Roam has the idea of sort of – individual pieces of info at the bullet level so when I put on my December 31st notes page: [[Delapouite]] lives in [[New York City]] – that single piece of data shows up on both the [[Delapouite]] page and the [[New York City]] page. The breaking of data down to the sort of like atomic level and the linking has massive value.
I love roam research, but I also hope for a great open source clone to de-risk it. Also, works great from mobile browser, so it is a great phone app too.
So it kind of automatically greps for \[\[.*?]] and shows all references at the bottom? And making it easy to create a new page? Looks pretty powerful and conceptually easy to replicate.
@danr basically – nothing in it is super-unique. I had a very similar system in 2001/2002 for a corporate wiki that we loved.
Roam is just an exceptionally polished implementation that focused on time to new “point of data” and keeping “points of data” uniquely targettable (each bullet) rather than at the page level that most wikis are.
I really hope it is cloned because I fear them going out of business, but not enough to kick them $500 for the true believer package.
I stumbled across Roam the other day while researching the “Zettelkasten” idea. Other related software appears to include Obsidian (which stores notes locally as Markdown, instead of in the cloud), and Neuron (a command-line app that nicely a directory of Markdown files to a pretty website). Now that I’ve learned a bit more about it, I think that my kakoune-ghwiki plugin was beginning to scratch at the same idea.
As a commercial project, probably nothing will approach the polish of Roam except for other commercial projects. If you’re willing to lower your standards a bit, though, I think you’ll find a bunch of options along the spectrum to “super-modular, minimal lock-in”.
I’ve heard a lot of good things about Roam but never got around to trying it. I wish there was a modern version of Notational Velocity, that was a universal app. im aware of NVAlt but even that is pretty dated, and not universal
Yeah, I have poked around a few other options – the sad thing is the polish is so important for daily use – so many of the little things about Roam work so well it is hard to give up. For example, page renames properly go everywhere, so no fear of renaming, which was one of the biggest problems with wikis. Even reworking single datapoints which are referenced very cleanly update the references to them. A lot of this is due to the graph DB underpinning.
I am happily paying the $15 a month, but my loyalty is non-existent, I just need a tool that makes entry very quick – a few months into Roam and I am coming up on 1,000 pages because the entry process is extremely easy.
Honourable mention for tiddlywiki which works well with the Zettelkasten method. I have it running on a Raspberry Pi so I can access it from multiple devices and back it up to a NAS. It’s open source and files are saved in markdown (though slightly strange dialect of it). Probably needs more tweaking than the commercial alternatives though.
One of the more interesting (and elaborate) examples is h0p3’s.
It has backlinking and Zettelkasten as well as being free and open source. It dosen’t look like it supports syncing, though. I haven’t really had a chance to use it but it looks pretty interesting.
Ah so you tried tiddlywiki before settling on Roam.
Roam looks great. But I was too worried about vendor lock in. Plus tiddlywiki was a solution I could use vim with, rather than having to go through a browser / app.
I’m using TiddlyWiki as personal wiki right now (NodeJS version on private server). I have a krystal plugin installed. It gives you a similar look & feel to the Evergreen notes. Roam-like Tw distributions are Drift, Stroll and TiddlyRoam.
I don’t enjoy using TiddlyWiki anymore. It does not scale well. Mine is already ~6mb (2mb compressed) and I’m using only since 2020-06. That’s 2+mb non-cacheable load every time I’m visiting my wiki. Performance is bad, especially if you are using a graph or full text search extension (or a lot of dynamic queries). Exporting tiddlers to HTML is overcomplicated. My initial idea was to use a subset of notes to generate my website, but I never managed to produce anything usable.
Notes are in plain-text format (good) but it’s not Markdown or any other widely used markup (bad).
I used to use TW for work related notes as well. I replaced it with LogSeq (for now). It’s more interactive and has less clunky interface. It’s still young (read: buggy) project. Since it’s build similarly to TiddlyWiki (all data in memory all the time) I believe it won’t work in a long run as well. My goal is to replace both wikis with something else. Most likely I’ll write something tailored for my needs from scratch.
When it comes to Roam: Putting a personal wiki inside a proprietary SaaS app seems like a bad idea.
It is – but it is the least bad. I have been keeping digital personal notes for – two decades about. I have about 13000+ pages of data, been imported and exported through generations of systems. In the past, it additionally had 5500+ links and cached copies of those pages, moved to pinboard.in with the archiving option.
The experience I would like back is the old Evernote experience with page clipping and notes at small scale. Sadly, Evernote scaled horribly and was incredibly buggy, but being able to have as rich or simple interlinked documents as I want from web content or not via a web plugin was fantastic.
I have gone through dozens of systems, paid and free, and thus far, Roam Research is the one that makes me the least angry.
My dream system would be open-source, graph-based, have FTS with faceting, intrapage linking (to datum, like Roam), have chrome and firefox plugins for web snipping, and have email sending to it…
Wow! I have the same sort of system - and the same problems!
A durable, open-source platform for my massive notes collection would definitely be welcome, but if Roam does appeal to me then I reckon I can try it out while keeping my notes backed up in other places.
I’ve wanted to do the work on my own personal platform for a long, long time.
Is not folder structure with markdown and find + xargs grep not enough for this? I did little bit of research but I would like to use my unix ide (kakoune + nnn + gnu scree + fzf + find + grep) to manage notes and so far it work quite good, only problems I have are
I did not find sufficient toolchain to use on phone, so for now I use samsung notes as inbox and then rewrite them on laptop
working with image or audio files is cumbersome
I suspect that I will hit performance issues when I will have enough notes, but I think I can fix that easily by creating some index file instead of always findxargsgreping
I have unsubscribed from Roam Research. This isn’t because it failed to do anything it promised. My primary reason was – I am easily distracted.
Google Keep and Roam Research (or Org Mode, etc) are polar opposites. Keep takes away the temptation to over-index, over-organize and over-formatting by not letting you do much of it at all – the bare minimum of features.
It turns out without this limitation I am weak and will spend a lot of time organizing, structuring and doing other formatting things that never yield a return on that investment. I do love the idea of “today” being the default space, so now I just make dated notes in keep – 2021-10-25.
Still think RR is great and would be amazing for a team or a specific topic (notebook) – but as a sort of “write down my life” – it ended up letting me get distracted too often. PEBCAK.