Re: [Usability] GNOME UI principle: All applications should save internal state?



On Mon, 2004-04-05 at 02:58, Ryan McDougall wrote:
> On Mon, 2004-04-05 at 10:31 +0900, janne wrote:
> > mån 2004-04-05 klockan 09.33 skrev Ryan McDougall:
> > > Do you agree that all applications should save their state such that
> > > they are no different from the time they were close to the time they
> > > were reopened? If you agree should this be part of the HIG or similar UI
> > > policy? What are the positive or negatives of such a policy, were it
> > > implemented? Should there be a unified API/framework for apps to reuse
> > > so they can store their state?
> > > 
> > > As an example of an app that goes a long way to save state in a real
> > > spatial or sessional way, see Eclipse's workspaces. Epiphany and Pan
> > > also save their current workload state in the event of a crash, but
> > > should they *always* save their state (ie: save the current web pages
> > > seen, or items queued for download), even if they are terminated
> > > normally?
> > > 
> > > I think they should, and doing so provides a number of UI benefits.
> > > Consistent state meshes *especially* well with a spatial design, since
> > > the "real world" always has a consistent state. When I leave my papers
> > > spread out at work, they are sitting exactly where I left them the next
> > > day. I also think there should be a GNOME-wide policy and method to
> > > allow apps to save their state data is a consistent and easily
> > > manageable way.
> > 
> > Interesting idea.
> > 
> > As you point out, though, in many cases it is pretty fuzzy what "state"
> > really should mean. Consider a CPU activity monitor applet, for example;
> > when you restart, should it remember and display the graph from before
> > the reboot? Or if a particular set of network interfaces where active
> > before reboot, should they (through the network device control
> > application) by default attempt to reconnect using the same parameters?
> > 
> 
> You have a point. I think that the definition of what "state" is should
> be defined on a app-by-app basis. Although its obvious that apps with
> trivial internal state, or which simply display the state of other apps,
> should refrain from wasting resources by saving state unimportant.
> 
> As for apps that manage resources they do not directly control or have
> implications beyong GNOME, such as system state, should not save state
> since this represents obvious system management headaches (as you
> alluded to). However as you know, we do have a way to save system state
> usually called init-scripts and config files. ;)

i also think in many cases it makes sense (for example mail clients)...
but not everywhere . and it's not just the case of system monitoring
tools.

for example it would irritate me VERY much if epiphany would restore the
last visited page(s) (tabbed browsing) everytime i start it.
there is a reason why i have closed epiphany. it is because i do not
want to see the webpage anymore.

maybe an easier way to browse the history, or to auto-bookmark the pages
which are open when you close epiphany would be the solution.

and the same also for gimp. and imho for many other apps.

gabor




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]