[Grants] Re: [LACTTE] Grant Application from Russell Coker for FOSS on Mobile Phones

[Russell Coker]

On Wednesday, 10 May 2023 22:47:19 AEST Jonathan Woithe wrote:
> After reading through the "FOSS for mobile phones" grant application I have
> a few questions to help clarify what is planned for this project.
> 
> The application outlines a significant number of tasks: LORA support,
> refinements needed to FOSS phones so they can be used as a "daily driver",
> security enhancements, and convergence to name a few.  Each one of these
> appears to be fairly involved and won't all be completed quickly.  Is it
> possible to provide a prioritised list of the work plan, a rough idea as to
> when the community might see the results, and an outline of how the
> community can expect to benefit from the work facilitated by this grant?

Part of the reason for the number of tasks is the differing goals that Yifei 
and I have.  Yifei's interests are radio and I18N of which radio will be the 
first priority and the one with the most obvious success criteria.

My interests are a phone that's usable for daily use (which involves 
improvements to battery life and getting the standard functions like contacts 
listing and integration working) as well as convergence and security.

Convergence has some dependencies on things that I don't have the skills to 
directly work on (particularly Wayland).  But some of the work has been done 
in this regard by Purism people (their CTO has blogged about having 
convergence working on the Librem5).  My tests of convergence on the Librem5 
indicate that there may have been some bugs in recent versions of the software 
that prevents previous functionality working correctly, and also that there is 
some hardware compatibility issues that need to be tested and tracked.  I 
currently possess three USB-C docks which all have different levels of 
functionality with the Librem5, having another phone to test with will help in 
investigating those issues.

My plans for security improvements have some elements that can be completed 
reasonably quickly.  Getting the sandboxing of applications basically working 
with firejail is easy.  I've started investigating the use of xdg-desktop-
portal and ways of integrating that into a sandboxed environment.  Yifei and I 
have been discussing ways of testing the sandbox (basically running a shell 
inside a sandbox configured in the same way as an application sandbox and 
verifying that it can't perform certain classes of inappropriate activity).  
One thing to do in the near future is to determine the best way of sandboxing 
"desktop" (or in this case phone) apps, snap, bubblewrap, and something like 
what Chromium does are all things to consider.  Documenting the differences 
between these different alternatives might end up being a noteworthy 
milestone.

While Yifei and I have plans that differ significantly I'm submitting this as 
a single grant application because one aim is to start a community for FOSS 
phone development in Australia.

TLDR:

Ordering for me will be:
1) Improving phone utility for daily use.  Finding ways of getting it to work 
well, documenting the configuration, and developing tests for regression.
2) Convergence.  One of the first things to start but will face some delays 
due to Wayland and hardware driver support.
3) Sandboxing of apps.  I'm already making progress on investigating this.

> It is noted that "The work needed to achieve the aim of making the PinePhone
> work well as a regular use mobile phone depends on what areas it has
> problems with."  Have you investigated what these might be, as I expect
> this would facilitate your planning?

Poor battery life is a known issue, and a "fix" to poor battery life causing 
problems in receiving calls is another known issue.  I will discover how bad 
these problems are when I get access to hardware.

> Another stated aim is to "develop a community of developers for mobile FOSS
> in Australia".  Are there specific ideas about how this might be done?

I think that when Yifei and I start getting some work done in this area it 
will interest other people to join in.  When I can make a FOSS phone my daily 
driver it will inspire others to do the same.

> The technical goals of the project are not clear from what was written in
> the grant application.  Is it possible to provide more clarity around this?

The problems with making a clear technical description are the difficulty in 
determining what the problems might be without having access to the hardware 
and the fact that the work will be partly determined by what aspects of the 
functionality other people have got working.

I think that the best thing to do is to have a list of things to work on, if 
one area of work is unavailable then I can work on the others and consider it 
a success if good progress is made on them.

> Finally, I note that there are other pinephone models which are half the
> price of the Pinephone Pro Explorer.  Both are Linux-based phones, but there
> must be specific reasons why the grant requests the "Pro" version
> specifically.  Could those reasons be shared?

The Pro version has faster CPU cores and more RAM which will be a significant 
benefit for compiling on the phone and for use in convergence.

But I have to admit that I just like newer fancier hardware.  So we can 
probably get it all done with the cheaper one.

-- 
My Main Blog         http://etbe.coker.com.au/
My Documents Blog    http://doc.coker.com.au/