mime apps specification

[Mark Edgington]

On Sat, Apr 5, 2014 at 6:11 PM, Jerome Leclanche <adys.wh at gmail.com> wrote:
> For what it's worth, I agree. However the use case of browsers asking
> to be default has to be taken into account, and we are very far from
> being able to offer them an api to *properly* become so.

Hi Jerome,

I recall someone mentioning on this list not to long ago the idea of
having desktop-environment-specific applications which are able to run
with slightly higher privileges than "ordinary user applications".  I
don't know what all would have to go into making something like this a
reality, but to me it makes a lot of sense.

For example, if such a separation of privileges were possible, then a
user could, via standard desktop-environment configuration tools,
choose among a set of applications registered for a particular use,
deciding which one(s) should be default (or that the user should be
prompted to choose between when opening some kind of
resource/document).  But, a mere web-browser or other application
would not have the necessary privileges to make such a change -- the
only thing they would be allowed to do is to launch the
desktop-environment "default apps" tool, and give the user
instructions on what they should do if they wish to make "browser X"
the default.  Of course "ordinary apps" would be allowed to register
themselves as handlers for different MIME types, but they would not
have the authority / permissions to make themselves default -- that
decision would be something a user would explicitly need to choose via
the desktop-environment's config tools.

Nonetheless, even if the above (separate permissions) scenario were to
exist, it would still be important to have some kind of standard API
for choosing a single default and/or preferred-set of apps to handle
different MIME-types, so that this information could be shared between
different desktop-environments, and so that desktop-environment
developers would have a standard by which to develop their
"default-app selection" tools.

What do you think about this?  How feasible would something like this
be to implement?  If it could be implemented, it would certainly
guarantee a lot more user-control over preferred apps, etc.

Regards,

Mark