Re: Moving away from app-centric mimetypes (e.g. kword)

Howdy,

> This is why, ultimately, no metadata system without
> kernel support can 
> ever be complete.

This is true, but if you consider how much work it
would take to get a kernel-level change like  a
filesystem to be widespread enough that KDE could
depend on it, the metadata feature will be dead before
it ever arrives.

 
> This is not to suggest that partial solutions aren't
> helpful. 

This is the approach I'd vote for.  Let's get it done
now, rather than take a long time to get it done
right.  Modern machines are overpowered, and it sure
seems like they'll continue to be.  If there were a
local search engine that integrate each app's metadata
to make some good indices, you could get very powerful
search results and virtual directories into Konqi very
quickly.  And given the robustness of the Kparts
architecture, KDE is well poised for this.  Plus
getting some kind of metadata searching incoporated
into a desktop and usable can help serve as a roadmap
for lower-level approaches.

To give an example:
	Friedrich mentioned marking a Kword doc a letter.  I
think that's an excellent refinement on Clippy.  If
you're writing a letter, have KWord pop up a little
window (inobtrusively, like the way the light bulb
pops up in OpenOffice) that with a drop-down box to
classifies the document as a letter.  Then enable the
find feature in Konqueror to recognize the
document-type metadata, and filter on it.  At the very
least, this would mean a column in the search results
called "Document Type"

It would not be the most elegant or complete solution,
but it would be working, and it would get people using
it and thinking about it more.

My $.02222222

Cheers,

Eric

__________________________________________________
Do You Yahoo!?
LAUNCH - Your Yahoo! Music Experience
http://launch.yahoo.com