On Sun, Apr 2, 2017 at 5:41 PM, Joanna Rutkowska
<[email protected]> wrote:
> On Sun, Apr 02, 2017 at 02:39:28PM -0400, Jean-Philippe Ouellet wrote:
>> On Mon, Nov 21, 2016 at 11:13 PM, Marek Marczykowski-Górecki
>> <[email protected]> wrote:
>> > On Mon, Nov 21, 2016 at 10:57:25PM -0500, Jean-Philippe Ouellet wrote:
>> >> On Mon, Nov 21, 2016 at 10:14 PM, Marek Marczykowski-Górecki
>> >> <[email protected]> wrote:
>> >> > Generally yes, especially the part about choosing GTK ;)
>> >>
>> >> Noted.
>> >>
>> >> > Since then we've moved to default Xfce, and GTK will be more consistent 
>> >> > here.
>> >>
>> >> Are you saying that Xfce as default is the long term plan now?
>> >>
>> >> Or saying that there will be more consistency between {GTK in Xfce vs.
>> >> GTK in Gnome} than {Qt in Xfce vs. Qt in Gnome}?
>> >>
>> >> I don't understand.
>> >
>> > Both Gnome and Xfce use GTK widgets, so GUI parts done in GTK (vs Qt)
>> > are more consistent with the rest of desktop environment.
>>
>> [1] has made me give some more thought to this. Yes, GTK is the
>> natural choice for things wishing to be first-class citizens in Gnome,
>> but this may not be the strongest consideration long-term.
>>
>> Moving to Gnome has stagnated, and is a very low priority [2].
>>
>> With the planned disaggregation of the GUI into its own domain [3],
>> iteration in the desktop experience of Qubes becomes much easier, and
>> I think it's quite reasonable to consider that Gnome may not be the
>> only final target.
>>
>> I think it is also worthwhile to consider the difference in experience
>> between Qt in primarily-not-Qt environments, vs. GTK in
>> primarily-not-GTK environments. On other non-linux platforms, Qt is
>> much better integrated into the native desktop UX than GTK. This is at
>> least true on Windows and OS X, and even Genode has native support for
>> Qt which continues to improve. This may may sound irrelevant now, but
>> I think is less so post-GuiVM.
>>
>> FWIW it seems that peoples' complaints about the current Qt
>> qubes-manager in Xfce have nothing to do with it being Qt instead of
>> GTK.
>>
>> I'd also like to bring attention to some relevant 3rd party
>> discussions on this: [4]
>>
>> I believe that limiting the inputs to this decision to "Gnome means
>> GTK, we plan to eventually target Gnome, therefore let's use GTK" may
>> be short-sighted.
>>
>> Thoughts?
>>
>> [1]: https://groups.google.com/d/msg/qubes-devel/jDHiZ9nhzIc/5DIRzcOUCwAJ
>> [2]: 
>> https://github.com/QubesOS/qubes-issues/issues/1806#issuecomment-280281571
>> [3]: https://github.com/QubesOS/qubes-issues/issues/833
>> [4]: 
>> https://askubuntu.com/questions/281092/why-is-canonical-choosing-qt-over-gtk-for-unitys-next-generation
>
> I think it's difficult and often counter-productive to try to over-generalize 
> or
> make code over-portable "just in case".

Noted. I certainly agree with that principle in general. [1] :)

[1]: 
http://opensslrampage.org/post/83031733755/remove-support-for-big-endian-i386-and

> So, I think that the choice of the GUI framework should be dictated
> primarily by the requirement to make our GUI stuff first-class citizens of
> whatever Desktop Environment we want to be the official one in the upcoming
> release. At this time we believe that for both 4.0 and 4.1 this will be Xfce4.
> The next criterion should be ease of development/integration with current 
> code.

Makes sense.

> Perhaps for Qubes 5.x we will want to switch to some yet-unknown desktop
> environment, but then, chances are high, we will want to rewrite any GUI stuff
> anyway.

The reason I raise this discussion is in hopes of reducing the
likelihood that we find ourselves in a position where we wish to
re-write things in the first place.

If you believe it is inevitable, then I agree that optimizing for the
next concrete plan makes the most sense.

> This is because the new environment might use different UX paradigms, or
> maybe in a few years time the general UX paradigm will change. So, ISTM, 
> trying
> to plan too much ahead (which framework/lang to use for GUI/UX), might often 
> not
> be worth it.

Noted. The current plan sounds good to me too then.

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-devel/CABQWM_BgfZtamgL-OZ%3Dr9x2VT7XYYGYV1aOMjKBsis8dZ%3D5%2BTg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to