Hacker News new | ask | show | jobs
by zokier 790 days ago
Honestly I think LibreOffice Base (and its brethren) are underappreciated and underdeveloped pieces of software these days. In 80s/90s stuff like dBase and Access seemingly were relatively popular, but that whole category of software seems to have mostly died out which is imho a shame. I suppose Access being somewhat notoriously horrible has something to do with that, suffocating the whole field. Considering how much business users are stretching Excel to fill the gap, I do think there would be demand for better options. I guess Airtable etc are the modern alternatives.

I find it just a shame that databases as a concept is something that has been relegated to be hidden deep in the backends and accessed only by specialized DB admins or through narrow and leaky APIs; SQL is considered arcane wizardry instead of being suitable for technical business users like afaik it was originally envisioned as.

3 comments

I haven't seen anything that beats the UI and UX of FileMaker Pro from the 90s.
Agreed if FileMaker would do what you needed. 1990s FileMaker was not really relational, right? To be painfully fair to Microsoft, Access was more powerful and did a nice job of balancing capability and UX.

For many contemporary use cases where one might have used FileMaker in the nineties, there’s SharePoint lists and PowerBI.

> FileMaker was not really relational, right?

Not until v3 (1995) or v7 (2004), depending on how low your bar is.

Truthfully, given the backwards compatibility of Windows and Wine, you could still get away in this day and age with sticking with it as a solution.
Has FileMaker Pro changed / improved / worsened significantly since then? The latest release is FileMaker Pro 2023.
It has created a how complex layer under the simplicity to better compete with databases. It’s a great environment but they seem to cater to experts more than people need to Crete a database for their own use.

It is great deployment platform for offices based on FileMaker/database workflows.

to give some perspective, there is an app i've written as a complicated google sheets script, which would have been way simpler to maintain as a desktop app, but what the end-users want is the ability to have multiple people use it from multiple devices at once. that's become table stakes for a lot of software now, and it's really what has killed desktop apps for stuff like this.
Very interesting point that multiuser killed desktop apps. Couldn’t desktop apps implement multiuser?
they could, but you would have to do pretty much everything from scratch - set up a server, figure out client-server communication, handle state changes, flaky networks, etc. with a webapp you get a lot of that infrastructure "for free" insofar as the entire stack is designed around the multiuser experience, and you are going with the grain of the platform. not to mention all the frameworks and libraries that help you with most of the common tasks you will face. you can follow a rails tutorial and get a small database with a multi-user read/write user interface working within a day or two.
Desktop apps can use HTTP(/grpc/rest/graphql) for communication just fine, that's not really problem.
true, and that saves you having to write the frontend in html/css/js, but you still need to write an api-only web app to act as your backend. the "going with the grain" path for a desktop app is to be running on a local machine, with all data stored locally, and interacting with the user over a keyboard and mouse, not the network.

you can absolutely write a desktop app as a web app client, but it's a lot less well-lit a path than either writing a standalone desktop app, or writing a full-stack web app.

Notion is putting the database in the Center in a way.