Hello all,

this is my first attempt at describing UI areas of KDE which have room for 
improvement. I'll try to provide links to related research.  C&C is welcome 
as always.

Well, after reading several papers on e-mail clients and document interaction, 
a general flaw in the way data is currently presented in desktop environments 
became somewhat apparent to me: Users only see abstract representations of 
the data they have stored. This appears to be one of the prime reasons of 
people still making heavy use of paper [1] and to some extent it also lets 
users become overwhelmed by data [2].

For those of you who do not have access to the ACM library, I'll try to 
explain the main points of concern.
Basically [1] (as well as [3]) states that paper in the real world can be 
regrouped and rearranged in order to support people with prioritizing tasks 
and reminding them of important items. A good example is paper on a typical 
office desk: Sheets that are close to the user have higher priority than 
those further away. Sheets at the bottom of a stack have lower importance 
than those on top. In addition, just by looking at a pile of documents, 
people can see its contents (through paper size, color, thickness, labels, 
etc.) and will be able to tell what the documents are about and what tasks 
they belong to. Further on, the fact that documents are in the same stack can 
imply that documents are related and belong to the same task.

Going to the computer, most of these rather simple real-world actions are not 
possible at all or only rudimentarily supported. Ever wonder what file 
"presentation-ces-03-fixed.odp" contains? Well, you need to open up a program 
and wait for a few seconds before it is even possible to know what the 
content of that document could be. Now think about transferring one of those  
20 sheet stacks into the electronic world and you will agree that getting an 
overview of a set of documents or picking the right one will not be easy at 
all.

Similarly, in the context of e-mail, [2] argues that all users typically see 
of their e-mails is a long list of sender/subject/date lines in a table. If a 
user wants to read an e-mail which is contained in a folder with a large 
number of other messages, it will be difficult to pick the correct e-mail as 
there are only few visual cues on message content.

Now, some people might counter that there already are file previews available 
and that my whole point is obsolete. Well, ever thought about what you 
currently see when your first page is blank? What if the most important page 
is in the middle of the document? Would research be conducted even in 2005 if 
this problem didn't exist?

With that being said, what would all of this mean for KDE (4)?
In my opinion, it is absolutely essential that users are able to instantly see 
file previews. And this does not only apply to the classic case of text 
documents or images: Previews should be available for all types of files. 
Have OGG files in your music collection? Why not show the album cover instead 
of a loudspeaker symbol! What about more exotic files like a UML diagram? It 
is certainly important to someone in a software project. Ergo it needs a 
preview. You like watching short films downloaded off the net? Was 
"final_med_snd.ogm" that funny clip with the dancing bear or was it 
"20050524.talktalk.shapes.ogm"? You get the point.

Looking at it from a more technical side, I assume that KDE currently relies 
on external libraries to create previews, e.g. for images (comments please!). 
If a large number of filetypes are to be supported, this would most likely 
become a massive dependency problem and therefore does not seem to make much 
sense.
Well, this is where I think Tenor would come in handy. Instead of having KDE 
create a preview for files, why not offload this task to applications, which 
in turn register their previews in Tenor? In other words, who would know 
better how to represent important aspects of a file than the application 
which created it?! Think about it. For example, if Amarok already downloads 
cover art for music, it could as well store it in Tenor which would make it 
available to all of KDE. If you edit a document in KOffice or use Umbrella 
for UML, why not register a preview? It would not even be necessary to change 
the file format in order to support previews.

To summarize, I hope that it became apparent that ubiquitous file previews 
should enable users to better group and manage their documents and therefore 
improve their workflow.

Now even if this was rather long, hopefully some of you can comment on the 
above points.

Cheers.

[1] Documents at Hand: Learning from Paper to Improve Digital Technologies
http://doi.acm.org/10.1145/1054972.1054990
[2] Designing remail: reinventing the email client through innovation and 
integration
http://doi.acm.org/10.1145/985921.985944
[3] How do people organize their desks?: Implications for the design of office 
information systems
http://doi.acm.org/10.1145/357423.357430

-- 
Markus Weiland
[EMAIL PROTECTED]
http://www.3danim.de
_______________________________________________
Appeal mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/appeal

Reply via email to