Gtk 4.0 is not Gtk 4

[Steve Dodier-Lazaro]

Hey Ali,

> what worries me more is the rapid API/ABI breakages rate. It is true that
Xfce port to Gtk3 did not happen yet, but I think it is sound to discuss
the gtk future plans here, and to share opinions.

Let's look at the reality of GTK+3. Had they already adopted this
convention, we would probably be looking at the first stable release of
GTK+ 7. This cycle has been a display of madness, they've demonstrated
exactly the value of software projects where people finish their design
prototypes *before* they deploy stable releases to their user base.

Of course there is a value divide between the Xfce and GNOME communities on
the balance between innovation and reliability, and ultimately it's up to
end users to decide what they prefer. One finding from that survey I
conducted last year (no, don't ask for full findings; I had other more
urgent priorities re: xfce) was that Xfce has a lot of users who joined it
recently. Whilst I don't know how well we perform on user retainment I can
say that we have gained a lot of users, most likely due to our commitment
to not screwing with our users' UX every six months.

Now. We know how fast GTK+ breaks. We believe we are valued for our
commitment to reliable UX. The fact that our upstream decides that they
will no longer break a stable release in such a way that Xfce users end up
suffering from it is *excellent news*. It doesn't matter if they release
GTK+7 or 27 as long as GTK+3 works. And we can then decide what the next
porting target is, based on which versions we consider to bring more UX
advantages than drawbacks. I am very critical of some GTK+ prototypes, and
very appreciative of others. Having more properly supported stable versions
allows us a wider range of options when it comes to providing our users
with what we think is best.

Is this the best course of action for Xfce? In absolute, no. In practice...
we can't write our own toolkit, and porting to Qt would be a hell of an
expense (which core devs voted against when the question was brought up).

> My overall feeling these days is that I don't trust those guys anymore,
and I fell myself demotivated to write gtk code that the gtk devs will make
obsolete in a second, with continuous breakage even between minor
releases...

I feel the same way. It's really hard to justify to myself writing new GTK+
code (and I don't on my standalone UI projects). However, the new
versioning is a guarantee that we won't have to go through such an
unreliable 'stable' release cycle ever again. Thus, it takes away our main
issue with GTK+, free of cost for us.

Another issue I've had is that the GTK+ devs couldn't decide how to design
one thing or another (e.g. theming, and input device handling) and ended up
writing 3 different APIs in just a few years. With more separate stable
releases, we can see what's coming in the next unstable version and decide
whether it's worth porting to a new API. If the GTK+ devs decide to throw
their new API away and re-re-re-design it once we're almost done porting,
at least that no longer means our current porting work is obsoleted -- only
the time left to finish the port is now wasted time.

So ultimately I think we should still aim to release a GTK+ 3 port ASAP.
Even though I'm still taken by my research and unable to contribute for the
next 9 months or so, I've got some work lined up which I'd absolutely love
to deliver to Xfce, including an alpha prototype of a fully sandboxed
desktop environment. Since we're unlikely to face such a challenge as
GTK2->3 porting for the next few years, getting this out of the way now
will allow us to start experimenting again on UX and new cool features! So
I think we should all aim to contribute to the best of our ability and
availabities! It's very much worth it.

Cheers,

On 30 July 2016 at 22:50, Ali Abdallah <aliovx at gmail.com> wrote:

> Hello everyone,
>
> Maybe you have heard already about Gtk 4.0 plan and versioning convention.
> Xfce depends on Gtk, and whatever the gtk guys do to the toolkit will have
> a big impact on xfce. The versioning convention by itself it not a big deal
> (a part that it has no sense at all to change the versioning convention
> that everybody is able to interpret), what worries me more is the rapid
> API/ABI breakages rate. It is true that Xfce port to Gtk3 did not happen
> yet, but I think it is sound to discuss the gtk future plans here, and to
> share opinions.
>
> My overall feeling these days is that I don't trust those guys anymore,
> and I fell myself demotivated to write gtk code that the gtk devs will make
> obsolete in a second, with continuous breakage even between minor
> releases...
>
> What do you think?
>
> Ali
>
> _______________________________________________
> Xfce4-dev mailing list
> Xfce4-dev at xfce.org
> https://mail.xfce.org/mailman/listinfo/xfce4-dev




-- 
Steve Dodier-Lazaro
PhD Student
University College London
Free Software Developer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.xfce.org/pipermail/xfce4-dev/attachments/20160730/bd076689/attachment.html>