[skidding]

> I'm not going to use a project that cares more about "hype" than about delivering value

I can assure you I care a lot more about delivering value than hype. I spent countless hours (from my personal time) over the last 6 years on making React Cosmos easy to use and compatible with as many codebase configurations as possible, and maybe two weeks on the website.

The project README explains what the project does and what it doesn't: https://github.com/react-cosmos/react-cosmos/blob/main/READM...

There is an open thread about website feedback, and I agree I could've done a better job and there is still room for improvement: https://github.com/react-cosmos/react-cosmos/issues/1111

[uxcolumbo]

Hey - thanks for putting your sweat and tears into this over so many years... OSS work is vital.

Us HN folks can appear a bit grumpy at times ;)

I've got some ideas how your website could be improved... I'll have a look at the open thread.

[jitl]

Sorry for the excessive negativity. These kinds of product page for software a big pet peeve of many - light on details, heavy on marketing does not endear anyone to the grumpy HN crowd.

I think the biggest single improvement you should make is adding a paragraph above the fold that describes in as plain terms as possible what your project does, and the problems it solves. Most of your bullet points on the website describe your project as being different or better than something - but the reader has no reference for what that baseline something is.

For example, “Don’t settle for localhost:3000” - this doesn’t give me any information about what the project does, or why it’s better.

Your README is much better than the website! Here’s how I’d summarize the bullet points at the top of the README into a paragraph you can put above the fold on your site:

React Cosmos is a component development, test, and preview environment separate from your application. Iterate on individual components in isolation. Build a library of test component props & states that you can run visual regression tests against. Cosmos is flexible enough for use with different bundlers and build systems.

[steve_adams_86]

Conversely, cosmos is targeted at me and the lander made good sense. I understand the frustration with “inside” comments like “Don’t settle for localhost:3000”, but I appreciated it and it implied fairly heavily that this tool disrupts a very slow and tedious workflow.

I do agree that some things are buried. The docs didn’t lead me to the information I needed as quickly as I think they could have for example.

Overall I’m excited to give this a shot. It seems like a great tool.

[jitl]

I think I’m also in the target area - I work on localhost:3000 every day building Notion, which is a big React app. But why does changing the port change my workflow? What is the “default” state of the service on port :3000? How am I settling? What workflow are you referring to when you say “this tools disrupts a [...] workflow”

[steve_adams_86]

This is my take on it, but I realize now that it’s going to vary a lot from person to person.

The workflow is all the jank that comes with HMR, or worse, reloading entire pages to see changes.

The port number isn’t a specific thing so much as an allusion to the workflow most people use. You’re focused on a single component quite often, but working with it in an entire page served through your entire app server. It’s overkill and slower than necessary, and sometimes comes with all kinds of state frustrations outside of the scope of your component.

Changing the port doesn’t inherently change the workflow, it’s just calling out which port most of us use for that workflow and suggesting to investigate what could be better than that workflow.

They’ve made some changes to the page now, so hopefully this isn’t a source of confusion anymore.

[skidding]

Couldn't have said it better myself, thank you!