Serve site as a compiled binary over express? And other wild ideas...

[nathanjhood]

I can't help but feel a little weary about certain aspects of the current tech stack from a philosophical standpoint.

Service, functionality, tooling, et al. is all currently fine; in fact, rather good really.

But I can't shake the feeling of being quite deep in a closed ecosystem, where some of my more interesting and creative ideas are not always aligned with "this way of doing things".

I am a little tempted to see what can be done with noderc and some other more customary approaches to building and maintaining my online portfolio.

I have some sketchy concepts about serving read-only resources from a binary-compiled library of OOP-like structures that define web elements as objects.

I would also just like to see more of what is out there today for the benefit of my own growth and personal development. This current stack, however, is mostly geared toward supporting its' official integrations, with little scope for much else (custom server? custom build script? etc...)

Not all is lost; I can easily write any kind of string content and return it as a Napi::String, then pass the result to my NextJs components. But, such an example again leads to creating source that only functions within this ecosystem.

A much more generalized approach would yield far more useful results. But do I have so many compelling reasons to tear down all that I've built from NextJs?

It is slightly demotivating that I'm reaching limits with how much I could actually do with this stack; it all amounts to using API's and following docs per whatever logical needs arise.

That is completely suitable and probably desirable in the context of a business, but a total buzz-killer in the playground.

One big aspect of being a portfolio, is to be a place where I get to "show off" a little bit more of the things that I would not be given the opportunity to even consider in a professional environment.

Not least of all, because it is the only place where I'd have the opportunity to try out some outlandish new concept on a total whim.

I feel that I would speak highly of NextJs and Vercel all day, but if I were asked to raise my biggest grievance, it would easily be this total immersion, lacking the scope of a more "mixed" approach as is possible with Vue/Nuxt - although, considering I have a nice bit of experience on that stack already, I have no regrets in learning NextJs in my own time. As a means for reaching a certain ..."next" level in my personal development as a coder, it has certainly been all gains.

But now I am looking ahead at the work left to; currently, I feel that I'll slog through the making of a realtime-enabled blog, seperate to the GitHub-sync'd projects pages, and I have some fun ideas about placing more things under the control of users, particularly if they are authenticated and logged in.

But those are the only real ideas I have at the moment for putting more of this stack - also looking at the supabase side, here - to good and proper useage enough to justify maintaining the whole thing.

Concurrent to the above, incorporating some kind of off-the-wall ideas with NodeJs addons and C++ OOP and such forth is a bit of a non-starter, due to tooling, building, limited CI/CD capabilities, and other restrictions that are present in the Vercel/NextJs platform, and which I am quite unlikely to convince anyone is a good idea to change.

Yeah, I know, sometimes things are restricted for very good reasons that are best not messed with.

But then what have we learned?

I guess I'd prefer to be writing this in a real-time enabled blog on my page, instead of here, on GitHub. But I also wish I could incorporate some of these new trains of thought into this whole portfolio, the only home that will ever accept them no matter their behaviour 🤣 and foster them long enough to have a better life cycle.

Harsh but true; it might pay me a little to scaffold up an Express-based draft of the same content (which can largely be recycled easily anyway) and same concepts, to weigh and measure the approaches; I have a feeling that, one way or another, I probably won't need to get too far into the work, before getting a clear idea of which approach is going to hold in the long run, and why, from a perspective that I don't have currently.

Another option is to try with Vue/Nuxt. This is also quite appealing...

Let's see... I'm struggling to find the fire to get back into this project in its' current form. This will probably pass eventually, but meanwhile I have the opportunity to ignite some new inspiration and drive a whole new learning cycle with some other popular stack.