5 ms·
First, it makes wonderful sense for certain projects even without productivity/learning-curve gains. E.g., anything that's not particularly supported by the HTM
by throwawayjava 8y ago
First, it makes wonderful sense for certain projects even without productivity/learning-curve gains. E.g., anything that's not particularly supported by the HTML+CSS paradigm. Think "stuff that would today be a native application".
But let's jump into the "finding labor/business case" problem.
> It is already very hard to find a decent backend developer
And why is that?
> In my opinion the super hype over Javascript is because of it's simplicity(compared to GO for example).
Agreed! You totally hit the nail on the head! The reason frontend is easier to hire for, is because HTML+CSS+Javascript provides a relatively small minimal viable set of abstractions that allow you to be productive.
Of course, I don't want to diminish the fact that there's an infinite rabbit hole of complexity here (see e.g., wavethrough); serious engineers can definitely thrive in front-end work. But for "productive enough frontend dev", the set of things you have to learn is relatively small compared to backend.
> if the frontend is written in backend language like GO, it kind of makes it worse.
On the contrary, I think WASM could enable a fix to your problem!
The question is, could we build a WASM-backed unified development environment that allows people to learn full-stack development in the same amount of time that it currently takes to learn just the front-end stuff? And can you produce productive full-stack engineers for a particular full-stack framework while skipping a lot of the fundamental CS stuff?
If the answer to both is "yes", then coding bootcamps could step into that environment that churn out lots of relatively cheap/easy-to-find full stack development talent!
So. Who can build a single framework providing a small set of abstractions (learnable in 6-12 weeks from scratch) that allows for both front-end and back-end development? If someone can do that, and build an educational commodity around it (i.e., a scalable model for bootcamps), then I could see wasm enabling a whole new approach toward web development in which "not top-tier but definitely good enough" fullstack developers are at least as easy to find and for the same price as today's "not top-tier but definitely good enough" frontend developers.
- ivrrimum 8y agoOkey, here's my observation: Frontend work is a hugely lower barrier than backend. You can take a dummy new guy and ask him to do same work in JS and he will do it with using google, etc. Backend languages is whole other deal. There margin of "oh damn, i fucked up" is a lot wider. The inert complexity of JS compared to any other backend programming language is huge. JS is pretty easy to learn(not master) and the experience is very interactive. With backend languages it is not the same. You can't really make a quality code with GO without knowing a bunch of complex conventions.
- marcosdumay 8y ago> Who can build a single framework providing a small set of abstractions (learnable in 6-12 weeks from scratch) that allows for both front-end and back-end development? Remeber the desktop RAD platforms, like Delphi and Visual Basic? There is no reason we couldn't have very similar things on the web.
- meredydd 8y agoObligatory plug: we've built one! https://anvil.works https://anvil.works - it's explicitly modeled on VB6/Delphi. All Python (front and back end, with a simple RPC interface from client to server), and learnable in a lot less than 6 weeks. Check it out!