Hi everyone, I'm a part of the Ton-err RunKit team and would be more than happy to answer any questions!
What makes this the exception?
After their two years at Stripe are up, they can go back to working on it if it still makes sense.
Edit: Not sure why this is getting downvoted. I'm only pointing this out to help.
I have two questions for you.
1) Offline mode, please! I've opened the feedback on uservoice at October 27, 2015 (https://tonicdev.uservoice.com/forums/273722-general/suggest...), there should be some self-hosted version for us to use. Which kind of leads me to what I think is blocking the offline mode
2) module-fs < is this planned to open source at one point? Sounds like a interesting solution, what I've done before is to hook up require to be async and then download the packages when they are being used, but not very friendly
Offline mode is an interesting problem. For example, module-fs is 1 part code, 99 parts infrastructure really. We have 3.5TB of installed packages, stored internally in a content-addressable way, and then exposed as a file system through FUSE. module-fs seems to be something that only becomes powerful when you have ALL packages, thus really having similar performance characteristics of having already installed the packages. The solutions I've seen that try to download packages in the background, etc. don't quite provide this magical "I only need to require it" feeling. Its similar to the issue we have with providing real time-traveling locally: in my experience "time-traveling" has never taken off because its eternally stuck in demo-mode, as all code eventually interacts with the outside world. By snapshotting the entire computer, I think we've pushed this feature over the edge to the point that people use it without realizing it on RunKit, it is of course natural that going back in a REPL also goes back to a previous state. But locally, this feature once again doesn't seem to immediately make sense. At that point the question becomes what part of RunKit would we offer locally? Just the object viewers? I ask this non-rhetorically btw, I am happy to hear what you imagine a "local" RunKit to look like, and if it makes sense we'd certainly consider pursuing it.
Do you plan on branching out to other languages with Notebook (Ruby!)? Also any plans to open-source any of your stack? In particular, I would love to have an on-prem version, then I could reach internal resource and store sensitive data.
Here is a simple multi-language scratch pad that can be saved to a URL endpoint: https://hook.io/editor
`module-fs` is mentioned in the article, is the source for this available?
Can you elaborate on your roadmap and what parts of the stack - if any - you are planning on open sourcing?
Thanks!
What can we expect next from you?