so there were few talks about UI, including some old Vyatta UI.
Just wondering what would we like to have there?
So many use cases out there, on what we should focus?
It's very interesting indeed. Its licensing is a total mystery. ;)
To start with:
- There is no LICENSE/COPYING file in the repository. There is AGPLv3 buried deep in docs/ dir, but see the next point
- Source files lack any copyright headers, so it's impossible to find out what exactly is under AGPLv3
- They use double licensing, which means contributors are made to transfer the copyright to the original authors and they can close the source and continue to develop it as a proprietary product any time
- AGPL in no way prevents anyone from using it for a frontend for a proprietary application. People can comply with it by giving people open source code for an application useless without its proprietary backend.
Also, it uses gevent, which is widely considered harmful. I would think twice.
I have might missed this in previous tracks of conversations but since we know that everyone will most probably want a different flavour of UI why not approach this differently ? Why not have API calls ( configurable with appropriate security ) and then enable people to build UIs on top of this base of APIs .
Then what we could focus on would be making let's say base dashboard UI which will be modular enough to be extended by needs of people ( kinda plugin architecture )
Hello @RafPe, it's never late to join.
API is something that considered apart, and your approach is logic in all means.
API part is challenging due to the architecture of backend, while we work to get something that we can call API, it's clear that it will not be an elegant solution.
This what we going to address in v2
Basically, under this question I looking to some feedback regarding UI for SMB/ROBO types of install
where people looking to have GUI
In this instance if we are going to generate config I would maybe implement something in lines on simple UI and config api validator ? Then anyone could extend UI and program against backend validator before it would be committed
IMO we should wait for v2.0 because of this. it would be a lot of work to get a half decent UI, and v1 isn't really capable of giving a nice API to it.
Put any sort of UI on the v2.0 roadmap, and make sure v2.0 has a good API would be my vote.
A question was more like conceptual
UI/UX design is also good chunk of work
I more concerned about general elements and how it should be placed together
I think we don't need yet another shity web gui,
i prefer if it delivers some value instead