I got the hypernote demo when Roam was still infant prototype on my local machine, but main benefit was showing me what NOT to do.
First day I tried it, did same test I had for my own tool - Euclid's elements.
It was still worse for this than nValt.
Conversation
The thing is, when youre at the frontier, seeing actual demos that help you clarify what you DON'T want is HUGE.
Graph.global was huge for my thoughts on "reified edges", which was 9 month dead end - but lead to backtracking and seeing need for page AND block refs.
2
9
To be more specific Graph.global was huge for seeing first interface where you could treat [conor friends-with mek] and treat that as object to build new relationships on
[Conor friends with mek] due to [Belief in Memex]
Quote Tweet
Ill be first to admit that both @mc2 and @jacobCole4000 were enormous inspirations which encouraged me in to create graph.global (open source) in 2015:
github.com/w2g/graph.glob
Show this thread
1
3
Notion deserves credit for a lot too, especially with respect to UI -- deserves much more credit than Notion though, because project that got him job at Notion (markdown parser that allowed arbitrary React components) solved HUGE bottleneck 4 us.
Quote Tweet
For years I've been working on a Doug Engelbart style "Dynamic Knowledge Repository" @NotionHQ beat me to it. Try it notion.so/invite/twitter
1
4
Just the idea of it. Whether or not code was open source was totally irrelevant.
Had 10 minute conversation with him about the side project, with a few days, got up in middle of night and built {{[[TODO]]}} and {{[[DONE}}, which opened door for {{diagram}} and {{table}}.
1
3
org-mode style TODOs in Roam is one of the most promising ideas I've struggled on:
1. todo.rip (which I attempted to connect to graph.global)
2. michaelkarpeles.com/essays/univers
3. Look at this mess: graph.global/?id=317
1
2
Fact that you could combine single markdown escape hatch for arbitrary display / interactivity - from
And STILL build off all the filter and backlinks power Roam already had...
That was the novel insight
Chet inspired {{todo}}
Rich hickey inspired {{[[TODO]]}}
1
6
So much of the hard work in design comes with figuring out details of the workflow and conceptual model - often only by months of working with users.
Open sourcing actual code for consumer tools is red herring, ignores other contributions to commons
Quote Tweet
Replying to @mekarpeles @RoamResearch and @AthensResearch
Open source was totally irrelevant - at least for me.
Never needed to read your code, definitely was never going to use.
What mattered was you pushed the conceptual frontier.
2
9
I agree there are many valuable forms of contribution:
One not needing a resource is not grounds for justifying it be closed. The world is built on resources I, in my privileged position, don't need and which others do.
All things equal; show work, share, & enable.
1
All things are never equal though mek.
2
4
I sympathize with Conor hereโit's already so hard to get a tool like this off the ground; adding more constraints (i.e. on production flywheel) seems like asking for trouble. I love that we've found some models that work (e.g. open core); I'd love to see a playbook collection.
I struggle with this in my own work. It's obvious that if it's to succeed in the long term, Orbit wants to be an open standard with OSS implementations. But it's also obvious that there are enormous path dependencies; it's not at all clear when and how it's best to move to open.
1
9
My current inclination is to rapidly publish sources under restrictive "source available" licenses, then slowly move modules to open licenses from the bottom of the stack upwards as I better understand my path to sustainability. But who knowsโฆ
2
9
Show replies