Re: [Usability] Re: ui-review of new modules

[Marco Pesenti Gritti <mpeseng tin it>]

> The problem here is lack of coherent vision and desktop wide model. Without 
> it, everyone does it a little different.

Here is a perspective from a maintainer point of view. I'm not a
designer but I have to:

- make a call between different design proposals
- more often ensure the user interface changes I accept are coherent,
basically convince someone to make a design for them or try to do it
myself

I think the problem can be seen at 3 levels. (but there is a strong
interaction between them).

SKIN DEEP LEVEL

I think GNOME is addressing this quite well. The HIG is an excellent
tool to be able to make decisions and to ensure consistency. Several
contributors are learning it and submitting bug/patches to address the
problems. I think the quality of software at this level is improving
fastly, even for projects outside the core.

INTERACTION DESIGN AT APPLICATION LEVEL

I think this is the level where we have more problems. Most projects
skip the step directly.
We lack:

- A documented metodology.

Comparing the way we produce code and the way we produce interfaces I
see two differences:
1 Most contributors has at least basic clues about code (they usually do
it for work or school) but they dont have a clue about user interface.
2 Sources are an excellent way to learn coding. Unfortunately there is
not the equivalent for interface design.

We need API documents (a documented metodology) and source code
(documented concrete designs).

- Designers

Maintainers or core developers often has simply no time to build a
rational design for features/projects, they tend to defend code and the
intersection between the two interests is probably not very high.
Obviously that doesnt mean a coder should not design an interface, I
think combining the two things can even be useful.
I think there a lot of potential contributors in this field. But unlike
for coding, they have no way to learn.
Maybe some of the people who just speculate about UI on mailing lists,
with proper tools and metodology could make a more useful contribution
to the projects.

- Power to the designers

In my experience working with designers is one of the coolest thing.
Being helped to develop something more usable is very satisfactory.
Though I think there a few factors that are blocking a more strict
cooperation:
- Often mantainers doesnt trust designers. I think that mostly depend on
the lack of evaluation criteria and on communication problems from both
parts (often we just talk a different language). Setting a metodology
and evaluation criteria could improve the situation.
- Mantainers are heavily influenced by their apparent user base (people
complaining on the mailing lists). As I see it user interface design is
an iteration process, often implying temporary regressions. Changes
itself generate a strong resistence.
- It's very important to have designers strongly involved in the
project, following the whole development process. This conflict with
their low number and the lack of trust though.
- Finally current maintenance model has his disadvantages. IHMO current
maintainers are mainly code maintainers and their power on user
interface decisions is too high. But I think this is due to the lack of
connection between the two fields. When you realize a designer is a
strong help to reach your targets (for example having more users), you
tend to trust him more and so to give him more power. Probably a
different maintenance model will be the result of a different
interaction between developers and designers. I doubt trying to impose a
different interaction by a different maintenance model has many chances
to succeed.

DESKTOP WIDE MODEL

Unlike Dave I think we are not that bad at this level.

IHMO a desktop wide model will inevitably be an iteration process and
the interaction between the single projects (and their different needs)
will have an high influence on it.

Still we need a common coherent direction to orient our work. I have the
impression there is some consensus on the basic lines to follow (for
example is there really a debate about document oriented interfaces ?). 
Debating it publically would have probably the effect to make it
stronger.
I think the usability team have a reasonably common vision, it's matter
of communicating it and help developers to translate it in concrete
designs of applications/features.

Hope it's useful.

Marco