Re: Moduleset review



Michael Catanzaro <mcatanzaro gnome org> wrote:
On Mon, 2016-01-11 at 18:37 +0000, Javier Jardón wrote:
This has been merged now

Thanks for working on this, Javier; I think our modulesets are much
better now. I posted a follow-up patch for the metamodules here:

And thanks to you for doing a much more thorough job than I did, Michael!

Having well organised modulesets that reflect the current state of the
project is a good thing in itself. That said, when I originally looked
at this, I had wider goals, and it is possibly worthwhile to revisit
what these are, and what other steps might be required to achieve
them.

 * Help the GNOME project focus, and help contributors to understand
the layout of the project, by providing a common organisation which
reflects our priorities and direction. (This implies that contributors
will be exposed to the moduleset definitions, through the organisation
of git.gnome.org, Bugzilla, etc. This goal won't be reached if no one
sees the modulesets.
 * Make decisions around modules (and specifically apps) more
transparent, and provide some kind of accountability for those changes
through the release team. So if we were to switch from one core app to
another, it would be clear how and why that decision had been made.
 * Promote exposure of the experience as designed, and enable testing
of this experience. So, for example, testers could download live
images featuring the core default experience, as we want it to be.
 * Communicate to distros what our intended default install looks like.

Perhaps there are other goals worth considering here?

Apps in Workstation, but not gnome-core:

This list looks good to me, with the following questions/comments:

bijiben

While I do believe that there needs to be a notes app in the core
experience, I'm concerned that bijiben isn't of sufficient quality
right now.

evolution

Not sure this should be in core to be honest: it's a heavy enterprise
thing, which might makes sense in some contexts more than others.

gnome-boxes

I like the idea of having Boxes in the default install, although maybe
that has size implications for install media? I think this could go
either way, to be honest.

vinagre

Isn't this superceded by Boxes?

Apps in gnome-core, but not Workstation:

empathy
epiphany
gnome-logs

I think Empathy should be removed from core unless a new maintainer
surfaces. Epiphany is under active development. We should probably add
Logs downstream.

Makes sense.

Now, the design team has a page with designs for GNOME core apps
here:

https://wiki.gnome.org/Design/Apps

It's worth bearing in mind that this list is something of an
idealisation: it's future orientated rather than rooted in the here
and now.

It is much more expansive than what we have currently. Ignoring apps
that don't exist yet, it includes the following which we do not have in
core:

gnome-documents (and gnome-books)
gnome-photos
gnome-music
gnome-clocks
gnome-maps
gnome-weather
gnome-calendar
bijiben
gnome-dictionary

It also includes Mail and Chat, but the designs are so far from Evolution and Empathy that it's quite up 
for debate whether or not we should include those apps until replacements exist. I'm curious to know your 
opinions on this (and Allan's in particular).

I agree that Empathy is a long way from what we want, and shouldn't be included.

My view is that Evolution - as designed - is a specialised app for
people in corporate environments, who have a need for a complete
Outlook replacement. It obviously overlaps with Mail, Calendar,
Contacts and Notes. I wouldn't recommend installing it by default.

The design team list omits:

evince
eog
gnome-font-viewer

I would like to implement the proposal from the design team. Some specific thoughts:

 * It's not lost on me that we just removed Dictionary from core. I thought this would be uncontroversial; 
Allan, was this a mistake? Should we put it back?

Having a dictionary is nice, particularly if it has a search provider.
But it's also not essential, so if our dictionary isn't up to scratch
it wouldn't be the end of the world not to have it in the default
install.

 * I will consult Richard to see if Color Manager is still needed by gnome-control-center. If so, we need 
to move it to core. It has a NotShowIn=GNOME desktop file, so this should be uncontroversial, but it is 
showing up in GNOME Software as an app, which needs to be fixed.

My understanding is that it's a control center requirement, indeed.

 * I will consult with the gedit developers regarding de-branding the
desktop file, as I think gedit should be in core. I do notice that
gedit is conspicuously missing from the design page, but I think it is
too important to omit.

gedit's role has always been rather fuzzy. It's a basic editor, but it
can also be expanded into something much more complex. It's treated as
part of the default experience, and yet it has its own identity rather
than being the generic "Text Editor".

I would love us to move in the direction of only having unbranded apps
in the default experience though. It would make the difference between
the stock apps and the 3rd party ones you install much clearer, and it
would make the default install feel much cleaner and well-organised.

It might be worth asking what our requirements are for text files out
of the box. Do we require a text editor to be installed by default, or
do we just require the ability to view a text file? Do we want
something super simple, or should it be more capable?

 * Lastly, Cheese. Cheese is currently in core; it's needed by gnome-initial-setup, gnome-contacts, and 
gnome-control-center. On the design page it is explicitly marked as non-core, but if that is the case then 
we would need to remove the webcam integration from these three apps. Allan, was this an oversight? Should 
rename this to Webcam or Photo Booth and keep it in core? If so, I will talk to David.

I think it's fine keeping it in core - it's nice to have something
that's a bit fun in the default install.

Thanks again!

Allan


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]