let's begin with two

[Fania Bremmer]

Am 12.10.2011 17:56, schrieb Marco Martin:
> Hi all,
>
> so far active one seems quite a success and that's great :)
>
> so, to not lose time and think forward what we want to do next, let's plan a
> bit for two:
>
> * first of all finally, recommendations, yay ;)
>
> other than that, things revolve around a concept:
> the idea was to have a release in december, and that's mean it's going to be a
> relatively minor update and that's good since at this pointwhat's needed is
> mostly ironing out the user experience.
> a target, was to provide use cases for an user "workflow" starting from a
> simple case (one idea was image management) that will be the basis of an
> uniform document browsing component
>
> to achieve that, to me the priorities are 2 not really visible foundations:
>
> * qtcomponents: using the old qgraphicswidgets (use of which i limited as much
> as possible) gave several problems in the ui since there still are the old
> layouting problems of qgraphicswidgets, they don't support anchors and all
> have mouse over events.
> having standard-ish components will speed up things a lot, i hope that with
> the meeting of tomorrow we can have a clear path forward (i would like to see
> a ban on import org.kde.plasma.graphicswidgets in pa2 ;)
>
> * nepomuk: besides the startup issues noted earlier today (that are a bit
> upstream to us) the way queries are encapsulated will probably need to change:
> the metadata dataengine was fairly convenient to use for simple things, but we
> can't do more complex queries without adding special cases in the dataengine
> c++ and most important is very inefficient, a more heavy duty (and lazy
> popuplating) model is needed for having good performance (and less memory
> waste). That's the prerequisite for a slightly smarter document explorer.
>   a separate thread will be start for this ;)
>
> besides that, to support the idea of the image organizing workflow an ui would
> be needed for that. initially should be a simple ui that i (hope) could be
> shared between the add resource interface and the images application,
> basically an ui that lets do filters on resources by type, tag, date, rating
> (and keep it very simple in the beginning;)
>
> other minor stuff that still are needed (kinda missing from pa1) is a
> beginning of configuration interfaces in qml, with basically a porting of the
> logic of the clock/timezone kcm to a bunch of qobjects that can be used from
> qml
>
> other ideas? (well, remembering that december is 2 months away :p)

For me, from a designers perspective, there is the big goal of polishing 
the current User Interface and its Interaction. This has been not 
accomplished yet. Just a few things to mention about that:
- A big goal should be becoming more and more touchfriendly, touch 
centered, offering a real flow while working with the product (for 
example the add items and cancel buttons in the add resource dialog are 
too desktop-like). We talked a bit about this topic in our last workshop 
- we should now concretize the steps.
- Switching: the flow of working between activity and apps, switching 
and opening is not mature yet - there was also a mail about "a lot of 
elbow" movement on the ML as well
- Shaping and polishing the User Interface and its experience; there are 
little issues like the activity switcher wheel that looks bad with only 
3 activity thumbnails, because they are too far away from each other; or 
sometimes a wording that is not optimal. For this point I would like to 
invest time in heavy user testing. With PA1 we have a version that users 
can test, try, sqeeze, adjust.... and we should observe them. I am 
100%sure, that there are still a lot of traps in the UI, that should be 
fixed
- ShareLikeConnect: the current context menu is not final in my opinion. 
It doesnt work consistently for all types of resources (there is "remove 
bookmark" but no "remove/unlink document") and is not intuitively 
understandable.

So, besides the recommendation feature I would rather use the time to 
polish UI wise. If there is time left we can try to think about a smart 
way how to implement the image management workflow and adapt its 
patterns to the rest of the system. But thats already a big tasks for 5 
weeks only ;)

Cheers,
Fania