Inbuilt tracing is really cool for rapid prototyping

[pbiggar]

A user says:

I've played around with Dark over the weekend, and I like it a lot so far. The team and you have put together something great, which I hope will fundamentally change how we think about developing software.

I'm writing a blog on my initial experiences, which I'll share when it's published. However, I have some preliminary dotpoints for you:

Things that make me feel like this is the future of software development:

• deployless is the logical next step to continuous delivery. it feels great.
• inbuilt tracing is really cool for rapid prototyping, getting immediate feedback on new functionality as i write it is awesome
• canvas components make a lot of sense. I didn't feel like I was missing anything so far.

Some things that make me not want to use it in production just yet:

• current lack of packages/libraries means reinventing the wheel every now and then (for instance HTML parsing, rate limiting)

• no convenient way to share code
• no observability tools other than tracing. i would like to see graphs depicting request-rates per endpoint, statuscodes, responsetimes. something that allows me to understand the health of my API at a glance. especially since im editing "live"

• lack of tools to give me confidence in the code and the changes i make (I'm thinking environments, unittesting, stubbing of third party services). some of this appears to be covered by the feature-flags which are currently unavailable if i understand correctly.

I believe most of these things are on your radar already, I've seen discussions in Slack on all of these points already, so I can't wait to see how Dark implements them :)