Show HN: Fastify DX, a full stack framework based on Fastify and Vite

https://github.com/fastify/fastify-dx

by jgalvez • 4 years ago

47 22 4 years ago

github.com

Semaphor• 4 years ago

This required quite some clicking to find out WTF fastify is [0]. Luckily I knew about Vite [1] already.

[0] https://www.fastify.io/

[1] https://vitejs.dev/

chatmasta2• 4 years ago

Very cool! I’m a big fan of Fastify. In fact I’m working on a SQL-over-HTTP bridge right now, published first as a fastify plug-in. And I’m using Vite, so this repository will be nice to study. :)

I originally chose Fastify over express ~2 years ago, due to benchmarks and newness, and it’s been running stable in production since then. It’s great software. If you’re coming from express there is a bit of a learning curve. Fastify felt a bit more rigid at first - for example with defining schemas for request body or not logging with console.log - but it was totally worth it.

From that perspective, I think this dx framework you’re making here is a great idea! Might I (selfishly) recommend a variant for library authors?

jgalvez3• 4 years ago

Thank you! There's a low-level library it builds upon:

https://github.com/fastify/fastify-vite

hsn9154• 4 years ago

As usual, I click the link, I try to read the page content but I have no idea what it's saying.

Buzzword buzzword, buzzword!

Just tell me what is it, and if you don't want to, that's your right, but then why are you sharing the link?

- What general category of tools is it?

- What makes it different from other/similar tools?

Don't tell me "it's fast and developr friendly". Tell me what makes it that? What makes it better than the other tools? What do these other tools lack that you are providing?

brailsafe• 4 years ago

Ya, sometimes I think people with this opinion didn't read the home page, but this is beyond me. Seems like a junior programmer's side project or something based on one of the deeply nested descriptions.

I guess at some point they were a contributor on another esoteric stack layer and decided it was too much, so they announced rebuilding it on an even more alpha and esoteric bit

string2• 4 years ago

I can only assume you're not a web developer? Because just the title of the post makes it clear to pretty much any web dev what this is. The package introduction sections accessible in one click are also very clear and concise.

pdimitar• 4 years ago

I'm a web developer for most of my 20 years of career and that page tells me absolutely nothing except that the software is apparently a full-stack framework, which operationally and concretely can mean hundreds of things.

I don't think you understood the criticism of the parent poster. I too got exactly nothing concrete from that cryptic generic README.

yuppiepuppie• 4 years ago

You shouldnt have to be a <whatever> to understand a tool and what it does. I think Richard Feynman had a technique around this.

hsn9154• 4 years ago

No, I'm a web developer (unfortunately?).

Maybe you don't realize there's a hundred ways to develop websites and just because you know a certain stack does not mean everyone else knows it (or cares about it).

jgalvez3• 4 years ago

You can think of it as a potential replacement for Nuxt.js, Next.js, SvelteKit, Remix etc, if you're running a live Node server and prefer structuring your app as a Fastify application first.

hsn9154• 4 years ago

I mean, let's suppose that I do know what Nuxt.js is (I don't). The page tells me nothing about what would make this an improvement over it, or over anything else.

swyx• 4 years ago

agreed. i really dont like this trend of branding products with DX. let people talk about your DX if they like you, putting it in the product name is just so presumptive.

nobleach4• 4 years ago

The primitives on which this is built are cool. Fastify and Vite are two projects I love. The moment I see TailwindCSS added as an opinionated default, I nope right out. Tailwind is one hype-train I can't wait to see run out of steam. I've been around enough years to see a lot of things that we thought were GREAT ideas, turn into "what were we thinking?" I believe we'll be doing the same with Tailwind in 3 years. "Why oh why did we think annotating all of our HTML elements with a bunch of class names - some which represent rule-value pairs, was a _good_ idea???" This viewpoint makes me very unpopular obviously. But it's a hill I've chosen to die on. It's bad DX and it's worse for maintenance. I welcome a lot of downvotes... but let's revisit this in 3 years and have a candid conversation about it. I'll gladly admit it if I was wrong. But I wasn't wrong about AngularJS, Bootstrap and a litany of jQuery plugins. And JS developers fiercely defended each of those at one point as "the one true way".

jgalvez3• 4 years ago

Tailwind is not part of Fastify DX.

It's just part of the starter template, through unocss.

unocss goes beyond what Tailwind does. it's general CSS processing engine and provides quick access to some rather essential goodies, like an icon set. I think it's a worthy addition to the starter template, but easy to remove. Literally remove one import and a vite.config.js plugin entry.

mgrandl• 4 years ago

I have been a frontend dev for 10+ years and tailwind is absolutely amazing DX for me personally, but to each their own.

nobleach4• 4 years ago

Hey yeah, that's entirely your call. If you and your team agree on it, I'm very happy to hear it increases your quality of life. Where I start to drift is when I see it included on any project because "everyone's using it"

I do not ever want to use it. I tried to see what all the fuss was about. I read the book and. No thank you. I promise I'll add it to projects on my own if I come across new info. Remix takes the right approach. They recommend it in their docs... but stop short of making it a product dependency. (If it's included in one of their stacks, I'm not sure)

CharlesW• 4 years ago

But I wasn't wrong about AngularJS, Bootstrap and a litany of jQuery plugins.

Although it's an imperfect measure of developer popularity, Angular has 82K stars on GitHub, Bootstrap has 158K, and jQuery itself has 56K. Both projects you love have fewer stars. Assuming your opinion is that those projects are "dead", you may have "shiny thing syndrome".

chatmasta2• 4 years ago

You might argue that popularity is an indication of how many people are stuck with it due to (for example) Bootstrap’s unmaintainable nature.

I agree with thread OP that tailwind is a mistake, but I’ve also never tried it myself, and I see a lot of people like it. I’ve avoided it because I have a rule of thumb to avoid any DSL embedded within another DSL. It’s a nightmare for static analysis, as demonstrated by the fact it needs its own compiler. That’s enough reason for me to stay away from it; I’ve got enough layers of compilation to worry about as it is.

nobleach4• 4 years ago

Please don't assume my opinion.

AngularJS (not Angular) was a bad abstraction. Many years later, we can all see it. It's so bad that the core members abandoned it to create Angular - which, many still dislike in the front-end community, but it's actually a very competent project. My gripe is about including bike-shedded "everyone's using it anyway" dependencies in any project. I wouldn't want this author including Styled Components or SASS or Less either!

So no, I was not wrong at all in 2013 when I called out AngularJS for being a bad abstraction. The amount of hate and downvotes I got back then is just like what I get when I say, "tailwind is a bad idea and I think we'll all see that in a couple of years." Nothing to do with "shiny stuff". I think it's more like, "I'm old enough to have seen this show play out a few times"

Show HN: Fastify DX, a full stack framework based on Fastify and Vite

Related Stories