DEV Community

Forem (Dev.to) is slow, so I del...optimized it.

FrancisTRᴅᴇᴠ (っ◔◡◔)っ on April 07, 2026

This is a submission for the DEV April Fools Challenge Disclaimer: This post is very very very very very very very very very very very very very ...

Read full post

EmberNoGlow • Apr 8

Excellent work. A few months ago, I was using Windows 7, and Forem seemed to work fine. Either I had plenty of RAM (8 GB), or the platform really did slow down. (Or, it's possible you have a weak computer. Although, it's hard to find one weaker than mine these days.)

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 9

Hey Ember. The platform really slowed down whenever I am on Window 7. The specs I have is usually 1 MB of RAM and I have to use a toaster box as my PC box. The motherboard I had was an Arduino hooking up on public electricity.

I could search up tutorials on how to build my PC but I saw 5 minute craft did it and it work so I tried it. Not sure what I did wrong.

Thanks! :D

EmberNoGlow • Apr 9

1... mb?! Wow..😪

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 9

Ikr

meow.hair • Apr 8

windows 95😁

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 9

Why 95 where it can be 100%? I would like Windows 100 lol

meow.hair • Apr 9

"Sorry, my reply wasn't clear. I meant Windows 95 the operating system — the old gray boxy interface from 1995. Not 95%. Just a joke about how it looks now 😄 I wish you further progress and success."🤍

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 9

Oh ik lol. Sorry if I was not specific

Harsh • Apr 8

This is the kind of deep-dive I wish more developers would do instead of just complaining about performance. 🔥

The lazy loading on user_profile_views and rating_votes was a smart catch — those aggregates can definitely become silent killers at scale.

Quick question though: Did you consider moving those counters to a separate Redis cache layer instead of lazy loading? That way you could update them async and keep the main feed queries even cleaner. Curious if you tested that approach or if Postgres optimization was the intentional first step.

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 9

Hey Harsh! I prefer moving the counter into Docker since it is faster than "loading". The reason is that I hate loading screen in websites because it takes too much space, so I tend to delete it.

Everything is too complicated now-a-days since I remember websites runs on HTML, CSS, and JS. Now, it is run on Redis, Ruby and such so I delete it. Having complicated programming languages is a skill issue since we can optimize websites with just Vanilla HTML and CSS and JS.

Thanks for reading :D

Sylwia Laskowska • Apr 7

Hahahahaha it's a masterpiece 🤩

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 7

Thanks Sylwia! Glad you enjoyed the new update! :D

BestCodes • Apr 10

"somehow miss the clues" yeah how the heck would I miss this 😭

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 10

Great question lol. I have experience where something were so obvious but for some reason the other person does not see it. Strange but gotta make sure because you never know XD

Archit Mittal • Apr 14

Performance optimization on a live platform with real users is always more interesting than theoretical benchmarks. The tricky part with Forem is the Ruby on Rails monolith - you're balancing N+1 query fixes against the ORM's convenience. Curious if you looked at adding database-level caching with something like pg_stat_statements to identify the worst offenders, or if the wins were mostly at the application layer. In my experience, the biggest Rails performance gains usually come from eager loading and fragment caching rather than architectural rewrites.

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 21 • Edited

Hey Archit! Forgot to respond to your comment!

I also removed the database along with things related to the application layer. The problem is that database takes too much space where it is much easier for us to just remember it

We rely on too much technology to the point where we got very lazy. Not only it frees up space, but it also helps us to become more productive instead of just Vibing it. If I was a Rails contributor, I would delete it too. Too many programming languages these days where we can just stick to one.

Thanks again :D

Dhruv Joshi • Apr 10

This was hilarious 😄 The “optimization by backspace” joke landed perfectly, but it also sneaks in a real dev truth: sometimes the best performance win is removing complexity, not adding more cleverness.

FrancisTRᴅᴇᴠ (っ◔◡◔)っ • Apr 10

It is true! The best performance is based on simple solution instead of complexity. The simpler the better. It's also good for readability as well, so it's a win-win!

Thanks for reading Dhruv!