Possible GSOC ideas

todd rme toddrme2178 at gmail.com
Sun Feb 24 10:33:50 GMT 2013

Although I will not be able to mentor, I have come up with a few
possible ideas for GSOC fitting with this year's "polish existing
things" theme.  They may or may not be appropriate or even desirable,
so feel free to ignore any or all of them:

1. Polish GHNS handling.  This would involve three parts, which could
be one project, two projects, or three projects depending on the
complexity.  First, it would be to support downloading multiple files
from a single GHNS project.  Many projects, especially on the artwork
side, will ship multiple variants of their work (such as different
color schemes or layouts) in a single project, but KDE can currently
only download one.

The second would be to support downloading files from external
websites using a popup webkit-based browser that would automatically
put all downloads in a particular location and install them.
Currently the user needs to open a separate browser, save the file
somewhere, then track it down in a separate interface from the GHNS
downloader, often requiring installing each file one-at-at-time.

The third would be to handle pulling from multiple sources at the same
time, with the ability for applications to add new sources.  For
example, the window decoration KCM supports aurora, but there are
other window decorations like qtcurve that are also provided through
GHNS.  This would allow them to be added to a single list so users
don't need to care what backend provides the decoration.

2. Improve mimetype association handling.  Currently KDE has two KCM
modules in two categories for handling mimetype associations: default
applications and file associations.  Further, there are limitations in
the current file association dialog: no drag-and-drop support, no
ability to move multiple applications at the same time, no ability to
copy associations from one mimetype to another, no ability to set file
associations for a mimetype category, etc.  This adds up to making
changing file associations an extremely long and complicated affair.
Most users aren't going to care about dozens of raw file formats, they
just want all their images to open in the right program.  The purpose
of this project would be an overall usability revamp of these UI's,
combining them and making it easier to make general changes to a large
number of mimetypes at once.

More information about the kde-core-devel mailing list