Hm.
So, without a kernel that can do heavily lifting, you're limited to basically what you get when have the browser console open; ie. html + js.
Is that the idea? A better / structured browser console?
Or am I missing something?
I mean, I completely appreciate the benefit of notebooks; but if you're limited to running code inside your browser window sandbox... how do you do anything useful with it?
How do you import a dataset?
> const data = await fetch(`https://api.exchangerate.host/timeseries?start_date=2020-01-...
Is great for trivial public datasets, but CORS is going to forbid you from loading any personal/private data right?
unpkg can give you public packages, but if you have any personal code, how do you use it?
It seems like you basically end up with an empty sandbox.
I somehow felt very similarly. In that this is a very cool project, but what does it really solve. People who work with data also actually need Python, do you really expect me to recreate all the Py libs in JS now?
I am really curious what was the author's thinking process here with regards to solving the problem, what in fact was even the problem. Perhaps the author could share their thoughts.
Edit: grammar
After looking at https://starboard.gg/about things are clearer now. The use-cases are clear and I can see the audience it targets.
Can you explain it, because even after reading the about page I think your objections in the first post stand unchallanged.
To me it looks like while Jupyter is good with Big Data, this may be good with Small Data that doesn't need the overhead that is Jupyter, but saves the need to reinvent all the wheels yourself manually.
I'm looking at this and comparing it to the situation I find myself in with looking for a syslog collector. I want something that can collect all my syslog and present them in a web UI in a sensible way.
My options appear to be limited to commercial solutions, ELK (Elastic, Logsearch, Kibana), or text files. There's no middle ground. I have ~20 VMs + containers, and no way to view all of their logs in a central sensible manner unless I roll my own (or buy more hardware to be able to run the ELK cluster).
I feel like I must be missing something here; you want to aggregate your syslog data; you have a bunch of files and a web browser.
Now what?
You beat your keyboard against the monitor while your browser sandbox steadfastly refuses to load your local file system files? [1]
You write your own server that serves data and the open source core? [2]
Yay, you now have the dubious honour of having re-invented jupyter notebooks, only, you had to roll it yourself, and it's insecure and really not very good compared to an existing mature product.
What can you actually do with this, that you can't do by right clicking on your browser window and picking 'inspect', and then clicking on 'console'?
You can certainly make interactive blog posts; absolutely, I 100% acknowledge that as a blog plugin, when you have already got a server to host your content, this would be pretty awesome.
If you do want to run your own databricks / notebook SAS, then I also see it as being a really great kickstart for you, absolutely.
However, out of the box? I'm... like, I think this is interesting, but I don't see how I could actually use it.
[1] - https://stackoverflow.com/questions/50007055/fetch-request-t...