Re: Recently Used Files Proposal

[Mike Emmel <mike emmel gmail com>]

On 6/6/05, Aschwin van der Woude <aschwin van der woude movial fi> wrote:
> On Sat, 2005-06-04 at 08:43 +0000, DANIELLLANO wrote:
> > Emmanuele Bassi wrote:
> > > What if you want to open a document you edited some time ago (but not
> > > too long), and it went out of the menu?
> >
> > Then it's not a Recent Document.
> 
> I have always found the "recent files/documents" list kinda useless. I
> use my computer for various tasks, some work related some private.
> 
> During work I might use a small set of files, or perhaps just even work
> on a single document during the day. In the evening I might pick up my
> digital photography hobby and review a bunch of photographs and perhaps
> even edit them. (Nautilus, eog, gimp)
> 
> The next morning I start working again and want to continue to work on
> the file I was editing the day before. Unfortunately it already fell off
> the "recent files" list, and I will have to dig out that file myself
> again. Which is pretty deeply nested in my cases, as I have lots of
> files to keep around.
> 
> Perhaps I am alone in such situation, but at least for me the "recent
> files" list is not very useful in most cases.
> Even if this is more common than my own personal experience, I am not
> sure what a solution is to this use-case. Perhaps grouping recent-files
> per directory or such...  I dunno... I am just pointing out my own
> observations.
> 
> -Aschwin
> 

A solution to this thats not too invasive and may be possible is to
record the application that opened a file and group by application
first. Next is to notice that multiple applications are being used on
the same file and create the meta group. In general you want to have a
design that allow you to build up a ontology. Too do this I think  its
important for the framework to be able to be extended to add new time
dependent tasks to gather new information and too allow groupings to
be created and saved.  Consider the proposal of a most rececently used
file menu to be one view into this ontology back end.

Another generic use case is lets say your also using documentation or
other urls as you work on something you would like the urls to also be
associated with the current files as current urls.  It seems to me
that one important point is you need to define a top level task to get
the system started. This can be done via some hueristics and would
probably work okay but it makes sense to allow the user to define the
current session and allow them to pick sessions as they change what
there working on.  You could also create anti-sessions that do the
reverse and ensure that no trace of the sessions stays on the
computer.

A use case I'd be intrested in is one where I tend to work via ssh on
a remote server with a web browser up for docs. I'd like to save that
"session" its directories etc.  This would require that the ssh client
report session info back from a remote machine too my session manager.

I'm not saying that you need to do all this stuff but I really think
that the proposal is just one aspect of a bigger problem thats useful
to solve and if done correctly it can provide the framework to build
out  cooler stuff :)

Taking the time to use DBus here in the right places might make a lot
of sense for example.


> _______________________________________________
> gtk-devel-list mailing list
> gtk-devel-list gnome org
> http://mail.gnome.org/mailman/listinfo/gtk-devel-list
>