[kde-community] Proposal One: KDE (Core) Apps and Suites

[Jos Poortvliet]

On Monday 28 April 2014 11:08:34 Martin Klapetek wrote:
> On Fri, Apr 25, 2014 at 10:12 PM, Jos Poortvliet <
> 
> jospoortvlietstanburdman at gmail.com <jospoortvliet at gmail.com>> wrote:
> > I think the idea of grouping releases ('Sigma'?) is a good one. How about
> > we
> > start there. Let's give Applications more freedom, yet allow them to be
> > part
> > of the 'bunch', yes? Calligra, Amarok, the Extragear apps - they should
> > be
> > part of the KDE Applications. Fold it all in there, with more flexibility
> > thanks to more regular (but non-mandatory) releases. Yes, everybody their
> > own
> > release numbers, no synchronization needed at all. Not every release
> > needs
> > every application, but perhaps for convenience of distro's we provide
> > everything in a tarball- just not with updated version numbers. They can
> > ship
> > KDE Applications 2015.6 (?) and be sure to have all of them, but many of
> > the
> > apps might not be different from those in KDE Applications 2015.2.
> 
> I think the release numbers should be all the same and perhaps even the
> number of the "Sigma" release (also, we should come up with something else
> than "Sigma" before it catches on and stays...like "SC"...unless we want it
> to stay). Otherwise it will be a mess imho - "KDE Applications 2015.6
> contains Dolphin 5.2.1, Calligra 7.8, Amarok 4.6.4, Kontact 5.3.1" -- "KDE
> Applications 2015.12 contains Dolphin 5.2.3, Calligra 8.1, Amarok 4.8.2,
> Kontact 5.4.1"....are those own version numbers really that important? It
> could just as well be "Dolphin 2015.6" or "Amarok 2015.12" (or some other
> numbers), but unified. More coherent, more clear, more simple. The
> downside I see is that the apps' developers would need to commit to this
> new policy, which might hit some resistance.

Look at what we try to do here: message that our applications are separate 
and independent. There is nothing about Ktouch that requires Amarok, and 
Words is just fine without Kanagram. The fact that, on a release engineering 
level, we release them in batches - that is irrelevant for users. They just 
get the one app they want, be it for Windows, Mac, Linux, Android...

Delivering it as a 'suite' with the same version numbers gives the impression 
they do belong together. But they don't - the only thing KDE software has in 
common is the people who make it. Functionally, you can use them anywhere, 
alone or in groups, separate or combined.

Also - most apps wouldn't release with every sigma release, so more than half 
our applications is out of sync most of the time. Having Kontact and Gwenview 
2014.8 and Words and Palapeli 2014.6 and Amarok 2015.2 all being the latest 
version seems more confusing than Kontact 1.8, Gwenview 2.3 etc etc on their 
own. That is what people are used too.

I don't see a strong argument for syncing the release numbers, the confusing 
part doesn't convince me. There's plenty of different version numbers on your 
system atm ;-)

But if anybody knows a good reason to sync, say so please.

> > And then we have Plasma, as it is now - the core desktop (netbook/media
> > center) experience. Kwalletmanager, Systemsettings - they are part of
> > this
> > already, aren't they? That makes sense. The criteria: you really need
> > them
> > to
> > use Plasma Desktop in a reasonable way (eg 95% usecase).
> > 
> > To satisfy the need of distributions (and users) to know what they should
> > have
> > for a basic, functioning, KDE-software based desktop, we define the KDE
> > Essentials. Very bare: Konsole, Kwrite, Dolphin, Ark, Okular, Gwenview,
> > you can imagine. The criteria: EVERY user (well, ~90%) uses these
> > applications, BUT you can swap them with another without everything
> > falling apart.
> I'm a bit skeptic about the metric here. I'd rather maybe define sets of
> goals the user should be able to do with our desktop and then make the list
> from that - "the user must be able to read a PDF; the user must be able to
> view pictures" etc.

Well, sure, but what criteria do you use to decide what criteria should be 
part of the core set? A DTP app won't be part of core following what I 
propose, but "The user must be able to do DTP" seems a perfect fit to the 
"user should be able to do X" list.

In other words, I'd argue that "the user must be able to do X" is a criteria 
that follows out of defining what 95% of the users use. It doesn't work on 
its own.

> > Example: You can't configure things without Systemsettings (gnome
> > systemsettings won't do the trick for you...), you can't save passwords
> > without kwalletmanager, but you can replace Dolphin with Nautilus and
> > Kate
> > is
> > most likely not needed by ~90% of our users. So Systemsettings goes in
> > Plasma,
> > Dolphin in Essentials, Kate in its module in KDE Applications.
> > Accessibility
> > also probably belongs in Essentials, not for 90% of the users needing it
> > but,
> > well, let's call it human decency that accessibility is something we
> > consider
> > essential!
> > 
> > We ship no duplicates in the essentials, and have a best-of-breed policy.
> > Let
> > the release managers decide what goes in, in consensus-style discussion
> > with
> > the application maintainers, that should generally work just fine. The
> > modules
> > - I think they can stay where they make sense for their respective teams
> > (KDE
> > Edu comes to mind) and just go away where they already don't really exist
> > (KDE
> > Admin) or make no sense.
> 
> +1
> 
> Cheers