Re: a list!!?!

[Brian Fleeger <brianfleeger yahoo com>]

From: William Jon McCann <william jon mccann gmail com>
To: Bryan Clark clarkbw gnome org
>One idea that I had that I quickly rejected, as requiring a mental model
>that is too complex, is for workspaces to be activity bundles.  Others
>seemed to have liked this idea but I remain deeply sceptical. 

 

Hi,

 

I am just a normal computer user, but I eventually came to understand other peoples conversations about "Activities" to mean app and doc bundling too.  Although it took me a while to get my brain around it, when I explained it to others (coworkers and classmates) as a way to quickly jump into complex tasks (like research reports, doing taxes, etc) that involve lots of documents and are repetitive over long periods of time, most people responded very positively to the idea.  In fact, many people (well... three actually) who are in no way technically inclined told me they would love a feature like that.  Whatever the designers decide to do, most design decisions have pluses and minuses.  My personal feeling on activity bundling is that it all depends on how it is introduced to users, and how easy is it to do.  Becuase the feature would be completely new, it would have to be easily discoverable and totally intuitive, with major graphical clues guiding people until they get the hang of it.  I am not saying that I know how to best present the feature, but I think it does have merit and is at least worth further investigating.  Unless the shell designers think it is not worth the trouble, that is.

 

On the other hand, from a hardware perspective, launching an activity bundle really would be taxing on system resources.  Even now, opening a singling document in open office takes a long time on my old 2004 notebook.  Opening a raft of applications, documents and web browsers sumultaneously might cause some systems to hang for quite some time.

 

>Another goal was to try to reduce the reliance on workspaces by defaulting to
>only one and allow adding them as needed.  I think that trying to
>design around the one workspace model is important for a few reasons:
>a) most non-geeks will probably only use one b) it greatly simplifies
>the navigational complexity of the system c) it forces us to be more
>creative in solving window / task management issues.  I think a lot
>more focus has been put on workspace management in the shell
>implementation than some of us in the hackfest would have liked (or
>maybe it is just me).

I don't know if non-geeks would only use just one desktop if they were put in front of a gnome-shell style interface.  FWIW, I believe user behavior is based largely on the visual presentation and spatial framework with which they are presented, kind of like those Japanese watermelons that are grown in boxes and so turn out cube-shaped.  Many people only use one desktop workspace now becuase it is all they know and they've never had a choice to do it any other way.  I remember in the early days of tabbed browsing, a lot of people (including me, until I'd used them) thought tabs were superfluous.  In fact, the early implementations by Netscape really were pretty non-intuitive.  But now, because of better UI design like in Firefox and Chrome, most people loves tabs, even grandmothers.  I believe the giant Plus sign in the Gnome-shell will encourage people to treat desktops more like disposable tabs, and that if done well it may not be at all confusing to people.  My feeling is, the benefit lies in users being able to spatially parse out their workflows more freely.  

 

Rui Tiago wrote:

>Workspaces are particularly confusing IME when you have more
>that one physical display on a xinerama(ish) setup which seems to be
>getting quite common, at least on my anecdotal list of friends and at
>work (where we use Windows).

 

On the other hand, I completely agree with Rui that multiple workspaces is really hard to get to work well on multi-monitor setups.  My wife is of the opinion it would be great if each desktop workspace could be assigned its own screen, but that has other complications too.  Should solving this issue be a point on the list of Open Design Questions?

 

-Brian