Re: FW: Nautilus and Setup Tools

[Arturo Espinosa Aldama <arturo ximian com>]

On 30 Oct 2001, Havoc Pennington wrote:

>
> Seth Nickell <snickell stanford edu> writes:
> > Definitely. GNOME needs to functioning and/or thinking of itself as an
> > operating system rather than a thin graphical shell over the "real
> > operating system".
>
> Neither of those. GNOME might think of itself as a component designed
> to drop in to an operating system and function seamlessly as part of a
> usable OS. Remember most users will _not_ be downloading GNOME from
> gnome.org, they will be getting it from OS vendors.

Yeah. This is still compatible with the idea that GNOME should provide a
common way of configuring the underlaying OS at a basic end-user level,
the same way GNOME is supposed to look the same across OSs.

> One bit of braindamage in thinking GNOME is an OS is that we start
> reinventing lots of wheels at the wrong place in the dependency chain,
> because we get the delusion that all apps that matter will use GNOME
> stuff. And that's just not true - it's hubris plain and simple.

In the case of the XST, we always code things thinking that other
utilities may be used to configure the system, or that XST may be later
uninstalled by the user.

> At some point, a decent user experience simply requires a lot of
> OS-specific integration work. I don't think there's much way around
> it. And certain kinds of progress are going to require new OS
> features, such as file change notification, and removable media
> handling, and USB/hotplug device handling, and PCMCIA handling, and so
> on. Replacing the kernel with gnome-kernel or the PCMCIA scripts with
> GNOME-pcmcia is just not a solution.

The solution is portable scripts with intelligent graphical interfaces,
that mutate depending on the OS capabilities. We've got this too.

> One of the pesky things about OS integration work is that it has to be
> done _before_ the OS is released. Which is pretty tricky for the GNOME
> project to do.

That's a problem: the XSTs have to "catch up" with the new releases, but
it seems like users are willing to wait; fortunately, not everybody
upgrades the next day a new release is done. I hope the distributions
eventually find the value of the XST project and start working actively
with us.

> The implications of these metaphysics for XST are less clear, I talked
> to Chema on the phone about it yesterday.

I'll ask him about your conversation.

Greetings,
Arturo