Re: GSOC idea

[Carlos Soriano <carlos soriano89 gmail com>]

Hello Анастасия!

That's an interesting project and it's so good that you had the
initiative to come up with your idea! I've worked on the past on
gnome-shell as GSoC student and now I'm part of the GSoC admin team,
so I'll try to give my input.

The idea sounds very good and your previous work you already did in
this area is amazing, so I'm sure someone would be glad to mentor this
project. The current maintainer of gnome-shell is Florian Mullner, but
I'm not sure if he has time for this GSoC round to mentor. I'll let
him chime in the conversation for that. If that's not possible, would
be possible to you any other time or internship like Outreachy in a
few months (if you are eligible)?

Regarding the idea itself, I can see there can be two phases. One is
to improve the current keyboard to what is missing from the designs on
https://wiki.gnome.org/Design/OS/ScreenKeyboard. The bigger tasks I
can find are:
- Search
- Emojis
- Popover for choosing the language
- Some way to open up the keyboard from gnome-shell (needs design input)
Those could take maybe the first phase of the GSoC (1/1'5 months) for
example. But then there are some smaller tasks, for example:
- Changing the color of some of the buttons to blue
- Ability to have permanent caps
And these two are pretty good for you to get started into the
contribution of this module, and in this way you will have more
contributions and a stronger proposal to be accepted (we require at
least one non-trivial patch, but definitely more will increase
enormously your chances for the mentor to accept you) :)
Then as second phase of GSoC, when you are already kind of an expert
on that area of the code, you could implement your proposal of
predictive input, which is something everyone of us will feel in love
with :)

What do you think?

Carlos Soriano

On Wed, Mar 15, 2017 at 8:47 AM, Анастасия Маркина
<asyamarkina2 gmail com> wrote:
> Hello,
>
>
> (writing here by advice of Bastien Nocera).
>
>
> I have an idea for Google Summer of Code related to improvements in the
> on-screen keyboard. I see this idea as a really important and useful feature
> to implement, and have already done some background work (including testing
> the input efficiency). Perhaps GSOC project would be a great help :)
>
>
> The proposition is concerned implementing additional functionality for the
> Gnome on-screen keyboard (caribou): heuristics-based autocompletion and
> spelling corrections, following the model of Android keyboard (taken from
> AOSP project).
>
>
> I'm a computer science student of Brest State Technical University
> (Belarus). During last few years I actively participate in the open source
> community. Particularly I’m co-organizer of the LVEE conference
> (http://lvee.org) - a central FLOSS event in Belarus and the largest
> general-topic Linux-related regular conference in Russian-speaking
> countries. Also I’m addicted to the ergonomics of a desktop software, and
> the main developer of the free/libre software suite for parallel usability
> testing, https://bitbucket.org/AsyaAliset/uxdump).
>
>
> More about GSOC Idea:
>
> I’ve used GNOME on a tablet for some period and have a suspicion that
> on-screen keyboard has more space to improve user’s touchscreen experience,
> in comparison with other parts. Currently it is usable, but... little bit
> lifeless (same is true for on-screen keyboards of other desktop
> environments). So I’ve done some testing.
>
>
> Tests:
>
> I’ve carried out the comparison of a typical on-screen Linux keyboard
> (onboard from Ubuntu 16.04 was used as a laboratory rat) with the Android
> one (default open source keyboard from AOSP). Typing experience of about 20
> people was tested with use of some biometric measurements to evaluate
> comfortness with different types of text.
>
>
> The result of the comparison was presented this autumn on the Russian open
> source conference OSSDEVCONF, so I have slides to share -
> https://lvee.org/uploads/image_upload/file/441/Keyboard_as_Co-Author.pdf
> This is a specially created English version :)
>
> It can be seen that in most cases a heuristic-driven keyboard provides
> faster typing, lower error rate and lesser physical load even in spite of
> some wrong autocorrections.
>
>
> I’ve looked through https://wiki.gnome.org/Design/OS/ScreenKeyboard and have
> found that I’m not the only person who thinks about adding to Linux desktop
> some useful features of mobile keyboards (of course, their open source
> nature simplifies this task a lot) :)
>
>
> Finally I’ve looked through caribou sources, and feel capable to deal with
> this task.
>
> I haven’t coded in vala yet, but it would not be a problem after a number of
> other programming languages, from rather universal C/C++/C#, python, etc. to
> some exotic (of course I can do some patch or bug fix as a proof).
>
>
> I would be happy to add my five cents to making GNU/Linux and Gnome viable
> on tablets :)
>
>
>
> _______________________________________________
> gnome-soc-list mailing list
> gnome-soc-list gnome org
> https://mail.gnome.org/mailman/listinfo/gnome-soc-list