Discussion: Runtime Performance #33
Description
While we initially kicked off reporting stats from the Dev Time list, the second list, Run Time, has many stats that will be really hard to measure fairly.
For example, two issues that come to mind off the bat:
- Some meta-frameworks run better on certain providers
- Not all meta-frameworks are built for the same thing
A basic idea for testing these could be to reimplement the same app in each (like our starter packages, but with an app). If you asked me a year ago, I would have thought this was a huge maintenance burden, but I think it can be nicely aided by LLMs. However, the bigger question is how do we do the testing:
- Run in a CLI and do real browser testing?
- Deploy to a provider, say Cloudflare and use Playwright and other tools to test
- Both
- One I have not listed here yet