Re: Proposed: Rhythmbox

["Eugenia Loli-Queru" <eloli hotmail com>]

Havoc wrote:
>Note, I do kind of like the idea of including rhythmbox in the core, but
>saying that GNOME has to have _everything_ in it just isn't right and
>probably including everything is counterproductive to progress.

Nobody said to include everything. But I do advocate to include "everything
needed by a modern DE" (that was my sentence earlier). For example, I
wouldn't put gLabels in the core. I wouldn't put gDesklets either. I
wouldn't put MrProject or Dia or GnomeCommander either. These are far too
specialized or duplicates of existing included apps. However RB is offering
functionality similar to what keeps Apple alive with iTunes.  Ditching
*this* specific functionlity out of the Core is a mistake IMHO.

>The argument for Rhythmbox IMO is that it has an important user
>datastore - music files - and I think we should be pushing for a vision
>of one UI for accessing and searching all your data, including music,
>movies, photos, documents, etc. Similar to the advantages of WinFS.

Yes, that's nice and I agree. But we are not on that point yet. Storage is
still a far fetched dream and ReiserFS-4 is still far from mature. In the
*meantime*, a traditional solution should be offered. When Storage is ready,
by all means, ditch what you have to ditch out of the core release to
accomodate the next generation of doing things. But for now, you have to
offer a solution to a very common problem.

>But the counterargument is that so many apps access data that we should
>only have developer platform elements for this, and the apps should be
>able to integrate while remaining standalone. Another counterargument is
>that we don't have this unified data access model yet, and we should
>wait until there's at least some attempt at it.

I do not agree on this. There is no point in "waiting" just because some
programmers do not feel that the code does not do the job as in their
utopian vision. Every minute you "wait" and don't provide a much needed
functionality, it is one user you lose every minute. If MS or Apple would
wait until every engineer did their thing and felt satisfied with it, we
would still be in the Win3.1/OS 7.5 days still.

>We need to keep the core GNOME release small enough to be manageable,
>or we won't be able to keep the release engine rolling.

If Colin and Bastien can commit to RB to keep it up to date at all times I
don't see any problems. In fact, RB is a good "real life test" for Gstreamer
and that would only help both projects. More users, more bugs found, more
progress done.

I am not saying thar RB doesn't have its problems. It does:
http://lists.gnome.org/archives/rhythmbox-devel/2003-December/msg00154.html
but I truly find _laughable_ the fact that people here spend their time over
here discussing as to include or not monkey-bubble or gdesklets to the core
and not Rhythmbox. I mean, come on guys, ask yourselves which app common
people *need* more... If "cool" is not what Gnome needs (as Jeff said
earlier), then gdesklets should not have even being proposed at the first
place.

>If Slackware doesn't like FAM, GNOME including it doesn't force them to
include it.

I beg to differ. Want another example? Evolution.
I would bother Patrick about adding Evolution for weeks on Slackware and
Patrick would just not include it because "it has too many dependancies" and
he got bored chasing these around after 1-2 hours of hacking. But I bet my
head, when Gnome 2.6 will come out with Evolution, Slackware 9.2 will ship
with it. Why? Because Gnome will have figure out the dependancy headaches
FOR him and he would just have to run "make". As I said in my previous
email, Gnome is a SOLUTION for these OS people. If the solution sucks, they
will look elsewhere. If the solution is incomplete, not everyone will be
able to fix the experience themselves, not all of these people know what
matters on a desktop. Slackware is not a desktop distro and so this might
not have being a great example, but other distros are and they are not
performing better than Slackware in this matter anyway.

>I don't think it's GNOME's job to second-guess OS maintainers.

On the contrary, I believe this is one of the main goals of the Gnome
project. And if it is not, it should have being. Otherwise, why do we get
all this cheering and self-congratulations when distros are picking Gnome
over KDE (or the other way around?). Havoc, you are in the position of the
most powerful person in the Linux desktop today (via gnome, RHL and
freedesktop.org). You might not want Gnome to second-guess yourself because
you already know what you want exactly or because you want the terrain clear
of competition for the company your represent. But that would not be too
honest for the rest of us now, would it?

> Speaking for myself, I feel totally free to hack things in or out of GNOME
in the Fedora or RHEL packages,

Sure. You know your Gnome stuff, because you are not anybody.  But the
VectorLinux guys are. The OpenBSD or the NetBSD guys are. Debian people are.
Heck, even the IRIX porters are. These people do not hack on issues, they
mostly have the following mentality: "Compiles? Good. Runs ok? Good. Ship
it". As I said earlier,  not everyone is a Havoc or a Waldo. This lack of
desktop knowledge is not Gnome's fault of course, but it will make GOOD in
the long run to the Gnome project. Because Gnome is a PRODUCT for these
OS/distro people. The better the product, the more satisfied these people
will be and they more will use it. They more they use it and support it, the
more users Gnome will get. That's how you get DE marketshare and kick KDE in
the butt. By using the distros/OSes directly and "sell" them more.

>What we're discussing in this thread is a block of things we coordinate as
a
>release unit, not what end users receive.

Yes, but the part you don't want to admit is that most of these distro/OS
managers only include what comes with the default  gnome installation,
limiting the scope and experience of Gnome in the long run to the eyes of
their users. Gnome should have some fail-safe options to go around the
narrow vision of some of these OS people. Gnome should include the
applications that are needed for a *basic* desktop experience today. And
yes, that does include both a (non-existant currently) good gnome CD/DVD
burning app and Music player/management. Let more specialized stuff, like a
spreadsheet or a diagram app to be included or not in a distro/OS, on the
decision of these OS people. But Gnome should offer the basic functionality
out of the box IMHO ("basic" as defined by the sign of the times).

>So the relevant question is how do we make high quality timely releases
>that are useful to the people building distributions and enable those
>people to make a great end user experience.

Yeah, because that's what you want for Red Hat and Fedora, because in your
mind you have already planned what you want in your company's OS. You know
what experience you  want to offer. You have done your study. But this is
not the case for *most* others and the fact that you work for Red Hat in
conjuction with the decisions that need to be taken here, only perplexes the
situation IMHO. Please do not misunderstand me, I am not accusing you of
anything, it is just that I don't find appropriate for you to vote down RB
(while monkey-bubble might make it in :P) and then watch Fedora Core2 ship
with Rhythmbox in it (as you did in FC 1) and then sit down and stare all
the other OSes and distros not including such a fundamental app by default,
out of plain stupidity or other reasons. You see, that would be the reality
of it. Sure, who ever has the brains will survive in this business, and you
do have the brains (I am a signed fan of yours, you know that). But as far
the Gnome project goes, Gnome should provide for these other people, should
do SOME of the thinking for them, because it is (should be?) one of its
purposes. It is how you get Unix on the desktop. These OS people are your
"retailers" and direct customers.

Also, you seem to think about the whole thing a bit too much from the
programmer's POV, you still try to make the life of the programmer easier
but not the user's. I know you like and read JoelOnSoftware from time to
time, so I suggest you read the relevant page about the MS policy on the
issue. There is an article somewhere where he was explaining his life at MS
and how a must-have user feature should be implemented even if it makes the
code or the architecture and the life of the programmer harder. The user is
the No1 concern in this profession/business, not the programmer (sadly of
course).

What I am trying to say is that Gnome's "customers" are these OS/distro
people, not directly the end user. We agree there. But by only caring about
giving them easily maintanable code is ONE part of the equation, not the
whole deal.

take care,
Eugenia