Re: Any news about Blends in tasks selection (Was: Debian Installer Jessie Beta 2 release)
On Tue, Oct 14, 2014 at 10:14:53AM +0200, Cyril Brulebois wrote:
> > In other words: I perfectly know the fact that Blends are widely
> > ignored even amongst Debian developers and that's not about you / the
> > debian-boot team - perhaps my "running around and tell people" is just
> > not the right way to convince people. At least I can tell that those
> > people who were listening started to like the idea [see 1].
> to clarify a bit: my surprise was about blends support in tasksel/d-i.
> I've known about blends for a while but I don't think that topic popped
> up in my debian-boot radar during the whole Jessie release cycle.
I admit I expected *you* to know about Blends for a while - but
considering the video recorded quote I think I was not wrong using this
chance to point this out for other readers of this mail as it is really
a fact that I always meet DDs who mix up this concept with derivatives.
> > Well, Blends and "the desktop situation" could be considered orthogonal.
> > The main goal of a Blend is not primarily to tweak the desktop (even if
> > this could be done). It is rather about the applications. In Debian
> > Med we even have a cluster task which contains exclusively those
> > packages which can be run without a graphical desktop (bio-cloud ).
> I meant the needed changes in tasksel to support both desktop selection
> and blends.
> > ...
> > earlier, yes. The reason why at least I stayed away from this since
> > 2003 (#186085) was that I have seen little chances to change the
> > refusal. However, since recently some Blends of some more general
> > interest like Debian Games and Debian GIS started or gained some
> > traktion resp. the idea came up to rise this question on IRC in the
> > DebConf talk.
> Blends… support in d-i (during this release cycle) was what I meant,
> sorry for being unclear. Hopefully that was covered by the above
> clarification. ;)
Yes it was. :-) However, I also had taken the chance to refer to an
earlier bug (perhaps also to review its old arguments).
> > I perfectly agree that you as the one person army keeping Debian Boot
> > alive (hey, do you like the Blends born idea to prove this point??)
> > should not spend extra time cycles into the implementation.
> That really isn't true, there are many other developers, reporters, and
> patch providers. I'm only adding glue or oil where needed… Of course we
> could do with more hands (look at the BTS), but I'm far for being the
> only one working on d-i.
I agree that my term was a bit in terms of a compliment in the sense of
a "friendly lie". I was not trying to underestimate the work of those
people who are providing smaller contributions. However, you really
find lots of graphs similar like which show the feature of one
dominant person at a certain time. Perhaps you take this as: Thanks
for the effort you spent obviously for debian-boot.
> > That's in fact a quite motivating incentive and I perfectly agree that
> > we really should start rather yesterday than today. The thing is that
> > it is not really clear to me, what we should do rather than adding the
> > packages
> > edu-tasks
> > games-tasks
> > gis-tasks
> > junior-tasks
> > med-tasks
> > science-tasks
> > debichem-tasks
> > ezgo-tasks
> > (multimedia-tasks is not ready according to their maintainer) to the
> > boot disks.
> > Joey Hess as tasksel maintainer mentioned "limited amount of space in
> > tasksel for blends" but this does not give a sensible hint of what exact
> > action we should do now. I think currently eight additional lines is
> > not that much. I also totally contradict to Joey's statement "The
> > 'Debian Pure Blends' effort has been around for several releases and
> > been publicised." and I take  as sufficient argument that it is not
> > the case. Blends were never ever regarded in practice as some Debian
> > internal thing and *every* time when I talk about Blends on conferences
> > and in private discussions I will be asked: "Why don't you do this cool
> > stuff right into Debian instead of a derivative?" It would *really*
> > help in this kind of discussion to point to the Debian installer ...
> > So if we would get some helping hand what exactly technically needs to
> > be done, we could try to come up with some solution.
> I'm not sure we have 8 slots at the moment. I'm pretty sure a scrollbar
> (if at all feasible) in a multi-choice menu would be a bad idea.
I agree here. However, I think it would be a shame to drop a good idea
(and as far as I understood the responses to the bug it is considered
good by several people) since we failed to find a sensible menu design.
> Maybe we'd need a separate prompt for blends.
I perfectly agree that some additional menu level would be the most
natural way in my eyes. I think I mentioned this before. Hmmm, just
wondering why I can't find this term in the previous bugreport(s) since
I always imagined this. May be there is no instance of this since there
never was a real discussion whether we should do it at all and thus
implementation details were not discussed at all.
> Joey will likely be able to tell
> you more about this.
I'm keen in hearing this. :-)