Is the ‘free’ part of open source not a priority? There is an AGPL license, but it looks like future contributors will be required to use proprietary services for this helping this project with Discord & Microsoft GitHub as the only communication options. If FOSS is good enough for your project, it should be good enough for the tools used to build your community…
How does a news site operate without plaintext? The body element is entirely empty. Is there no need for SEO or accessibility? At some point last decade the text-only version of the BBC was discontinued, but instead you can only compare to the likes of the NPR Text-Only Version https://text.npr.org.
…Just don’t commit your .envrc to a repository & add it to the .(git|hg)*ignore. Provide an example if you want, but don’t expect everyone to want to use your exact settings. This is for your personal environment.
I’m on Posteo & it’s fine. I can recommend it, but I can totally understand wanting a host that lets you use your own domain name for portability (they have arguments against it).
Great to it’s hosted on a free software forge too not locking in contributions!
Not sure I always agree that feeds should have the full post tho. This not only (obviously) bloats the size of the feed, but there are valid reasons to want to drive users to your site--especially if you have demos or you write about code & have your code blocks syntax highlighted (statically, never do this with a JavaScript) as it provides a better reading experience. You can put styling technically in Atom/RSS but even then, a lot of readers won’t be applying the styling. That said, I definitely appreciate the full post if your site is full of trackers, ads, marketing garbage or other bloat since I can skip the site. Is this some site engineers giving us the nod on a better UX? I read a gridiron football news site & boy does that feed become take a site from unusable to pleasant (good photography).
Good thing there are alternatives to Matrix that use a lot less resources. It’s not just UX… Matrix by design is wasteful by requiring replication of all messages & attachments of all users in all chats across every server. Most can’t afford to host a multiuser server with the storage bills for this.
Users that need special clients (accessibility, hardware, etc.). Users blocked by US sanctions. Users that have been moderated off the platform for something not in your community (account bans even happen accidentally). Users with privacy/anonymity concerns about the data collection (& now ads)—especially the chat rooms that require a SIM card. Users that take their FOSS or otherwise ethical software views or anti-corporate views to heart & want something built on those principle—from wanting to use free software to make free software, to wanting to outright avoid what some now call enshitification where a free (gratis) account clashes with the idea of freedoms, etc.
If open source, free software is a good enough ethos for your code base, it should be good enough for your community communications. Supporting only a proprietary locks out a swath of users & Discord in particular is an information blackhole—and now with ads!
The cost of compression isn’t nothing. A lot of Linux distros have moved off of it for packages since they heavy Zstd compression was close enough on size while saving a lot of time & CPU bandwidth.
If it’s static & can be edits to content that could easily render out as static (such as some reStructuredText files) I always reach for Soupault so I can create something super lightweight, bespoke for the job, but extendable with plugins & tools that don’t lock me into a specific system… it’s mostly just the composition of applications delivered by my package manager, piped thru stdin/stdout, & glued together with this tool. The last thing you need are tools that are complicated, inflexible, aren’t built to last a decade of low maintenance, & whose outputs or dependencies are hard to make something reproducible when just building a website.
Now that Discord will start showing ads, is there any chance Bun will support a communication platform that is open & ad-free like IRC, XMPP, or Matrix?
These projects do other non-gaming livestreams for certain events, meetings, etc. Twitch was just a game streaming platform for many years with more incidental gaming / just chatting now than not. Unless there’s a lot of pushback, this will likely just be the tip of the iceberg before allowing more ads in more places in the future.
RIP to all those FOSS projects that thought it was good to build their communities solely on proprietary chat. I don’t expect the average folk to understand the ethical, privacy, or ownership issues, but in the software community, I do. Now your members get ads, & those ads are almost certainly expecting to siphon user data (as if Discord isn’t doing enough of that as is & not buying “seem minimally intrusive”).
The home base server should be free or otherwise ethical software the community can own, ideally decentralized (in practice, not theoretically decentralized)… & then use bridge/gateways if you want to support these proprietary, privacy-sapping platforms.
You can a) set up a gateway/bridge or b) contribute to those platforms to make them better so users like them more as the open+secure options are more worth the long-term investment into instead of making users choose between privacy/security or having the ability to communicate.
Why are users, especially technically-inclined ones, hellbent on using these proprietary services? Throw them all out if they are going to be this way since there are great alternatives. The last thing communities should do is build UIs around these types of services instead of building ones for systems/services that aren’t so hostile towards users.
This was my knee-jerk too. It’s great to see some of the good bits of ML flow into new languages, but this isn’t some Rust-specific feature. As an annoying side-effect tho, it’s been harder to pitch FP in these existing spaces since now “we have FP at home” in the existing language (even if the ergonomics are worse, it looks uglier, the language idioms & existing programs don’t match the style, & ’the bad parts’ + footguns still exist).
What makes folks thing landing pages like this should require JavaScript? Nothing here would need the dynamic capabilities of a scripting language, yet accessibility & SEO are hurt in the process.
Darcs has rebase & Pijul doesn’t which can really help ergonomics for fixups or if you do a lot of WIPs that need amending. Darcs also has a send command for easily mailing patches to others. Darcs also lets you override the diff output which is compatible with decades-old tooling rather than something bespoke. Pijul despite better performance, an awesome user identity system, & channels is still not a Darcs replacement IMO. That isn’t to say folks shouldn’t use or follow the Pijul project (they should) more than it is to say one should still feel okay using Darcs today too as it’s still getting updates & has features Pijul doesn’t; what’s wrong is folks calling Pijul a ‘successor’ as if the Darcs project had died. You could still use Darcs in 2024.