2022-05-04 - Rufus Notes
2022-05-04 - Rufus Notes
#done/process everything moved out to INBOX, PLAN or elsewhere (or done).
Tasks (for Rufus)
- Notes for Ola with initial plan of work 🚧 see below
- Outline of key streams and how they fit together
- Detailed next steps
- DataHub Enterprise refactored with main content as project and description of DataHub Enterprise in there ✅2022-05-05 refactored and split out DataHub Enterprise Website
- Create plan of work in Flowershow
- How does this relate to DataHub Pages (isn't it basically the same)? Yes, so can we converge them? DataHub Pages is a special case of Flowershow.
- Work out (? with Anu?) whether we go monorepo and how we split things up
Minor
- Access to next.datahub.io for Ola + Khalil
- Register flowershow.app @rufuspollock ✅2022-05-09 done
- Future DataHub monorepo
/portaljs # aka toolkit / present / showcase. /publish # what is currently datapub /flowershow /cli # command line tool
- On naming: think toolkit is too generic - more the product name covering a bunch of this
Plan for Ola
#todo/process into PLAN.
Potential (Work)Streams
#done/process moved to PLAN
- Flowershow: Flowershow is our product name for a publishing toolkit and app for "digital gardens" and, more broadly, markdown-based content rich sites
- web3 site is an example (CMS x wiki built on markdown JAMStack with rich functionality)
- DataHub Enterprise is the enterprise data management system from Datopian. Main work here is building the marketing site and materials for this.
- "Tao Plus" (the envisioned future community space/notebook/wiki/guide). Project outline: https://github.com/life-itself/community/blob/main/projects/digital-garden-2022.md
- Current tao will merge into this
- Could be (another) try-out for Flowershow
- Web3 improvements
- Ecosystem improvements
Someday
- lifeitself.us to lifeitself.org migration 🔽 (not urgent or that important. can wait for an opportune and quiet moment e.g. the summer)
High Level Order of Work
- DataHub Enterprise Website specifically landing page at datahub.io/enterprise for the product so that Datopian can already start referring to this (this is needed now and (relatively) easy to do)
- Basic flowershow. This means:
- Landing page for offer on datahub.io datahub.io/flowershow
- Offer distilled
- Mockup and implemnetation
- Skeleton app
- Tutorial (could be a demo in gitpod or whatever)
- Landing page for offer on datahub.io datahub.io/flowershow
- Implementation / dogfooding of Flowershow
- Tao implementation based on Flowershow alpha
- Web3 improvements
- Ecosystem
- etc
Next Steps
Preliminaries
- Clone these repos locally and set up first one with Obsidian
- https://github.com/datopian/product - documentation and notes repo for Datopian/DataHub products including Flowershow, DataHub DMS etc
- https://github.com/datopian/next.datahub.io - powers future datahub.io and already runs the blog
- https://github.com/datopian/nextjs-tailwind-mdx - starter template for nextjs (default we use atm)
- Explore product repo and make notes and questions
- Read PLAN
- Read this document and links from it to key product streams
- First steps
- Get next.datahub.io consistent with our other sites in terms of setup e.g. config/siteConfig.js, navbar setup etc (suggest with khalil as he knows the structure)
- flowershow skeleton
- ❓ what about using portaljs. is there really a difference? plus the monorepo approach has attractions.
- converge datahub pages with flowershow
- our standard template for nextjs in there
- think about how we add features to a bare-bones nextjs template
- Do we just make more elaborate templates? Problem with that is as add options we have exponential blowup in examples.
How would I introduce someone new to Flowershow?
Big Picture for Flowershow/DataHub Pages (70%)
- probably to Flowershow **✅2022-05-09 processed to flowershow-why-scqh **
Situation
- Context
- Needs: Digital gardens, Data Literate documents (in turn evolution of wikis etc). Generally: combination of simplicity and power.
- Simplicity: plain text documents. Can write in anything.
- Power: markdown-plus + javascript-based rich components for anything you want + workflows (e.g. integrating bibliography)
- Lego-ization, open-source, better UI of powerful tools (like git)
- Tech: JAMStack approach especially with markdown and nextjs. Plus things in background like git (the database layer).
- JADStack = Javascript, APIs and Data! (Data is CSV)
- Needs: Digital gardens, Data Literate documents (in turn evolution of wikis etc). Generally: combination of simplicity and power.
- Providing a tool that makes it easy to publish yet provides you the raw access and power
- Obsidian is a perfect example. Uses tech that was (and is) quite geeky … yet making it more accessible (e.g. WYSIWYG editor for markdown) whilst at same time preserving the power (plugins, templating and more). Example (perhaps) of innovator's dilemma type dynamics …
- Markdown
- Plugin architecture
- Aside: innovators dilemma for CMS: JAMStack gradually overtakes more "monolithic" CMS (or document editors in case of Obsidian)
- Several convergences
- Content and Data: diminishing differences at least on publishing side
- Online and offline. Desktop and cloud. e.g. using electron apps to build desktop.
- Javascript: everything done in javascript. Always bet on javascript.
- Plain text.
- Git. DAG-style models for versioning
- Cloud-first (?)
- (Personal) Knowledge Management and Content Management e.g. dissolving difference between a raw notebook and wiki and "polished" website
- This approach has several features that are hard to reproduce in existing systems
- Proper versioning support which then supports innovative contribution workflows (e.g. merges, including merges of trees rather than single documents)
- Portability of content (markdown is plain text and remember unix philosophy: text as a universal interface)
- Content is processable with standard tooling
- Good decomposability of parts (componentized vs monolith like microsoft word or even wordpress)
- Built in scripting language with massive adoption and support
- Basic point was that UI was a major barrier in the past e.g. really hard to use git, hard to edit markdown etc … tooling is getting better and better …
Complication
- Still painful to publish for many people (in the way they want).
- What does it take to publish your obsidian vault?
- Either use the built-in app. Can't customize. Standard template. Limited features.
- Or put in git(hub), setup my own theme and publishing system using JAMStack system, set up deployment … and maintain it.
- How do i turn a markdown file into an elegant site?
- Can go to medium (?) or some markdown oriented publishing app (are there any?). Tweak theme. Probably pay or not own my content …
- Set up myself (see previous point)
- How do I put my CSV file or dataset online?
- Steps of putting a CVS nicely online:
- Nerdy approach: 1) create repo 2) commit CSV and README.md to repo 3) setup some kind of publishing JAMStack system 4) add support for showing the CSV in html 4) setup deployment (e.g. github pages, netlify etc) 5) share …
- Steps of putting a CVS nicely online:
- What does it take to publish your obsidian vault?
- Want something that is open-source-y in that people can extend and improve it … and a way to build a sustainable business
Question: Is there a way that BOTH gives me ownership and control of my content (and flexibility how I create and maintain it e.g. with content in standard format) AND allows me to publish quickly and elegantly and reliably (and without too much technical knowledge)?
Hypothesis: FlowerShow / DataHub Pages
Once upon a time version (0% ignore)
Once upon a time …
People made notes on paper
They published newsletters and newspapers and journal articles.
And then
Digital technology arrived and they moved online. Often reproducing the old "paper" world online.
And then
And because of that
And finally …