Introducing Homerun

[Aaron J. Seigo]

On Tuesday, November 13, 2012 17:05:02 Alex Fiestas wrote:
> So a 2level tree with the amount of branches and leaves per branches we have
> in kickoff is scientifically proved to be better than a grid with the same
> amount of leaves?

this question is malformed.

it depends on the number of leaves[1]; the accuracy of the branch labels [2]; 
how well applications can be sorted relative to the values used to find them by 
the user[3]; if there is partitioning within the grid (which is another form 
of branches and leaves) and how that partitioning is chosen and displayed; 
etc. a ton of research was done in the 90s and early 2000s around these 
issues, though they mostly focussed on website navigation.

as for typical usage of launchers ... 

on Plasma Active, there is the grid of applications approach in the launcher. 
i find that as soon as i have more than a couple pages, this rapidly becomes 
very craptastic to use. this mirrors both personal experience and observation 
of people using similar launchers on smartphones.

with kickoff, the categories are limited by the app menu organization. this 
organization was accomlished with, at least to my knowledge, no usability 
process such as card sorting exercises applied to it and the results reflect 
that -> people seem to struggle with finding things in that hierarchy.

once aware of search possibilities, people seem to quickly find applications 
based on search criterion, but this is slower than selecting from a focused 
list.

one big question, which we shockingly do not have an answer for[4], is what 
people use these launcher listings for in actual, common, real-world usage.  
once that is understood clearly, a strategy for serving that need can be 
derived. they can even be tested, once we know what we need to measure for.

if may turn out, for instance, that a simple search field that lists the most 
commonly launched application beneath it is all anyone needs or wants. or it 
might be horribly inadequate, even if it might look really sexy.

[1] human list scanning has limits; on a carefully managed system there might 
only be a dozen or so application in the app menu and that is likely to be 
very easy to manage in a grid listing. if, however, there are several dozen 
applications on the average system, a grid can become inneficient

[2] is "Internet" really a meaningful category for today's applications and 
users?

[3] a list of countries sorted alphabetically is just fine for finding a country 
quickly (ignoring possible issues related to localization of place names :); a 
list of countries sorted by lattitude is probably not so easy to use, however.

[4] personally, launchers don't interest me much. for other topics which i do 
find more interesting, i have done more "field" experiments and reading on. i 
wish that those who find launchers interesting would similarly apply themselves 
to the topic and gather real data so as to uncover the questions that need 
asking and then work on evidence based answers. opinion based design is 
probably why the launchers we have are serviceable but not great.

-- 
Aaron J. Seigo
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20121114/52112ddc/attachment.sig>