Glance: A self-hosted dashboard that puts all your feeds in one place | Hacker News Reader

Glance: A self-hosted dashboard that puts all your feeds in one place

github.com

211 pointsulrischa35 comments

There was a golden time years ago.. when many sites had rss easily discoverable, and I used netvibes, pageflakes and my.yahoo as a startpage - was really great to have many topics and sources auto sorted by time/date of publish..

but the net started going more walled garden, places that had easy to find rss started to hide them and people started putting more important data being thicker walled gardens..

Imagine how much better it could be.. I suppose the only way this would be possible these days is to give crew AI your logins and let them scrape fbk marketplace and whatever news and bring it to your feeds place - remaking rss and filtering.. from those places that don't want you using them via rss and control of filters..

Maybe it's self-filtering then. Stick to the sites that still have RSS (for your feeds, that is).

My experience is that the websites without rss support are usually the worst and the ones you don't want to visit. It has become easy to ignore those so "walled gardens".

I have a soft spot for this sort of thing, and have built several. The challenge always comes with all the personal stuff I want on it- account balances, personal calendars, todo lists, etc. I’ll get it set up just so, then some api changes etc.

I think it would makes much more sense, to have a dedicated project which is focusing on just collecting the data and have a community to maintain updates. Then all those frontends could connect to whatever API this would offer.

There are many fancy projects like this for high profile environments in companies or specialized areas (like home assistant), but they all are very cumbersome to start and maintain. Something simple which could run with a oneliner and some config-file, in good old unix-mindset is really missing in this space.

> think it would makes much more sense, to have a dedicated project which is focusing on just collecting the data and have a community to maintain updates. Then all those frontends could connect to whatever API this would offer.

It already exists: https://woob.tech/

Yes, something like that, but not like this. This is too inflexible and limited. The datascrappers here are written in code, which makes it impossible to exchange them with other projects. And it seems there is not even a server which makes requesting data from an external process possible, to feed it to your own code. There are many tools like this, with a collection of specialized datascrappers for their own demand, and they are all on their own, isolated from the rest and everyone is constantly recreating the same.

It is code, so it is the most flexible that can exist. You can call the code with an application that includes it, you can make your own server with your own API although that isn't really useful if the server is going to be local anyway (most of these thing deal with private data, you don't want a central entity managing private data). You can call a process from another process, I don't see what's stopping you.

Perhaps if you have an example of what you want to build I can understand how woob doesn't fit what you want to do

Why do these things always have a stock ticker front and center? Who wants to look at stock tickers all the time?

Neurotic people with gambling addictions?

Maybe water cooler chit-chat.

People that make a percentage of their income from company stock?

I don't really want to be a stock owner, but the difference between paying my bills and doing more comes down to how people feel about the value of my companies stock.

Some people obviously care, but beyond that, I think it's an easy low hanging fruit when building them. There's lots of API's for it, the data is extremely regular, you get to show a graph and colorize some +/- and it has enough "bits" (symbol, name, price, delta) to look showy and interesting without being too complicated to actually design - the symbols even basically fixed width at 3-5ch.

BigTech employees often have 50%+ of their total comp paid in stock. Startup employees can be drastically more.

You get real twitchy about stock market movements around vesting time...

bank employees who often get these kind of customized dashboards setup.

Might be because it's a reliable and stable source of data, so it's not hard to implement. And it's something which is updating often, so it makes sense to have it as some kind of noise on such a tool, even if you do not care much about.

I know so many developers (including me) that build this very thing. There's clearly demand for it, but no incentive for any large services to integrate their stuff into it when they can demand that you go directly to them :(

Basically RSS all over again. Closed Web keeps data and users on their platform. Or an insanely expensive api, cough reddit, will do the trick.

Reminds me vaguely of Windows Desktop Gadgets, anyone remember those? https://en.wikipedia.org/wiki/Windows_Desktop_Gadgets

They are back again as Windows Widgets: https://www.microsoft.com/en-us/windows/tips/widgets

There's even a Developer-specific host for Windows Widgets in Windows Dev Home with some relatives of the GitHub widgets in Glance here, plus system performance things like CPU and RAM utilization. https://learn.microsoft.com/en-us/windows/dev-home/#dev-home...

curious about the UX really - feeds are a lot of things, hackernews, linkedin, twitter, instagram, techcrunch, bbc, my bank account, stock prices, crypto and so on

how do you make it extensible?

I was thinking of those "Magic Mirror" frameworks — and apparently MagicMirror² [1] is the name of a popular one. They're modular and do more or less the same thing. (I confess, I consulted an AI for the name of this one.)

[1] https://magicmirror.builders

Those interested in this might also like MagicMirror https://magicmirror.builders/ . It has many plugins, an active community, and renders in an Electron UI or in your browser.

I thought one of the points of building things with Go is that you get "single binary" deploys. Why is the preferred deployment tool for this a docker container?

Many people have entire setups around deploying just docker applications, it's as if it is a build target, so to say.

This looks great. The preconfigured page they provide as an example is almost exactly like I would want it to be configured. I must be the exact target market for this!

at first I thought this was Glances[0]; a python-based htop alternative that has some incredibly useful outputs (like statsd, influxdb etc;)

[0]: https://nicolargo.github.io/glances/

Only point: Why can't this not be made with php? This would be great to run on a cheap webhosting

People love it, but I could never figure out how to use it (either in Windows or Linux). The actual TUI was also fairly unstable (as in long hangs, etc).

Coworker had the same experience.

I was able to get it to run in minutes. (copy over the sample glance.yml file in folder and do a 'docker run'.

Haven't tried other methods.

Are you thinking of a different project with the same name? Like glances? Or Openstack glance? This is a new webapp, which seems to have no TUI. And it's indeed not much work to set it up with the example config.

It kinda looks like the 'wtf' app... maybe that one?

https://github.com/wtfutil/wtf

My bad. Yes, I was thinking of glances.

I don't think this project has a TUI at all, it's just a self hosted page of your social feeds, right?

I think previous commenter hasn't visited the link and mistake this for the glances[1] project (with a trailing s) which is a glorified top command written in python which provides both a tui and a web interface.

https://github.com/nicolargo/glances

Reply on news.ycombinator.com