I don’t know if this is the right place to post this. But I don’t know where to put it on this forum. And I think this discussion is important. It needs to be opened. So…
Notebooks, Framework and now Canevas. So many new features. But Observable cloud depreciated. Notebook gratuity, then no, then yes. So many zigzags too. Where is Observable going? I confess I don’t know anymore. For me, Notebooks are a really great tool. It’s Observable’s strong point. I love them. Notebooks create a community. They enable everyone to progress and help the ecosystem to grow. But I get the impression that it’s not evolving any more. No new libs. No new features. It’s as if the Observable core team has stopped making it a priority. I hope I’m wrong. Anyway, I think that would be a big mistake.
That’s my feeling. Answers and thoughts on the situation are welcome.
The reason I am building a single file notebook format which does not require a server and is compatible with Observable is so that if the worst happens, I have a bunch of working, self-editable notebooks to recover from, in a format that is 100% local-first, which works the way I love.
The repository holding the source code to the file format and backup infrastructure is here. My personal notebooks are backed up here.
I am quite happy with using Observable as my primary code store. They have communicated the notebooks will not receive new development but will stick around(don’t think they actually said that, sorry for confusing things), but I also think its wise to always have an insurance policy. Lopecode is my insurance policy. If you would like to use it too and you need docs/features/support, open a ticket on the Github.
Some open questions on the Framework discussion board as well. In hindsight what really took some air out of the community was the flip-flopping on the notebook publishing defaults and paid plans as well as less prominently featuring trending notebooks.
On the tech side, I remember the mobile editor being more stable in the past (pre 2022), as for a while it was kind of difficult to place the cursor and insert linebreaks, often jumping back and forth (Chrome Android, I’ve sinced used the mobile editor less, so this may have been fixed).
I think monetisation could have been handled with a user Marketplace instead, offering plugins, templates, bespoke visualisations, etc. for others to try or buy (a bit like buying custom theming from e-commerce websites). We know from notetaking tools as Obsidian that offering a standardised plugin platform can draw in a lot of would-be users and would make userland UX modifications less hacky than they are.
Larger clients would likely pay more if some form of distributed (GPU) compute was offered as an additional back-end to notebooks, to entice the model training crowd. E.g. tight integration with existing libraries, or an easy bridge to interface with other language ‘cells’ so that non-Javascript packacges could be used. Framework is on the right track here with its dataloaders, but may not offer customers much that need to distribute their compute workloads. This is where a cloud service can come in.
In any case, I also am onboard with notebooks still, and hope to see them maintained in the future.
No new developments, this means the downfall of notebooks in the future. Will I continue to teach this to my students and continue to use notebooks myself? Not sure… I’ll have to think about it.
Sorry, I can’t find any evidence they said no future development. Last noteworthy thing was that they would find a way to update Notebooks to be Framework compatible but I can’t find that either.
thanks @neocarto for posting this. I have the same feeling you expressed and am also evaluating how much to invest in new notebooks (which I love).
Thanks @tomlarkworthy for sharing jumpgate - I will happily try it out!
I wrote about this at some length on my Framework blog. I guess that’ll be gone once Observable.cloud goes but the key points are
Observable Framework is an awesome new product that excels at producing fast, response, interactive web-based visualizations,
the Observable notebook is an older product that provides an incredible community environment for quickly prototyping the same kinds of visualizations in a notebook, and
the two products could and should complement and enhance one another beautifully.
By this point, we should see an Observable Notebook 2.0 with syntax that’s fully consistent with Framework. Users should be able to export their notebooks to Framework projects that are ready to be compiled to production quality websites. That setup would provide a unique path for new users to get into serious development.
Tech companies have, for decades, integrated new products with existing ones to leverage their developed tech stack and user loyalty to promote those new products. The strategy has been so successful, that governments have from time to time had to intervene. In my view, it is a tremendous loss that Observable has not found a way to do that with the Notebook.
Having said all that, there are some truly amazingly creative folks behind Observable and the open source ecosystem surrounding it, i.e. Framework, D3, Plot, etc. That’s where my energy is going these days.
Will chime in as I’m a very causal user so I see changes late and after the fact. Notebooks, Plot have been absolute rockstar products IMO. No complaints and I have loved them. I started on D3 on my local machine and graduated to Notebooks (which had a learning curve, but it was worth it).
Framework is ‘ok’; I find it confusing and non-intuitive (I am NOT a SWE). Didn’t use or understand Cloud at all. Ditto with canvas. I find it unlikely people are collaborating developing visuals (?) but am open to being wrong.
I was hoping there would be a local only Notebook or some sort of standalone integrated IDE, Ala Rstudio.
Will follow this thread for opinions. Thought Tomlarkworthy was a Observable employee this whole time; thanks for the help over the years Tom.