RE: Gnome Packaging

["Matthew S. Rinehart" <msrinehart cs millersville edu>]

On Sat, 2003-12-06 at 21:55, Murray Cumming Comneon com wrote:
> > On Sun, 2003-12-07 at 02:17, Murray Cumming Comneon com wrote:
> > > Apart from debian users, who seem used to choosing their 
> > packages in 
> > > boring detail, I don't think this is likely to make much 
> > difference to 
> > > most people. We want GNOME to include everything that the 
> > user needs, 
> > > and that's what the distro wants as well. So I don't think most 
> > > distros will even bother to ask the user whether he really 
> > wants all 
> > > the things that he's likely to need.
> > 
> > Ergo the distros would include everything irrespective of how 
> > things are packaged.
> 
> Yes, we hope they will package everything (everything that is part of GNOME)
> because they have the same aims. So, yes, packaging is irrelevant to distros
> and users.

First, I'd like to say that I and a couple of my friends are Mandrake
users, and we all prefer to choose only those packages we need.  Our
desire to cut out packages isn't because of lack of disk space, but in a
lot of cases out of a desire to eliminate needless redundancy.  If I
only use command line tools like scp and ncftp to transfer files between
machines, I have no use for gFtp and it would be pointless to have it
installed.

Most installers I've encountered will generally install whatever
packages the distro feels the user should have, unless the user
requested to select individual packages.  As you stated, packaging is
irrelevant to the distro and to some number of users (not all, however),
so for those of us who like a per-package level of customization, why
not make Gnome subcategories?  It'll have no effect on those who don't
use it, but it will be very helpful to those that do.

>  
> > > Why would people want "cut out evolution". This suggests either
> > > - There is a shortage of disk space. I don't think there is.
> > > - We are forcing something on distros. I think distros do what they 
> > > like.
> > 
> > I've just read it a fair bit since the proposal to include 
> > Evolution was announced.  Some people like lean, mean 
> > machines.  This type of applicaiton association would cater 
> > more to them I guess.
> 
> Personally, I don't care about those people. "Keeping my machine lean and
> mean" is not a goal that I think any representative user cares about.

The desire to cut out evolution may not have to do with disk space, but
that possibility does exist.  Just because today's hard disks are rather
large doesn't mean that a person has disk space to install anything and
everything.  Suppose it's a multiple-boot system, for instance, or old
hardware, etc.  Perhaps whoever is installing Gnome wants a minimum of
user apps because he's setting up a kiosk for Internet browsing or
demoing something.  Evolution would be pointless to have installed on a
such a machine.

I'm curious to know what you feel is a "representative user."  My
opinion is that the users of Gnome are so diversified that a single
hypothetical representative user would be difficult to come by.

>  
> > Another benefit of such a devolution would be that authority 
> > can be easily delegated based on each group.  I could ramble 
> > on about what that means for quite some time, but the crux is 
> > that the formal relationships give a little more organisation 
> > to the whole Gnome structure, thus it is easier to manage.
> 
> Yes. But we already have Gnome Office, don't we. It just needs to get with
> the programme. It seems to be getting better recently. Still, I don't see a
> release schedule.
> 

Yes, we have Gnome Office as a separate entity.  Why not also have
Gnome-Net be a separate entity, for instance?  An individual can only
effectively manage 20 or less people.  By separating the parts of Gnome
into separate categories and having each category have its own manager,
the development process might become more efficient.  I think that's
what Charles meant, at any rate.

> > > I do personally think it's worthwhile to have a separate 
> > release set 
> > > for big things like Office
> > 
> > But is Evolution an Office application or a base desktop 
> > application?  
> > If Evolution and it's groupware functionality is present, 
> > then surely a suitable word processor should be present?  
> > Wash, rinse, repeat...
> 
> Again, If we accept that it's just for the sake of release organisation then
> we don't need to worry that the boundaries will always be vague. A user
> doesn't care whether evolution is a core/desktop application or an office
> application - he just wants to read and write emails.

A user who doesn't want Evolution will care where it is packaged so that
he can find it during an install process (or perhaps afterwards in a
list of installed packages) and remove it.  Personally, I'm of the
opinion that Evolution is an Office application, if only because Outlook
is in the Microsoft Office package, and users used to MS Office would
expect to find Outlook's counterpart in an Office package.

>  
> > The inclusion of Evolution is treading on some unsteady 
> > semantic ground!  A simple bit of devolution would solve that 
> > and introduce some more structure to Gnome.  Whether extra 
> > structure is needed or not, well, maybe it isn't now but in 
> > the future?  Who knows. :)
> 
> People are only bothered about these things when they think it means
> something that it doesn't. If Gnome Office is a well organised and scheduled
> release set, then whether Evolution is in Desktop or Office makes simply no
> difference for the user or the developer.

For the same reasons I mentioned in my last paragraph, where Evolution
and other packages are grouped does matter.

I think the argument basically boils down to this:

The users that DO NOT care about package organization have nothing to
lose and nothing to gain from dividing Gnome into subcategories.  It is
simply irrelevant to them, as you said.

The users that DO care about package organization gain a better
organizational structure for customizing their systems as finely as they
wish.

Neither group has anything to lose from doing it, so why not?
> 
> Murray Cumming
> www.murrayc.com
> murrayc usa net
> _______________________________________________
> gnome-list mailing list
> gnome-list gnome org
> http://mail.gnome.org/mailman/listinfo/gnome-list
-- 
Matt Rinehart                              http://kitsune.homeip.net/