[Fedora-marketing-list] Fedora usability : a new project?

[Nicolas Mailhot]

Le mardi 08 août 2006 à 16:08 -0400, Máirín Duffy a écrit :
> Nicolas Mailhot wrote:
> > Also I have my fingers in enough pies today I hope someone will take up
> > the program and make it real, but I've little time to contribute myself.
> > Though this is a direct result of trying to engage upstream alone as
> > we're supposed to do nowadays.
> 
> Are there any other distro-specific usability groups out there?

Novell I think has an active usability project

> I'm kind of confused at the goal of a distro-specific usability group. 
> Doesn't it make more sense to work with an upstream usability group such 
> as the GNOME Usability Project (which is somewhat dormant AFICT)? Has 
> someone tried to engage upstream via projects such as this and met with 
> resistance? (this is the feeling I get from your email?)

The problem is actual users won't have "GNOME problems" or "KDE
problems" or "OO.o problems". They'll have Fedora user problems.

So if on Monday a user hits a usability problem in OO.o we ask him to
integrate the OO.o ecosystem.

On Tuesday frustrated with OO.o he tries koffice (he didn't set up
fixing OO.o after all - he want to make it work within Fedora not wait
months before the next OO.o release). It works but then there is another
problem which wasn't present in OO.o. We'll tell him "finished digesting
OO.o upstream - try KDE now"

And so on.

(you'll note that when a user reports a crash we don't ask him to report
upstream directly - if we followed this logic we know perfectly well a
lot of problems would never be fixed satisfactorily)

Who there has enough time to follow every single Fedora list and IRC
channel? On the usability front we're asking users to do an effort
several orders of magnitude higher - engage with every single upstream
they may use as part of their everyday Fedora experience. That's not
realistic.

Red Hat/Fedora moved from a "patch everything to accommodate Red Hat
customer wishes" to "use as vanilla upstream as possible". This is all
well and good but I don't see why using vanilla upstream should prevent
the Fedora community as a whole from telling upstream what we wish in
the next release. Other distributions are not afraid to do so.

This requires actively collecting and discussing problems at the distro
level. Sadly almost never done in Fedora - the sole exception I remember
was the new print tool requirements. The influence of Fedora as a user
community over upstream is so diffuse one can ask if it even exists at
all.

> Where I see a distro-specific usability group being useful:
> 
> - Software update tool usability (because different distros handle this 
> differently)
> - Infrastructure usability (e.g. usability of the Fedora website(s), 
> account system, bugzilla, etc.)
> 
> But neither appears to be the goal...?

Apart from the general "help bridge the gap between upstream and users"
you need to consider the following.

A usability problem means "it works technically but it's awkward for
me". So:
1. there is a way to get the job done nevertheless - making it not worth
spending the time reporting if it's difficult (and unless you already
work with the affected upstream it will be)
2. lone reports may be subjective

You can only do good usability work by correlating many user reports. If
you make it hard enough for reports to percolate upstream, the few which
make it will be dismissed as "non representative" especially if they
don't fit in the user abstraction upstream devs have (if they did fit in
it chances are the problem would never have happened in the first place)

One major usability problem is inter-app inconsistency.
By definition this kind of problem impacts many different upstreams.
And it's dependant on the actual software collection the user runs.

The Fedora context is very specific - we refuse to ship many apps which
are good enough legal-wise for competing distros. Therefore our users
won't have the same problems as a distro who ships mplayer instead of
totem for example. Similarly the userbase of a project like OO.o will
span several operating systems, so Fedora usability priorities won't map
to OO.o usability priorities, and if Fedora does not make the effort to
make its case loudly OO.o will just develop for their biggest port -
Windows.

And so on. 

-- 
Nicolas Mailhot
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: Ceci est une partie de message num?riquement sign?e
URL: <http://listman.redhat.com/archives/fedora-marketing-list/attachments/20060808/b945bcde/attachment.sig>