A slow start is better than no start...

Progress in very slow baby steps, that nobody can see… yet.

Almost six months have gone by since my cryptic ‘something is going on over here’ post and the initial rebirth of this site, and honestly, so far, I have nothing I am ready to show off or report to the world at large.

I may have both seriously underestimated the effort needed to ‘learn a whole new tech stack and build a service from the ground up’, while also overestimating the number of hours life would allow me to devote to this project. I know this is fairly common occurence in life, in everything from home improvement projects to project cars takes twice as long as we ever expect, why would a whole new business be any different? Now is not going to beat myself up over delays, but instead, keep forging ahead.

I initially sat down with an excellent 8 hour long Youtube video on building with NextJS and NextAuth, and figured “8 hours, no problem, time well invested to learn the basics”. Because I’m very much a ‘learn by doing and building, not by reading man pages’ kind of person, I figured a video tutorial I could follow long with would get me off on the right foot. But completing that 8 hour video took me probably 3 months, because, well, Life up and decided that my desired schedule would be taking a back seat to other things, and, at the same time, as I did find some hours to invest in my learning, I kept getting distracted by wandering down various rabbitholes.

See, about 20 minutes into the tutorial, I decided to restart from the beginning and actually start building out what I’m looking to build using the video as a guide, rather than just duplicating what he does, and then tossing everything out to start ‘my platform’ and well, since I’m building ‘my baby’ as opposed to ‘his demo’… Of course I decided to make some different tech choices than he does, because again, “My baby”… For instance… I’m not a fan of the modern day trend of outsource everything I can to someone else methodology. I get that many developers don’t want the burden of running infrastructure themselves, so they pay more in costs to let someone else worry about such things. But, see, infrastructure, that’s where my roots run deepest, its the place where I’m actually the most comfortable… so when the video says “go sign up for an account with Neon.tech for your database”, what I actually did was:

• Go review the supported database list for primsa
• See MariaDB/MySQL are listed. Say “hello my old friend, good to see you again”
  • Remind myself this project was supposed to also be about getting out of my comfort zone and exploring new tech, and MySQL and I have a 20ish year relationship at this point, it hardly counts as a new experience at this point.
  • Table MariaDB as ‘the safe backup plan’
• See CockroachDB listed. Remember that some of my peers have been saying great things about it in regards to scale, and one day I hope to have ‘scale problems’ that need to be solved. So why not see if I can get ahead of the curve by building with scale in mind.
  • Build out a local CockroachDB Cluster in the homelab.
  • Do a bunch of benchmarks/testing, be disatisifed with the performance I can wrangle out of it.
• Decide to split the baby and roll with PostgreSQL. It’s something I have passable familiarity with, but is just foreign enough that mastering it enough to run a platform on will count as ‘a new experience’ for me.
• Realize I’ve spent a week of free nights on just this one small design decision, and chastize myself for letting “perfect be the enemy of good”, and promise that next time I’ll get to a ‘good enough’ answer quicker.
• Proceed in the tutorial until a step where I once again find myself going Well, yeah, I could do that… or… and start this process over again on a new topic.

So yeah, that 8 hour video was incredibly instructive and educational, and it no doubt was time very well spent. But I clearly way underestimated how long in calendar time it would take me to complete.

The good news is that starting to build out the actual platform has started the brain juices flowing, I’ve started building mental lists of you know what would be a good feature? and I bet people would like to have this feature…, and I always wanted an X that could Y… which is great, and lots of fun, but leads me into the second issue:

I’ve reached the point I need to actually get organized, figure out a platform roadmap, define my “MVP”, and start building something shippable. In short, I’ve reached that point where it all starts to feel overwhelming, yet exciting, at the same time. Everything is possible! But everthing needs to be done!

So there’s only one thing to do: Put one foot in front of the other and start moving.