New Comps Groups

[Jeff Spaleta]

On 11/27/06, Jeremy Katz <katzj at redhat.com> wrote:
>  And what's "better" about searching through a
> long list via browse than search?  I say search is better as you can add
> _other_ qualifiers.

Pre-defined structure using any self-consistent taxonomy is very good
for browsing a large collection when you aren't sure what you want,
but you know are in the mood for something different, something you
don't already know about.

Keyword searching on the other hand requires some amount of personal
understanding of available options as a sorting tool. Don't let
Google's 'Feeling Lucky' button fool you, searching through indexable
material is as much a matter of being an inform searcher than it is
about metaphysical index searching algorithms.

It's exploration versus directed search. Its the difference between
going to a bookstore and choosing to browse the semi-organized stacks
versus walking up to the information desk and asking the bright shiny
ultra-caffiened bookworm wearing the star trek label pin to help you
identify which Babylon 5 books available for purchase include a
Technomage in the story.

Personally I think we should consider different comps for different
user cases and let the users self-determine which comps to use. If its
meant to be an organization tool aimed for specific user roles, lets
provide different views of the package universe that best fit each of
those roles.  Its the Fedora equivalent to the
Choose-Your-Own-Adventure Book.
If you are casual desktop user.. turn to page 10.  If you are a
sysadmin to to page 30. If you are a software developer, go outside
for 5 minutes and get your yearly alotment of sunlight, then turn to
page 21.

-jef"that's it, its time to implement the Library of Congress
Classification System to packages. Hey digital software catelog, I
want package QC878.54"spaleta