Gnome Accessibility (Here At Last?) - Bill Hanemen, Marc Mulcahy

The answer is somewhere between yes and no (here at last)

what is this all about, eliminating barriers to access, means gnome desktop
and utilities are usable by people with diasabilities
users who can not use mouse or keyboard or whatever

includes people who dont think of themselves as disabled, those with rsi,
elderly, colour blind, explains the one in six figure sometimes cited as
number of disabled people who use or need to be able to use computers

why not only make specialised apps accessible, because you want equal access
(tro new features etc), choice is good, also shared machines (reduce cost make
things more available)

there are laws around the place that require softwre to be accessible, this is
why companies who need to comply with these laws have been pushing this (sun
have been behind a lot of the push and doing a lot of the work to make gnome
accessible)

the law in the US is 508 (and some others) there is a 508.org or something
with information. a lot of very specific stuff in the law, no software
complies with all of the law, so the requirement is that the most accessible
purchase is made.

key aspects of accessibility
keyboard navigation (this would make the desktop easier/faster for other
people also), theming, exporting/enabling a consistent api, other stuff

the theming issues is because of the need for things like high contrast or
low contrast desktops (almost black on white was demoed)

windows is accessible, but mostly only because of long term revverse
engineering approach put in by third party companies (every time windows
released all the accessible software would break or something) now days
windows has a pretty limited support for accessibility, MSAA is an api
approach, limited, also not used much at all by MS Office.


GPE - GNU PDA Environment, Nils Faerber

Motivation, there are pdas with linux on them, but no common environment,
until now there is no prodcutive PDA environment for linux, most developers
seem to think it is like a miniature workstation, this is not the case, using
tiny window manager and miniature xterms just wont work (QTopia and OPIE do
exist, though they are GPLd which Nils sees as a problem) Qtopia and OPIE both
are a windowing toolkit and the gui server all in one, unlike X.

GPE is about
good env. a basic set of apps, framework for developers, design and style,
apis, tools (daemons, libraries, helper apps (data storage and synchronisation)

Diversity is nice however there are not much resources on a PDA so having both
GTK and QT installed as we do on desktops is not an option

compaq iPaq is a baseline pda power/memory wise available for the linux
market (the sharp sl-5000 (Zaurus) has recently become available also) this
is the reference level platform

all of nano-X, pixo-GUI, GTK-FB, QT/embedded, and X11 have been evaluated, X11
proved the most powerful and featureful and comparably least resource intensive
system, so X11 was chosen

most data will be stored in an relational database system, this sseems silly at
first howrever they claim the use of SQLite library for this backend (~250KB)
will work well, all data access is through a daemon on top of this rdb.

sql advantages
   - well known interface
   - no need to invent new file formats
   - redundancy reduced with relational storage, no need to store same data in
     many places
(I would wonder why they could not simply use something like tdb (see samba
cvs) for a very small compact and fast storage mechanism, though I suppose
arguments of complexity with using DB style things may come into it)

xml was decided against as the data storage format for various reasons (see
the slides)

There is a rather cute feature available due to the use of the new X11 small x
server, they can have the orientation of the display changed at runtime from
landscape to portrait.

the uri for the project is gpe.handhelds.org


Bastien Nocera - Coding apps for the gnome desktop

design an app, do not code immediately, have questions to answer (?)
(bastien says hacks live a lot longer than you may imagine they will)

rhythm box project used as an example (rhythmbox.org)

goals of the application? (creating a full fledged digital music juke box)
plan of features at given time you want those features

look for existing applications (learn what is good and what sucks)

draft gui design
should look like a real gnome app, menu bar, status bar, tool bar with stock
gnome icons, a minimum of custom widgets
maybe use glade (it does all this), also allows sharing of the interface
creation with non programmers
glade also allows you to see very quickly what the interface may look like
(rapid prototyping)
(there are gnome style guides that should be looked into here probably)

do not use icons from windows programs, macosX programs etc, especially dont
use those hard coded within apps, they make things harder make software
consistent and you may be sued by someone

how to be a lazy bastard
hunt for librarires, eel used by nautilus, gal used by gnumeric, libegg (?),
widgets from some apps... directory selector in gimp, etc
get code from other places
use the search engines, google, freshmeat, sf, kde cvs (if so inclined).
lots of existing code is out there to do things you may need, such as mpeg
decoding, midi, etc

bastien has a brilliant accent, as he said at the start of his talk, he is a
french cockney, he is french but lives in london and has a cockney accent on
his french sounding english. it is a must hear :)

GOB, easy way to write GObjects and GtkObjects
boilerplating for you
this creates gobjects and gtkwidget objects boilerplate code for you, also can
do the gnome corba idl stuff

some gob examples, how to do the nice looking OO thing that gob will turn into
big and ugly headers and c files doing the gobject OO stuff

Componentisation, using/creating libraries, bonobo components
do not simply make programs that sit on top of some command line programs,
communicating with command line process and all the handling of it is far more
difficult than simply using libraries such as those the command line programs
themselves would use.

be nice to people around you, buy developers beer (or I suppose pizza as the
case may be)

getting your program known, announce things, sf.net, gnome-announce, gnotices,
freshmeat.

Maybe just maybe there is already a gnome application doing what you want but
it is currently not as nice/good as it could be so you may want to work on
that, and as you will not be the primary maintainer you will not get a lot of
the annoying stuff involved in that role (bugs, email from users, etc)

never underestimate the time it takes to program and the amount of effort
involved in creating a good program that people will use and will become
popular
the gnome (and others) community should look to work on exzisting apps not
creating new little apps all the time, do not make a new irc client, modify an
existing one, 15 new crappy irc clients are not as worthwhile as 1 or 2 good
clients (the same applies to other applications)
it is very easy to write quick hacks, it is very hard to write good stable
usable apps
things you assume could not happen do in fact happen on other people's
machines, avoid making assumptions, as always assumptions are a bad thing


The new gnome website - Jeff Waugh

jeff almost didnt want to do the talk as too often people talk about the new
gnome website and then never do anything about it

they seem to be planning the debut of the new website at the time of the
release of gnome 2.0 desktop environment (may possibly)

he has gone back over some of the very old websites to get an idea of what is
there and what changes (also what does and doesnt suck) he showed off a series
of old www.gnome.org and developer.gnome.org websites going back to the
inception of the project)

not much wording on the site has changed for a long time, a few weeks ago they
put the guadec logo up but otherwise not much

the directory that contains the website on the server is about three times
larger than the cvs history as nothing is ever removed and junk is added often

the new website needs to look at different classes of users
users, developers, potential developers, new corporate interests, press,
international audience

for potential developers it is really hard to say to someone who wishes to
help out, just go to this page and you can see stuff that needs doing, they
sometimes tell people to go to see bugzilla but until recently it didnt have
good stuff, and is still is hard to use.

corporate interests, they need to say to their customers here is gnome our new
desktop see how professional and good it is go and use it

no press kit, the information about a whole lot of gnome is internalised in
the minds of developers, there needs to be more stuff for press to easily be
given information they can use about different aspects of gnome

international audience, the current site is only in english, not any other
languages, too hard to make the current website internationalised

the architecture of the website the current site is small chunks of html, with
an evil perl scrtipt that processes them and adds all sorts of random stuff.
the new system every single page in the website cvs is an xhtml page, so all
standard web pages effectively that work from all editors, after that it all
gets processed by xslt style sheets, the build system is also easier, rather
than using make files that need changing it just picks up stuff as it finds
it, also for documentors, the docbook xml is simply uploaded, docs generated
automatically from it and authors no longer have to generate the html
themseleves and it will no longer be out of sync