Nyxt browser annotations beat pen and paper
nyxt.atlas.engineer
nyxt.atlas.engineer
There used to be a number of extensions of this type on old Firefox. I cannot remember exactly this one, but here are a couple to prove what was possible [1]:
- the original Firebug which became developer tools and was cloned to every other browser was just another Firefox extension(!)
- an extension I used a lot was Scrapbook, which let me crawl and store websites (I used it for symfony docs back in 2007-2009 before mobile broadband was available or at least not within reach for me. It would download pages recursively according to simple, useful rules like maxrecursion, only subdirectories and only a certain domain and then rewrite every url on the page to be relative instead of absolute. You could also edit the pages locally.)
- Alerts (or something similar) would check web pages according to simple rules I set up and notify me if it saw significant changes.
[1]: somebody will probably tell that it existed too and was even better, I just didn't know of it.
The Orion browser that I like very much takes a similar approach: they let you install both Firefox and Chrome extensions.
Hopefully they and/or Librewolf will start providing APIs that will let us create even better "native" extensions for those browsers.
Seeing the functions of Nyxt, they look exciting and perhaps I will try this for short-term knowledge (e.g. brainstorming on a topic, where I won't need to access this knowledge again after e.g. one year). For everything else I will stick to nested folders and markdown files, where compatibility is almost guaranteed.
Curious what others think about this issue.
(Mostly because it is suitably ubiquitous -- it's an operating system feature for me.)
As I consider moving to a more low-budget dev environment again I do question whether I need to jump the remaining step to text files in some cloud storage. But I'd want immediate background sync (not having to wait for the note to arrive on a device or explicitly fetch it).
I've been using the xBrowserSync extension to handle bookmarking across browsers; it also uses tags / descriptions + search for organization and recall. Its made a huge difference. But Nyxt now takes things even further; locally saving a copy of the page you're annotating. Great job!
Depends on the person I think. My mind is intuitively comfortable with graph/tree structures. Despite having used Gmail since beta, I was never successful in leveraging the label/tag+search structure into something manageable. It was always chaos despite recurring attempts to tame it.
Having switched to self-hosted IMAP with the traditional directory approach, I feel in control of my e-mail for the first time in over a decade.
Similar experience with bookmarks.
Great that you found something that works for you but I think it's a mistake to believe it's possible or desirable to have one structure for everyone.
Had similar thoughts in this thread, where so many were arguing if "users love settings" or "users hate settings", and which settings are loved or hated. Stop assuming you can reduce every human into a single profile.
https://news.ycombinator.com/item?id=30181291
(Apologies if I come off as projecting at you here, just got triggered a bit after yesterday)
I've been wanting to write an implementation of an annotation server, and it'd be cool to immediately interoperate with nyxt browser. Then I'd work on the same server supporting most other browsers (e.g. Firefox) via WebExtension API building on my old https://github.com/gobengo/web-annotation-extension
Conote is currently in early beta and there are a lot of rough edges eg. it currently only supports Chrome and Firefox desktop to create highlights (soon to be fixed). I have plans to add much better knowledge retrieval searches etc but for now am focusing on making the core ability to add context to your thoughts.
I welcome any feedback you might have :)
I assume that any given useful website may be gone in a month. This may be a holdover thought pattern from the 90s, but for me if it's worth knowing, it's worth writing down in my own system - which is app-independent, and could be printed from the source files without loss of information, should that become necessary.
Oh, and if the page author is here - please, reduce the resolution of that painting image. It's unnecessarily massive.
Will future Nyxt enter the zettelkasten space?
For example, some power users already integrate Nyxt captures with org/org-roam, so it's "already" Zettelkasten friendly, in a sense. Whether the devs plan to branch towards that, I have no idea, but it should be already possible to integrate Nyxt with org-roam (maybe even some spiffy interfaces like the one demoed in OP).
1-Store the information very quickly, without typing any additional text, so no tags and no annotations
2-Retrieve the information based on some word I might remember from the previously saved text or based on an approximate date
#1 happens a lot so I don’t want to spend time annotating or tagging information (ie click save and forget…)
#2 is much less frequent and when I really need the info, I can spend a bit of time to craft my criteria for a search
Making #1 laborious or painful is a nonstarter for me. Anyone else on the same boat?
See, your memory is using some words as index to access stored information. And it's indexing it asap (well, some re-indexing probably happens when we sleep). Thus, I don't think an efficient digital memory extender can work without active human participation in indexing. Otherwise, it will produce a lot of garbage.
If you have access to the office 365 ecosystem, I would recommend checking it out (which that said, I strongly discourage the use of the native window10 version of oneNote, its a dumpster fire on wheels)
There's a couple of Browser extensions that have sorta tried to accomplish this before. Epiverse[1] seems to be the most polished one so far. It originally intended to allow any user to comment on any webpage. But it found that without existing content, few found it useful. So eventually the creator just parsed to see if the webpage was posted on Reddit or Hackernews. The original purpose of it ended up being too expensive to host so it ended up just becoming a HN/Reddit parser. Which, tbh, is basically what I want to build at this point. I'd love to contribute to the project, but I don't have much time and it's closed source
The other similar extensions also just parse HN/Reddit like Newsit[2] (which is open-source) and Thredd[3] (which only parses Reddit). My only real addition to this is that I'd like to include the ability to parse more than just Reddit and HN. I wanna create a discussion aggregator. There's similar sites like Lobste.rs and Lemmy.ml that could also be parsed, but obviously that's not the full extent of where discussions happen around a webpage.
I don't think I have it figured out, and I don't know if anyone ever will, but I think there's a lot to gain if someone is able to someday harness that feeling that you get when you read something really good or find something really cool and wanna see how others responded