Posts from crabmusket.net this week:

Nobody knows about Zulip

14 Nov 2024

Zulip is a team chat app, broadly similar to Slack or Microsoft Teams.

When the Future of Coding community started considering Slack alternatives earlier this year, I was reminded that Zulip exists. They didn’t end up moving to it, but I got interested.

The Changelog podcast migrated their community after an interview with one of the Zulip team , so I’ve been using it near daily for the last few months.

And you know what? It’s really excellent.

I am in several Slacks, for my day job and a couple of community groups. The difference when I open Zulip is noticeable. Zulip has a snappier and denser UI. But it’s the way messages are grouped into topics that makes it really easy to ignore a bunch of messages that I don’t need to read. That makes it much faster to get a broad sense of “what’s happening” and then focus on chats I actually care about.

On a personal level, I wish groups with ideological commitments would consider supporting Zulip, a small independent company funded by its customers whose owners are intimately involved in building the product. Rather than Slack (sold to Salesforce) or Discord (raised hundreds of millions of dollars).

I really hope to see more communities and companies considering Zulip as a viable alternative to the corporate players.

Paradicms, a minimum viable CMS

13 Nov 2024

I found out about Paradicms when searching for advice about working with RDF data in JavaScript.

(The name appears to be a combination of “paradigms” and “CMS”.)

The pitch is that they’re building a CMS for digital humanities projects like archives and museums in a minimal computing way.

• Data is stored in non-coder-friendly systems like spreadsheets.
• The final result is a static site, with no server to host or babysit.
• The same data can be built into different site layouts depending on the project’s needs, and Paradicms provides several templates.

RDF comes in between the data sources (e.g. a spreadsheet) and the static site generation. Spreadsheets in a conventional format are pulled into a linked data graph, which is then queried to create the data on each page. Critically, this happens at build time, not at page load time.

I like:

• Using RDF as an “integration” or “enrichment” layer between different data sources. It feels like all data models eventually become RDF anyway, when they are asked to accommodate enough use cases.
• The focus on minimal computing, avoiding complexity for non-technical users.
• We will truly never escape spreadsheets, and this approach leans into this truth and meets users where they are, rather than attempting to reinvent data management.

I worry about:

• The UX of editing linked data in spreadsheets is… not good. I’m not yet convinced “real people” can work with this. A significant issue is having to pick IDs from dropdowns.
• Is static site generation from RDF “simple enough” or is this where all the complexity moves to? The project provides some templated site types based on their digital collections data model , but what if I wanted to build something different?
• It uses Python instead of JavaScript, the universal scripting language of the web. (Yes, this is a silly nitpick and slightly tongue-in-cheek.)

I wish RDF were more popular as a concept, with apps providing schemas/grammars for their data export.

I’d like to try out the Paradicms approach on some small projects for local groups I’m part of that aren’t technical. I hope this could be a way to responsibly deliver a small data-backed site or app to a group that aren’t coders.

That's all folks! Thanks for reading ♥