[kde-community] Why were there no talks about Ubuntu Mobile at Akademy?

Michael Zanetti mzanetti at kde.org
Fri Aug 23 07:31:14 UTC 2013


On Friday 23 August 2013 01:10:09 Aaron J. Seigo wrote:
> On Wednesday, August 21, 2013 09:14:58 Michael Zanetti wrote:
> > I for one work on Unity8, which just works
> > and looks so different in every way than plasma does. We don't need
> > Plasmoid containers, you don't need search scopes.
> 
> How do you know we don’t need search scopes?
> 
> The functional similarity between “search scopes” and Plasma::AbstractRunner
> is pretty stunning, actually.

Ok. This is news to me. Feel free to evaluate if you can build on top of the 
100 scopes project. AFAIK it's all GPLv3.

> 
> > - Once there is something which might make sense to be shared, it requires
> > the exact people working on it having interest in collaborating. Which
> 
> Yes .. and no.
> 
> If we leave it up to each and every individual to make a decision on their
> own in a vacuum, then you are correct.
> 
> If we make thoughtful community/corporate encompassing goals then it is not
> left up to each individual: there will be a point of guidance that people
> can use to guide their decisions.

I agree that this would be desirable solution. But I think doing the "big 
thing" in this regard would for sure break all deadlines and time plans. 
Here's again the totally not matching time to marked requirements from KDE and 
Canonical which make things hard to even get started.

> 
> > means, the responsive KDE person needs to accept that a certain API needs
> > to change for requirements NOT needed by KDE
> 
> This is a non-sequitor. If someone is using KDE libraries, then their
> requirements become part of what is needed by KDE. KDE is not a castle on a
> hill with a moat around it; it is an open marketplace where the edges are
> defined by participation.
> 
> > and the responsive person in
> > Canonical needs to have interest in pulling in something that most likely
> > can do way more than Ubuntu needs at this stage,
> 
> There is that word again: “likely” :) Every time I see that I think “.. it
> means we haven’t done the necessary homework to know for sure and so people
> are making assumptions”

You have a point here... noted.

> 
> There are two important points to consider:
> 
> a) Probably most libraries used do more than Ubuntu needs. Is every glibc
> call used by Ubuntu software written by Canonical? Is every Qt API used?
> (QWidget, e.g.)

No. But the difference here is that those examples are quite stable or have a 
huge community by themselves while KF5 is still in early stages and before 
Blue Systems announcing its support for it, it was mostly 3 (or so) guys doing 
KF5 in their spare time. That does make a difference in terms of needed efforts 
to incorporate something.

> 
> b) If a library in KF5 is poorly modularized, resulting in something that
> does so much more than it ought to that we should reconsider the division
> lines within it, we need to know *now*. So if there is a KF5 library that
> would make sense being used in Unity (e.g.) but it does “way too much”, we
> can fix that. But we need to know.
> 
> > needed. It is not possible for me or Albert to go to some API guys and
> > tell
> > them: You have to share code with KDE. This needs to happen from inside
> > the
> > team. The person doing the work must drive it.
> 
> There must be leadership that can set engineering mandates?

There is... But exactly those are the ones that need to worry about deadlines 
and additional risks. So unless the developer says "Yes, I can do this" on its 
own, they won't force anyone to go through the additional efforts I mentioned 
before.

> 
> > Now, coming from the Gnome/Gtk area, Canonical's people mostly are aware
> > what code could be shared with Gnome, but not many of them have a clue
> > what
> > KDE frameworks actually is.
> 
> I’d just echo Thiago’s questions here, as they are truly insightful and key

I don't know to be honest. Haven't done the homework to go through KF5 and 
check what it offers myself yet.

> ..
> > Same the other way round. I'm quite sure very
> > few here know how the Ubuntu's architecture is built up.
> 
> Not many, I’m sure, but they exist. We have the Kubuntu folks and then there
> are crazy people like me who do look at what ends up in Canonical’s public
> repos and who have even done things like port apps written for Ubuntu Touch
> to other QML component sets. So the ignorance can be dissipated through
> effort!

Exactly, we're ending up again at *additional* effort. Which is really 
something that does not go well with tight deadlines.

>
> > Then again, we actually do share and reuse some code. Take all the lightdm
> > stuff for example, the dbusmenu stuff and many more libs which in history
> > have flown into both directions already.
> 
> The status notifier stuff was a success, yes. I’d like to see that built
> upon so we have many more like that.

Fully agree.

That said. I try to make a good example by doing the Marble port for Ubuntu 
Touch in marble's git repo, by writing my phone apps in a way they can be used 
as a Plasmoid and a phone app and promoting this way of working on the ubuntu-
phone mailing list. Still, I can't force other people to do the same and I am 
not Canonical management. Also keep that in mind when reading my mails. It's 
my personal views on things trying to explain the situation as it currently 
is. I'd be happy to be proven wrong and see more of those good examples 
upcoming.

Br,
Michael



More information about the kde-community mailing list