infapi00 – Page 4 – I have seen the truth and it doesn't make sense

ATK Hackfest is coming!

Just a reminder, the next GNOME hackfest will start in a little more than a week. As usual, we are organizing it on the wiki, but in summary, the tasks that will be the basis for the debate are managed by bug BG#638537.

As I explained on my previous post, the objective of this hackfest is a kind of ATK 2.0 kick off. ATK has served well during these years, but was mostly unchanged during almost ten years. After all these years it is time to use the experience from implementors and AT developers to improve it.

New sponsors

We have confirmed new sponsors to this hackfest, joining the original ones, GNOME Foundation and Igalia: GPUL and Xunta de Galicia.

GPUL supports the ATK HackFest organizing an open session about accessibility. GPUL also provides economical support. This contributions is possible thanks to the budget from the agreement between Galician Free Software Asociations and the Xunta de Galicia (Galician Regional Goverment) in 2011

In related news:

One important new from the last a11y weekly meeting.

One of the drawbacks of at-spi2 was that due lack of resources, there wasn’t any plan to create a replacement for C-SPI. That means that only pyatspi2 was available, so any AT tool that wanted to interact with at-spi2 was supposed to be written using python (note that I’m talking about the AT side, application accessibility support based on ATK is still mainly written in C).

This has changed, as one of the last changes made by Mike Gorse on how the python bindings are created is write a core C library, and use gobject introspection. So now any AT developer has libatspi available. It is not API compatible with C-SPI, and it is not planned a compatibility layer, but at least now it is in theory possible to write an AT tool in C, and any existing one would require to migrate to it.

The next step would be improve libatspi documentation, in order to upload that to GNOME developer documentation page.

Announcing ATK/AT-SPI Hackfest

This will come as no surprise to many of you at this point, as we already talked about it during the weekly accessibility meetings, but I would like to officially announce that we are having an ATK Hackfest in May.

For the new people, ATK is the accessibility toolkit. Right now the main and more extensive implementation is GAIL (GNOME Accessibility Implementation Library), although it is also used in Clutter (Cally), Java (via the Java Access Wrapper), and others. AT-SPI is the accessibility service provider interface. Accessibility tools “uses” AT-SPI in order to interact with the applications. ATK works as the common language that target applications uses to expose themselves to AT-SPI, so a deep change on ATK would require an update also in AT-SPI. In the same way, we would like to get a one-to-one relation between the interfaces on ATK and AT-SPI as much as possible.

Applications and toolkits do not all implement ATK consistently. This has a negative impact on the assistive technologies ability to provide a consistent cross-application user experience. Even in applications and toolkits in which the ATK implementation is complete, the information obtained from a single event and/or object is not always sufficient for an AT client to proceed immediately; instead it is often necessary to perform further queries and make decisions based on heuristics rather than concrete data. This has a negative impact on both performance and reliability.

Accessibility team has come to the conclusion that this is time to improve on ATK no matters if that means an API break. The primary goal is to take what we have learned from years of ATK/AT-SPI and make things better for all involved, and create the basis of what ATK 2.0 should be.

Location: Igalia (A Coruña, Galicia, Spain)
Dates: 9-15 May

Who Should Attend?

Developers of applications and toolkits which implement ATK
Developers of ATK and AT-SPI
Developers of Assistive Technologies which rely upon AT-SPI
Developers of testing solutions which rely upon AT-SPI

For more information, including the proposed task, pre-event “homework,” venue, and associated costs, please see:

Live GNOME ATK Hackfest page
Metabug “Towards ATK 2.0” (feel free to open more bugs or comment on the current ones)

For the moment the event is sponsored by Igalia (Venue, part of the food) and the GNOME Foundation (travel+hotel costs as far as the current a11y budget allows it). If any other company want to fund this hackfest somehow, please contact me or Joanmarie Digss (and thanks in advance)

My 5th anniversary as a Igalian

5 years ago, 13th of January 2006 I became a Igalian.

2 years earlier I had a student internship on the old offices, and it started my ‘brainwash’. Although I started to write Pascal (my first real programming language, as I don’t include my little LOGO experience when I was 10) since I was 12 years, during a lot of years I was really skeptical about this “free software thing”. And sincerely, the reason was because the first people that tried to convince me about Free Software were using this typical condescending “Free Software is just cool and will save the world from the Evil” attitude.

Whatever, forget that off-topic rant.

My first day on this internship was funny. After my first car accident I arrived to Igalia office, and Alejandro García (Alex, our beloved Igalian in the shadow) started the tutoring. “Do you know something about gtk+ ?”, “No”, “Do you know something about autotools ?” “No”, “Do you know something about glade ?” “No”, and finally with a desperate face “Do you know something about C ?” “Yes”. (Because in the end, we still have C)

The summary is that this “free software thing” was the best environment to keep doing something that I beloved since I was 12 years old: program. Decide to do a program, and with your big or small knowledge of a programing language and the remaining power of your last operative brain cells, conclude the right steps, write a program and feel satisfaction when you found that the program does what it was supposed to do, learning from other people code and sharing with them because reinventing the wheel is just a waste of time.

5 years and a internship. A long time related with free software, with a lot of projects involved. First with Fisterra. And after Fisterra, several other projects, Maemo (now Meego,) Hail, Cally, the Accessibility Team and so on.

And on the best place to do so. This post has become somewhat disperse, but remember that the title is “My 5th anniversary as a Igalian”. All this experience during those years showed me that I was not wrong about how I saw Igalia when I entered and about their advantages and philosophy. Igalia tries to obtain the real value of the workers, and share the concept of the company among all of them. They don’t work because someone give you a order, the people work because is their company.

In the end, probably this is the best idea, involve all the people. All the companies ask their workers to “work for the company”, make an extra effort as you “should think in the future of the company” without giving you any extra value or greetings. As Igalia share the responsibility, and the final purpose of Igalia is become a partner (a real igalian, in the end), you “work for the company” and “think in the future of the company” as YOU are the company.

Igalia real value is the group of people that conform that, and the synergy of these people working together. A company I’m proud of belong to.

Concluding, I’m old enough to say that I have worked 5 years for an (almost) 10 years old free software company. Some people says that 5 years on computer science is a lot of time, enough time to be obsolete, but as Jethro Tull says, I’m “Too Old to Rock ‘n’ Roll: Too Young to Die!”. So as my plan is not die today, lets rock and roll 5 years more!

GNOME Shell accessibility status

These days some people have started to ask about the current GNOME Shell accessibility status, probably a collateral effect of the Boston Summit, as Joanmarie Diggs and Alejandro Leiva (Orca maintainers) were talking there with GNOME Shell developers, mainly about the current Universal Access UI and how fit Orca on the new ui experience. They are planning to make a summary post so stay tunned.

Anyway, I would like to share my view of the current accessibility status on GNOME Shell. Using the GNOME Shell roadmap as a guide:

Magnification: Joseph work has been integrated on GNOME Shell. Not as a extension or something like this, it is a real feature of GNOME Shell. So good news.
Design and implement keyboard navigation: Some months before GUADEC Dan Winship started to work on on keyboard navigation, and it has been really active these past months. This includes support on the Shell Toolkit itself (already on master), and then use this on the different regions of the shell, like the overview or the chrome, which provides a promising keyboard navigation. Note that in the case of the overview this is somewhat on a idle, as it is planned a overview relayout (check the overview-relayout branch).
Hook up accessibility to ATK: basic ATK support is right now directly provided without loading any Cally module, as Cally is now part of Clutter, and part of the stable Clutter 1.4 release. It is true that Cally can be more complete (this statement applies to any library), but the current status should be enough for GNOME Shell purposes. Of course, new bugs can be detected, and I should need to take a look to this bug, but I don’t see Cally as the main issue here. In the same way, I provided a patch proposal to load the atk-bride on GNOME Shell. The main problem with this patch is that although working, it is not clear if it is the final solution.
Comprehensively make controls accessibility: new elements are being defined on the Shell Toolkit, and most of them would require to export appropriate roles and actions. I have started to review this, so I have already created a patch for StLabel, and I have some other items in my TODO, like StWidget tooltips, and the new objects declared on the Javascript code. This is related with the previous keyboard navigation task, as I’m detecting most of then while testing the keyboard navigation patches. So I reported some additional bugs related to provide extra ATK information on the overview navegation (with patches) and the chrome. As there isn’t anything equivalent to gtk widget mnemonics those patches are using direct calls to ATK methods.
Theming: it is already reported as a bug, but AFAIK, there is not immediate plans to implement it.

So, if you apply all these patches, you get Orca speeching out most of the events from GNOME Shell. Part of the sections of GNOME Shell can be used using just the keyboard. In the cases that GNOME Shell is using the new core keyboard navigation support Orca reacts to the focus change, although the information exposed is only meaningful on part of the cases.

Apart from this, GNOME Shell includes an accessibility tray icon, with the intention of switch on/off the accessibility features, or start the Universal Access UI to configure those tools. But it is not still complete for the moment, and it seems more a mock-up. The screen reader switch doesn’t starts any screen reader, and although you can start the GNOME Shell magnifier you can’t configure it on the settings dialog (right now you need Orca to do that).

Conclusion

Most (all?) of the missing accessibility holes are detected, and a lot of them have provisional solutions. The process to manage the GNOME Shell with just Orca and the keyboard has started. In fact in my short-medium term plans was updating and finishing the current additional ATK support required on the keyboard navigation support, and create something like a branch or PPA and ask some user feedback (so new bugs). This will be postponed, due last changes on keyboard navigation and the overview-relayout.

But, most of then are provisional solutions, and it is still required to analyze them. Although it works, it is not clear if the current proposal to load the atk bridge, or start to add atk calls to the ui javascript code are the best options. Those should be also good solutions. Using Owen Taylor words “In my mind accessibility has to be held to the same high standards as everything else in GNOME”.

So as it was already stated, it would be really unlikely that GNOME Shell will ship with full accessibility support for GNOME 3.0.

BUT, as the roadmap states, accessibility is being taking into account. People are working on it, with several bugs, patches, and continuous patch review. So what now is a “rough draft”, at that moment will be a “real draft” of the required accessibility support on GNOME Shell, enough at least to start to ask users for feedback, and with a perspective of what it is missing, with the purpose of polish them for the following releases.

GNOME Accessibility Hackfest is over

Hackfest Logo
As you probably already know, most of the gnome a11y community were this week on Seville, on the GNOME Accessibility Hackfest, taking the most of the fact that some of them were already planning to assist the first international AEGIS conference.

Not too much blogging during this week, sorry. It was a really busy week. Assisting the conference or trying to hack something (I’m cooking some patches, imminent bug report).

Although it was a exhausting week (and somewhat a jinx week), I really think that it was worth the effort. Being face-to-face allow us to have meetings, coordinate ourselves, check the current status of GNOME accessibility and planning the priorities having in mind GNOME 3.0. And of course code. You can see in the following picture that Mario can work on WebKitGtk a11y in any place, and that he can’t stop in the middle of a task:

Mario Hacking

In the same way, the conference allow us to promote GNOME, and GNOME accessibility in particular, to a different public. Several AEGIS presentations were related to GNOME projects, some of them about specific GNOME modules and other about projects using GNOME, like the Guadalinfo project.

For any reason I received the award “Best accessibility project in the spirit of AEGIS”, due my work related to Cally and gnome-shell accessibility. In my opinion other people deserve it, like Joseph Scheuhammer, as his gnome-shell magnification solution is already integrated on gnome-shell, while my patches are still WIP. In fact, IMHO, this award should be granted to the GNOME accessibility community as a whole. Probably I fit more in the “spirit of AEGIS” because the most repeated word on AEGIS presentations were “prototype”. Who knows.
Aegis Award

I hope that all the presentations and the extra publicity of this award becomes a push to the interest to GNOME from the people attending the conference.

Finally I want to say “thank you” to a lot of people. I will try to not forgot anyone. If this is the case, please forgive me.

Thank you to the GNOME Foundation. Without his sponsorship most of the attendees would lost the hackfest. Thanks to the local team, mostly Emergya people, to coordinate several parts of the hackfest and behave like the perfect hosts during our time there. Special remark to Alejandro Leiva, as the in-theory common task of find and book the hotel was finally a little hell-in-earth. Thanks to Emergya and Yaco for sponsorship the BeerFest. Thanks to Joanmarie Diggs, as I’m always learning new accessibility things from the wisdom she irradiates.Thanks to all the people to make the effort to come to the hackfest.

And finally, thanks to the GNOME accessibility team for this really big and heavy present that they give me during the hackfest (if anyone doesn’t catch the joke, please check this video).

Cat herder

Wow, this was a really big post.