Hacker News new | past | comments | ask | show | jobs | submit login

I don't get it. Why not just use Plotly and Dash directly, neither of which are especially hard? This seems like a pointless layer of abstraction.



Mainly because you won't have to write as much glue code.


I think it's two pieces:

1) Removal of the need for complex glue code (filters, callbacks, etc.). This is nice so less sophisticated users can develop dashboards.

2) Easy hosting / deployment. I don't want that in the same system (I use dash in my own apps, and it's nice I'm not tied to much of anything else)

That's me extrapolating from a little bit of poking. I could be wrong.


Thank you for your feedback! I am one of the engineers on team :)

So we introduced this layer to serve as a configuration driven assembly system which implements built-in visual styling and an opinionated approach in terms of both application architecture and UX, in order to enable standardisation in code and user journeys for multi-screen applications and complex interactions between components.


This looks nice! Critical feedback:

- Spending about 5-10 minutes on your page, I also couldn't figure out why or what you were trying to do. The copy you wrote above is perhaps clearer, but not clear.

- You have less than 5-10 minutes to sell me on what you're doing.

- I began to get it when I clicked on the screenshot, and saw controls with vm.Filter(column=...

- To use this myself, I'd want to be able to add it incrementally to dash / plot.ly apps. You don't want to force yourself to be "the framework." You want to be able to slot into diverse places. The different pieces should be modular enough to do that.

- In one of your other posts, you wrote: "while it is currently optimised to leverage Plotly and Dash, it could theoretically utilise components from other packages such as StreamLit in future." This seems like a very bad idea. Plotly and dash are well-architected and expandable (it's trivial to integrate components from other sources with React). You don't want to reinvent the wheels plotly/dash invented. If you want to slot in other components, please use their architecture.

In other words, clearly communicate your value-add, and focus on that value-add. Don't try to do / be everything. Things outside of that value-add should be PRs back into dash / plot.ly, or independent, modular pieces in their own repos.


Thank you very much for the feedback!

We are definitely taking these points into consideration.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: