Re: focus! (was Re: Focusing on innovation re: mono, python et al)

[Havoc Pennington <hp redhat com>]

Murray Cumming wrote:
> > The word "desktop" is like a cancer. Its problems include:
> >   - it's vague as hell
> [snip]
>
> The "desktop" includes stuff that everything (apart from very tightly
> focused embedded stuff) needs. Vendors who don't need some part of the
> desktop usually don't want any part of it. So, it's just a "base" that
> isn't yet a development platform.

I don't agree. I think it's stuff oriented toward the niches that have 
already been adopting GNOME. That's why they've been adopting it, and 
others have not.

I mentioned these niches in my earlier mail; the computer lab / thin 
client; the unix-to-linux transitions; the Fedora/Ubuntu tech 
enthusiasts; tech workstations for animation and software development; a 
couple others perhaps.

It doesn't mean anything to say that some of these audiences need "a 
photo app" and some other possible audiences also need "a photo app" - 
the issue is not specific tech items, it's the whole package and the 
benefits it offers as a whole. The current audiences are the ones who 
need "a desktop" as GNOME has defined "desktop."

Much of the "very tightly focused embedded stuff" based on GNOME
tech is IMO focused on larger and more mainstream audiences
than the so-called "general purpose desktop" is. So why does GNOME get 
so stuck on "the desktop" (by which we mean "the 
enterprisey/thinclienty/unixy desktop") and act like everything else is 
some kind of distraction? If GNOME were Apple we'd be sitting here going 
"gee, the iPod seems too tightly focused, we need to make a desktop, not 
just a music player"

You have to talk about "need" in the sense of "person XYZ already has a 
bunch of stuff, why do they need what we offer also? Or why would they 
bother switching - what's so compelling? And how does it relate to what 
they have?" - and then you start thinking about what to offer that they 
_don't have already_ in any form. To the niches GNOME has been 
successful in, GNOME _does_ offer stuff they didn't have before, and 
without breaking what they had already.

Either the goal is to spread open source and build stuff people want and 
will adopt quickly; or the goal is to "make a desktop" for "everyone"

"make a desktop" is about as clear as "make a website" IMO. It doesn't 
say a thing about what sort of desktop.

Don't get me wrong; again, GNOME de facto has more focus than that on 
the specific audiences I've mentioned. The whole point here is: is that 
really all GNOME aspires to? And if not, when is someone going to 
acknowledge that it's the de facto reality of the desktop release (as 
opposed to the rich ecosystem of related projects)?

Because the current audiences ain't gonna reach the 10x10 goal, I assure 
you of that.

> As for bringing in new functionality and allowing varied focus, I still
> think this could be done with additional release sets such as
> - Productivity:
>   Spreadsheets, Word processing, Slides, Databases, Publishing.
> - Creativity:
>   Photos, Graphics, Drawing, Video- and Audio-editing, sharing, mixing,
> augmenting, collaborating.

This is splitting by codebase, or some kind of abstract taxonomy. The 
split has to be by _audience_ and _benefit_.

The splits may well involve _overlap_ in order to do this.

It's also critical to avoid terms like "photos" or "spreadsheets" as if 
they are all the same. Photos for who? Spreadsheets for who?

For example, one large set of people using photos might want some very 
simple way to 1) share photos with family 2) make a slideshow 
screensaver and 3) order prints online.

Someone else, a serious photo hobbyist, might want 1) detailed photo 
cataloging/organization 2) Flickr integration 3) Photoshop.

You can't just say "photos"! It's too vague.

We all do this, but that doesn't make it right.

(Can the same app serve the two photo audiences I mentioned above? 
Maybe, but "can we share code?" is a second-order question, not the 
organizing principle.)

> In reality, however, all end users and vendors will want everything. But
> the vendors will just prioritise on some of these parts.

That's only the case if by "all end users and vendors" you mean "most 
end users and vendors of linux distributions" and just ignore anything 
that isn't a linux distribution.

There's also Windows apps, "embedded" (focused?) devices, online 
services, all kinds of stuff that could serve the goal of bringing an 
open source computing platform to the general public.

Havoc