What is currently the most accessible, lightweight (clientside) and simple to implement, HTML+CSS framework for web-apps?
Primarily for a web-app for mobile use.
To be integrated in a fully serverside rendered web-app.
@bpepple Bulma looks perfect.
I'll look into it, mostly into the accessability. So far: looks exactly what I need. Thanks!
@flockingbird I think most most modern frameworks do a pretty good job with regards to accessibility, but truthfully I haven't looked very hard at the issue.
@io I did not yave this on my list yet. It looks really clean and simple. I like that.
It does use some classes, though: e.g. to turn divs into grids; I presume because there's no "semantic" tags for those. But also a.button to turn a link into a button.
I'll certainly look into it, esp. because it advertises with "mobile first".
beyond that, would help to know what the most complicated UI feature you intend to have would be.
@flockingbird i actually have a whole list of css frameworks, i can check to see which ones have the component you most need and give you the filtered list
@zens The goal is to build our entire web-app, but that is far away; and I don't mind switching to something else along the route (react, native, wasm, some other CSS framework).
For now, the direct aim is to build a Proof of Concept and maybe even the MVP. Screens for the far-away goal are on figma: https://www.figma.com/file/CgDIaLgjwVLPzw1ggmrZzy/Flockingbird (work in progress!)
@flockingbird I wouldn’t want to overprovision a framework with more complexity overhead than you need. underprovisioning with less than you need can cause issues too. some of what i see here would probably eventually need a reactive framework. I personally am not a fan of react.js.
for a prototype though, yeah, a css framework will probably be fine, but maybe more UI oriented than tailwind or bootstrap. you’ve got mobile app looking things there.
@zens Our most important requirement is not visible: SEO. If privacy preferences allow, we want peoples' profiles to be indexed and ranked high.
That means serverside rendered HTML (with microformatting) is almost a hard requirement. Added benefit is that this makes the app fast, compatible and accessible.
This requirement is so important, that if forms and interactions are too complex to achieve with plain-ol'-HTML forms and a sprinke of JS, i'd prefer redesigning it, so it fits that need.
> That means serverside rendered HTML
In 2010 perhaps. The world has moved on since. Make sure the product offers a pleasant experience and that you have the best sales guy you can buy, borrow or steal. The rest will fall into place.
Also, take too long and someone *will* get there first, leaving you with no market to sell to.
@zens I'm familiar with react, but don't particularly like it. I love redux, though. Also familiar with angular 1; onfortunately.
Haven't used svelte or vue. But all of those are, for this, I guess, overprovisioned. Maybe in some future when we can afford to hire dedicated frontend-devs it will fit.
But even then, serverside rendered HTML will remain of paramount importance; as outlined in previous toot about SEO.
@flockingbird after doing whizzbang SPAs for years i have been falling back to a more traditional approach, rendering the whole thing out server side like a slow sldeshow and only using js to, i. essence, draw the inbetweens.
@zens I'm a big fan on DHH and his approaches.
Been doing Rails development since before 1.0 :).
Yet for flockingbird I'm using Sinatra: ActiveRecord is just too unfit for our event-sourced (reactive?) use of ActivityPub.
And indeed, JS would be "sprinkled on top of the HTML+CSS" to make it work just a tad smoother or dynamic. Vanilla JS or even jQuery is probably more than enough.
@zens I've been using semantic-UI for some courses and love it so far. Thanks for confirming!
Don't know the other two, but'll give them a spin next week. Thanks for your time, also!
One word of advice: just fucking do it.
Discussing on the internet the different alleged pros and cons of every possible choice under the sun is not going to get you one inch closer to having a product delivered.
Pick one option, any option (ok, not #Angular 😁) and start coding. You can always fix your mistakes later, but keep in mind that perfect is the enemy of good enough.
@0 An Open Source project should leverage the community. Some choices can be made better with imput from others.
But that advice is neither taken as dicate, nor asked on all choices.
E.g. when someone explains to me how to do SEO in 2020, I'll nod, take a note and go with my own choices made on that matter already ;).
And indeed! "getting stuff out" (shipping) is by far the most important. All questions asked are to ensure were shipping the right product and shipping as fast as possible.
@0 In any case: Thanks a lot for the input. It confirms what I needed and gives some food for thought.
And just in case: we're building already. Nothing shippable or even usable yet: mostly PoCs/research and demos.
If you're curious: current PoC is to find out how AP will integrate in an eventsourced architecture (like Pleroma): https://github.com/Flockingbird/roost/
@CaffeinatedTech Thanks. You refer to the PHP framework, I presume?
Because we are already building the backend and AP engine in Ruby (not rails) so Symphony is not on the table for that, currently.
If I misunderstood and there's a frontend HTML/CSS framework called Symfony, I'm really curious what the URL is.
@flockingbird Yeah I was talking about the PHP framework. Bootstrap is quite a mature frontend framework. I’m not a fan of following the bleeding-edge trendy frameworks. I prefer the tried-and-tested ones with years of documentation and tutorials :)
Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.