To me it would be so fun to use AI to make the next version of this software. It could be done in a few months. And it would be so much better. Probably you'd want to run it off of Laravel and React Native and it would be really cool to use and easy to upgrade. I thought about just doing it myself If something doesn't get done pretty soon. It's honestly not that hard anymore.
Last edited:
I don't think it would take that long, to be honest. $5k and let Fabel at an installed version, so it can read the entire thing, understand it, then make a newer modern system... I reckon it would have it done in under a week. I reckon you could get it to convert just the frontend in under 12hrs, using the backend and its options.
Tell me you don't understand software development without telling me you don't understand software development:
Within 5 minutes you would see:
Code:
Fable 5's safeguards flagged this message. The safeguards are intentionally broad right now and may flag safe and routine coding, cybersecurity, or biology work. These measures let us bring you Mythos-level capabilities sooner, and we're working to refine them. Switched to Opus 4.8.And Opus 4.8 can be as dumb as a rock.
That it can... it does some amusing stuff at times.
Codex can handle it no problem. I'm not saying you're gonna knock it out in a week. Definitely not. But you can upgrade everything to the latest software over the course of a few months especially someone knowledgeable like you. There's no need to write code yourself anymore. I just upgraded a similar system (CRM about twice this size) over the course of about 6 months. The results are really good. You could build something that would be valuable for the years to come.
I would recommend Laravel, React and Tailwind. Move forward now while you've still got a good user base.
Just for your education..
Laravel, out of the box, takes about 30ms of time to simply do it's internal gymnastics routine before outputting hello world.
Xenforo, on the same CPU, can boot and have a page drawn in 2-3ms. Most of that time is spent connecting to a database and grabbing data. Otherwise it's damn lean.
Here is what Laravel performance looks like versus using a minimalist PHP framework. Xenforo's may be a few milligrams heavier than Zerolith, so it's a decent comparison point.
So just by switching your backend technology, out from Xenforo's you're going 10x slower for a simple request. That's before you do any real work.
How did that work for flarum? see their support form...
https://discuss.flarum.org/?q=performance
Don't even get me started on React. And Tailwind is a very bad idea for bandwidth reasons for forum software.
Use this combination of technologies and you will end up with a forum that is not feasible to run on the public internet at scale unless it happens to be funded by a fortune 500 company.
Laravel, out of the box, takes about 30ms of time to simply do it's internal gymnastics routine before outputting hello world.
Xenforo, on the same CPU, can boot and have a page drawn in 2-3ms. Most of that time is spent connecting to a database and grabbing data. Otherwise it's damn lean.
Here is what Laravel performance looks like versus using a minimalist PHP framework. Xenforo's may be a few milligrams heavier than Zerolith, so it's a decent comparison point.
So just by switching your backend technology, out from Xenforo's you're going 10x slower for a simple request. That's before you do any real work.
How did that work for flarum? see their support form...
https://discuss.flarum.org/?q=performance
Don't even get me started on React. And Tailwind is a very bad idea for bandwidth reasons for forum software.
Use this combination of technologies and you will end up with a forum that is not feasible to run on the public internet at scale unless it happens to be funded by a fortune 500 company.
Last edited:
GO backend, either PostgreSQL or MySQL, Redis, Opensearch/Elasticsearch, HTML from GO templates + TypeScript islands. Preact + TypeScript for interactive bits. Tailwind CSS. Obviously there are similar alternatives to most.
Saying all that... architecture matters more than trendy tech.
Saying all that... architecture matters more than trendy tech.
Last edited:
The weird thing is that they don't have to be bloated. There's just a lot of 'well, that's how we've always done it' or 'some guy said this is best practice' in frameworks. There's also a lot of programming practices inappropriately imported from Java that add to this bloat.
PHP is just C in a suit. Object oriented functions are slower and use more memory.
The trick of C++ 'Zero Cost Abstractions' is a very CPU-expensive compiler that, at any given opportunity, breaks your object oriented code into basic functions and variables. You pay during iteration time, not execution time. Not a bad trade.
But PHP can't do these tricks. You pay for objects in runtime. You also pay for the virtual monolith that gets built up and destroyed in frameworks like Laravel. In Java, we boot our monolith once. In PHP, not so much.
In 2020 i evaluated whether i want to keep programming in PHP. I didn't like the culture or ecosystem anymore. It all got so bloated and overly complex. This made programming very un-fun for me.
It lead me to read the writings of John Carmack and Rasmus Lerdorf, the inventor of PHP.
I formed a thesis: what if you write a framework in PHP around the underlying language? We can't really ditch object oriented entirely. But with a light OOP approach, it could be feasible to make PHP go brr.
And also while we're at it, let's apply all the John Carmack wisdom about C, if that's what we're interacting with to a good degree.
The first thing i built was a mini-profiler and debugger that would measure what the framework was doing. This way, i could could quickly prove/disprove my thesis. With this tool, i was able to optimize the core of it before anything else to see what the full potential looked like.
The performance result was, and still is hard to believe at times. And even stranger that we weren't paying for modern conveniences.
For this reason this thought experiment became something my company has been building on for the last 5 years.
Through this experience, i learned that PHP is a very cool language. If you know how to massage it, it can do a decent impression of a lower level language. You get that and a sizeable reduction of work the programmer needs to build a web app.
Go is a good language, and the web libraries for it are well developed. The concurrency features are very nice. But still, that tool doesn't fit in the hand like PHP does if you are building a web app. We now have a bunch of icky low level stuff to do, and we also lose many of PHP's superpowers. For me, it'd be hard to give up PHP because i like the have your cake / eat it too factor too much.
Thanks for coming to my TED talk.
Well, at least the AI will develop more than you guys are doing. I prefer having a potato being improved each day rather than eternal promises
Now that's funny...I can't help but picture this every time someone says this
TL;DR: PHP boomer was right, somehow.
Rasmus looked at web programming and went “this is tedious, can we make it less tedious?” Then 25 years later half the ecosystem went “great idea, what if we built a mega-cruise-ship-sized framework on top of it and summoned fourteen service providers to render a checkbox?”
One check box and fifteen beers ago.
