Re: [Ring] Future of platform-specific clients?

[Anthony Léonard]

Hi Tirifto,

> 1. Integration. Cross-platform software with a single unique
> interface
> tends to look and feel quite out-of-place on some systems, making it
> uncomfortable to use. But if handled well, this shouldn't be major.

That's right. By adopting a unique look-and-feel there's a great chance
that you'll not respect design guidelines if those are in conflict
between 2 systems. I'm not a design guy so I don't know how to handle
that but I suppose it is about making compromises.

The advantage of using Electron is that the HTML/CSS/JS stack is easier
to learn and is good at making UI. Moreover, I tend to think that
people involved in design are more HTML/CSS-aware than C/C++ but I may
be incorrect. If I'm right on that, it is also a way to make Ring more
accessible to contributors potentialy more knowledgeable about design
and ergonomics which is nice.

> 2. Performance. I've read often enough that programs using Electron
> tend to be slow and resource-heavy, also owing to that the program is
> running in a dedicated “web browser” for itself. I find web-based
> programs slower in general, so I imagine this could present a huge
> issue.

I see at least one point on which the performance concern could be
challenging. It is how to pass video frames efficiently between the
daemon and Electron. I don't know about Electron very well but we
currently decode video frames and then pass those raw frames (by buffer
pointers for efficiency) to UI frameworks. I don't know if we can do
that easily with the Electron/WebKit renderer.

Another way could be to rely on decoding facilities from the web engine
 itself but it's a major change in the whole media pipeline which, if
even doable, will certainly break compatibility with other clients (if
we don't want to support 2 different pipelines).

> Ring should really strive to be light-weight.

IIRC an Electron “Hello World” is like 20-30 MB so that may be not very
light-weight.

> 3. Freedom. Again, I trust that Ring will remain entirely free, but
> I've read some doubts about the freedom of Chromium and Electron.
> They've even been removed from the Parabola repositories. In that
> case,
> it sure wouldn't be used for the official client; I just hope the
> developers are well informed on the matter and have better plans
> ready.

Again IIRC, even compiling Chromium, which is supposed to be the 
free-ish brother of Chrome, requires downloading blobs in the process.
I don't know how much of Chromium is used in Electron but if it also
has this requirement that's definitely a problem.

Anyway, somebody with more knowledge about that should probably answer
those concerns and/or correct me if I'm wrong. I know very little about
this new project.

Anthony L.

signature.asc
Description: This is a digitally signed message part