On Thu, Jul 22, 2010 at 09:32:53PM +0200, Andreas Tille wrote:
On Thu, Jul 22, 2010 at 07:01:07PM +0200, Jonas Smedegaard wrote:Above Samuel says that in a *lot* of cases multiple screen readers do not make sense. But does not go as far as claiming that in *all* cases is that true.I therefore do not like conflicts here.Perhaps a mechanism like the XDM/GDM/KDM/NODM one would make sense here?That is, allow multiple packages to be installed concurrently, but have them coordinate to only by default enable one of them. That way in esoteric cases[1] where it might make sense to use multiple screen readers, it is possible to hack the startup script (and evolutionary convince package maintainers to instead improve that script) instead of being forced to hack the binary packages.When I wrote my mail about using a debconf question in the metapackage I've thight about something like putting a configurable variable into/etc/defaults/screenreader which contains lines like SCREENREADER_A=yes SCREENREADER_B=no ... and use something like test -f /etc/default/screenreader && . /etc/default/screenreader if [ "$SCREENREADER_A" = "yes" ] ; then startupcode fi (or something a bit more sophisticated, but you see the principle).That should be quite easy to implement because all relevant screenreader packages are under control of the Debian Accessibility team and this change could be implemented quite easily.
Thanks for the details.I do not doubt that team to be able to handle this issue. But since we are discussing it here, I focus on *principles* of packaging, and consider that particular set of packages an example only.
Also beware that configuration files are owned by a specific package and are not allowed to be edited directly by other packages - no matter if maintained by same team!
- Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private
Attachment:
signature.asc
Description: Digital signature