[Portland] Summarize current plan?

[Lubos Lunak]

On Tuesday 14 March 2006 15:11, Jeremy White wrote:
> First, let me apologize; I snapped, and there was a lot
> more froth and spittle in that last email than was needed.
>
> >  Short summary: The menu spec doesn't work on Mandriva, Kubuntu has a
> > bug, some desktops doesn't support the spec and the spec is imprecise in
> > some details, so in practice it doesn't work in 100% cases.
>
> Yes - the menu-spec is good; I want more
> areas of the Linux desktop to work as well as the menu-spec does.
>
> My key point is that even in areas that people think are good
> and/or solved, we have a 'last mile' gap.
>
> >  Now, what exactly would you like to happen here?
>
> That's a fair point; my email was long on venting and
> short on constructive comments.
>
> My worry was that you will have persuaded others that no
> work is needed, and I think that is very far from the truth.

 No no no ... that'd mean risking I'd have to do all the work once somebody 
finds out there's still lot to do :). And there is.

 Actually, extending the Wiki to include more info about what's needed would 
be probably helpful. I've ignored installing menu items because (besides the 
fact that I consider that to be outside of DAPI's scope) I really consider 
that one solved. There's a spec that should work and there's not much to do 
here about somebody not implementing it.

> I think work on documentation, on aggregating existing solutions,
> and on 'blessing' simple solutions is critical and needs to be done.
>
> >  As for the last item, the spec needing some
> > clarifications/fixes/whatever, if you really could have spent man months
> > on it, why didn't you instead try to spend noticeably less time on fixing
> > the spec and doing what'd be necessary? It's not like you're not allowed
> > to.
>
> Um...I'm not sure what you're suggesting.  You seem to be saying
> "don't bother trying to improve the specs, just do whatever you have to
> to make it work."  That is what we've done; we've adapted to all of
> the sick and wierd ways that Linux distros like to arrange their menus.

 No, I meant it the other way around, to improve the specs instead of working 
them around. If you really spent such awful amount of time on this, you could 
have instead joined the XDG list and said "look, this part in the spec is not 
right, we suggest fixing it like this" - you'd be simply another contributor. 
Even if that would mean the extreme case of completely driving the process of 
updating the spec and fixing the KDE/GNOME implementations you still wouldn't 
possibly spend man-months on that.

 Of course, reality would probably dictate supporting at least some 
installations that wouldn't support the latest fixed spec (yet), but you 
could shrug off at least some of the problems with making your release notes 
contain something like "X and Y don't support this spec, Z intentionally 
breaks it and A has this small bug, sorry, complain to them".

> >  Good news for you, OpenUrl is the first call in dapi/doc/API.txt . And
> > your last sentence describes the motivation - there's no spec and it
> > depends on the running desktop, so there can be no simple shared
> > implementation. You'll just call OpenUrl and the active backend daemon
> > will take care of it.
>
> Aha!  Now we come to an interesting point of debate.  See, I think that
> the existing methods to open a URL are 'good enough'; I would just like
> to standardize around the existing methods.  I find Kevin's
> implementation to be the ideal.
>
> But I'm not sure of that; I can imagine that more sophisticated
> meanings of 'Open a URL' will require an alternate implementation.
> So I agree that competing approaches is a good idea.  And I think it's
> vital that even the simple solutions be framed in the awareness of
> the larger problems, so that they can eventually scale to solve
> the hard problems.

 Good.

> I'm just obsessed with seeing that the simple problems not be neglected,
> and that we help ISVs make it over that last mile.

 Sure, that just needs somebody to do it. I can't do everything :). And as you 
can see from Kevin's posts there are people like KDE devs here to help with 
the desktop side of things.

-- 
Lubos Lunak
KDE developer
---------------------------------------------------------------------
SuSE CR, s.r.o.  e-mail: l.lunak at suse.cz , l.lunak at kde.org
Drahobejlova 27  tel: +420 2 9654 2373
190 00 Praha 9   fax: +420 2 9654 2374
Czech Republic   http://www.suse.cz/