CORS: People often want to connect to web resources by URL using either JS fetch() or Python pd.read_csv('https://...'). They'll run into CORS issues. Starboard works around this by providing a free CORS service, similar to https://cors-anywhere.herokuapp.com/.
Excessive traffic: In notebooks everywhere (Jupyter, etc) people often re-run cells multiple times and this is so expensive but not well understood by people new to the concept/field. It's difficult to cache the download and avoid that traffic from within the notebook runtime.
Solution
Use Deno to make network requests, and also save the requests to disk to act as a cache.
Hiccups
Cache invalidation is hard. The users will know best (sometimes) if they need a fresh copy. There needs to be a thing in the client to show what assets are cached (date, size, cell, references, etc) and the ability to clear items from the cache.
Problem
There are two:
fetch()
or Pythonpd.read_csv('https://...')
. They'll run into CORS issues. Starboard works around this by providing a free CORS service, similar to https://cors-anywhere.herokuapp.com/.Solution
Use Deno to make network requests, and also save the requests to disk to act as a cache.
Hiccups
Cache invalidation is hard. The users will know best (sometimes) if they need a fresh copy. There needs to be a thing in the client to show what assets are cached (date, size, cell, references, etc) and the ability to clear items from the cache.