of hours at the start of the month, working through the typical developer interview type questions. After about 10 days, I gave it away to spend more time on Mudmap but in the process did brush up on some of the basics.

I guess, this is a two-part reason as to why I stopped studying; I don’t really want that job, mostly because they want me to stay local and work in the office. That’s not something my family, or I want anymore. We’re keen to head back to the coast where we lived for ten years before coming here. Also, that crap is utterly boring and saps my energy - if it means I never work for a Big Corp, so be it.

Recommendations

We started watching The Expanse on Prime and after three seasons am totally hooked. If you like space, it’s actually really good.

I like to listen to music when working but get distracted by certain genre’s. It means I have to pick music that is rather mellow. As much as I love 90’s/2000’s hip hop, it is too stimulating for me and makes it hard to concentrate. This month I found a really chill playlist which has replaced my Interstellar and Tron soundtracks.

Wrap up

I feel like I am back this month after a sluggish August-September-October. It might be the end of lockdown, the increasingly nicer weather or just because I took some time to slow down but whatever it is I feel much more motivated.

What can I do better?

• More exercise
• More engagement with customers

What have I done well?

• Focused on completing one task well before starting the next
• Prioritised what matters and brushed off meaningless tasks

Next month’s goals

• Plug in the user authentication to Mudmap
• Email customers about upcoming changes to Mudmap

Analytics

Twitter