FC4 slimfast slimfest
Michael Favia
michael.favia at insitesinc.com
Tue Feb 22 21:54:38 UTC 2005
Elliot Lee wrote:
>Hi all,
>
>Here's a heads up that we need to get rid of about 300M of packages to
>make sure that FC4 continues to fit on 4 CD's. Right now eclipse, xfce,
>xemacs, cfengine, and all the games are leading candidates. We also may
>try to start removing stuff from Core in hopes that it will appear in
>Extras if someone cares about it.
>
>Nothing is set in stone, but if you want to write in favor of keeping
>something, you'll have to suggest something else of equal or greater size
>to axe instead.
>
>You can view the list of new packages since FC3 (sorted by package size)
>at http://people.redhat.com/sopwith/new-packages.txt
>
>FYI,
>-- Elliot
>
>
>
*Abridged Version*
Developers can be expected to work a little harder to establish a
working environment and have the knowledge skills and abilities to make
that happen. Consumers neither know how nor want to know how to work the
magic that makes everything the way they like it they would just like it
done for them. Towards this goals improve the end user experience so
that the developers continue to develop for a growing audience instead
of a shrinking one. I don't know if this is possible on short niotice
for FC4 but may be worth review for architecting fc5. Thank you and good
luck.
*Unabridged Version *
this seems like a discussion that is held frequently and is always
contentious. Instead of arguing the acceptability and merit of a
particular package perhaps our time is better spent working up a rough
framework of what our *specifications* are first. I think we are
designing the perfect solution to the indescribable problem here. I
would like to propose the following for consideration:
New/Average consumers represent the least experienced and the largest
growth segment for Fedora. They are most fickle in their adoption and
most skeptical in their approach. Developers, server administrators and
every other use case is commited to fedora for more substantial reasons
in most cases (development model, release cycles, etc) and have the
knowledge, skills and ability to take the time and perform and extra
step to transform the default working environment into their desired mode.
* New users that have little or no Linux experience seem to be
interested in Fedora at an increasing rate. I argue that they will
choose to adopt fedora if we make the dfeault environment as smooth as
possible for these users who have no knowledge and arguably shoudl
require no knowledge of repos and configuration files for things to Just
Work.
* Developers are basically knowledgable and communicative. If something
is not as we expected we ask each other on IRC or in a mailing list or
forum. We know how to configure repos and upgrade and install new
packages. If we don't it can at least be argued that we should be
able/required to learn.
* Fedora adoption of both consumers and developers is a good thin and
there are many more consumers than there are developers out there.
* Fedora relies on developers and will do so at an increasing rate with
the launch of fedora-extras
No I'm not arguing to purposefully make life more difficult for
developers. Instead i am arguing for the "greater good" created by the
developers ease of overcoming simple deviations and how detrimental
those same obstacles would be to the average/new user. To that end i am
suggesting that the user runtime environments (including a java runtime
that is differentiated from eclipse if they are currently bundled for
some odd reason), default desktop , and default applications (office,
web, file, mail, message) should take priority on the first CD['s]
(including required desktop libs to satisfy default apps). Once the
requirements for that are met than either place the alternatives on the
next/remaining cd's followed by the development tools or vice versa
depending on which user segment you think will benefit greater.
-mf
More information about the fedora-devel-list
mailing list