Tag Archives: GNOME

Tracker RDF database performance

I have spent time on-and-off the past week looking at the performance of the Tracker RDF database. Tracker, I believe, started out as a desktop search tool for Gnome. I never used it in this incarnation, and it has only come to my attention  since version 0.7, when the developers implemented a general purpose RDF storage engine at its core. I wanted to know how this newly implemented RDF database compared to a widely used RDF database in terms of query performance. In essence I was interested in whether the Tracker project had spawned something that could compete with Virtuoso and 4Store.

You can find my complete results and analysis at the tracker mailing list archive, but the headline statement is that tracker had roughly 9 times the query performance of Virtuoso. The graph here shows the breakdown by query.

This is a drastic difference in performance that greatly favours the home-grown database utilized by Tracker. However this stellar performance comes at the cost of flexibility. Its obvious that the database has been tailored very much to the needs of Tracker itself. Unlike Virtuoso it is not ‘schema-free’. A description of the data (In the form of something called an RDF ontology) is required for storage. In addition to this, the data formats are more restrictive, and some common elements of RDF are missing.

My general impression was that Tracker has great query performance, especially considering a tiny memory footprint. Unfortunately it is not suited to storage of pre-existing RDF data sets, such as those generated for semantic-web applications. This could well change in the future. Tracker, and its RDF database, are in heavy development. They already have speed and seemingly stability in the code-base. It might soon be time to add the new features that make it more generally applicable.

I should add that when I started this work I was heavily sceptical. Codethink have been highly involved in RDF, but I have not joined in. I have learned a-lot in the past few weeks, and this has made me more positive. I still believe that RDF might be too flexible for its own good, and I’ve found that the ontologies are onerous, complicated, and not very well specified. I did however come across a great post which explains some of the advantages of RDF over other data models; SPARQL is far more intuitive than its SQL cousin. If used to its potential, with highly interlinked data, I think it may be possible for the benefits of RDF to outweigh the tough learning curve.

Funding Gnome a11y

As many of you may have heard, from blogs by Eitan, Mike and Joanie, as well as an e-mail to the gnome-foundation-lists by Fernando, the Gnome a11y community is having a tough time.

I have been interacting with the a11y community for over two years now, and in that time the funding situation has never looked good. I do not wish to insult or demean companies that are no-longer involved in funding Gnome a11y. Companies and individuals have their own priorities that they must follow. Work they have done in the past on Gnome is very much appreciated by me, even if they cannot continue that work in-to the future.

That said, I believe that in the past two and a half years Gnome a11y has lost a huge amount of funding. First from IBM, which, to many peoples dismay, pulled out of a11y funding before I started work on AT-SPI. I was glad to hear that Mozilla is providing $10,000 to the Gnome foundation for a11y work. I’m extremely grateful for that, but I do not believe that Mozilla are providing the level of funding that they have done in the past. Our work on AT-SPI D-Bus has been funded jointly by Codethink, Sun, and another un-named benefactor. None of this funding is likely to continue past the end of February. All of this would seem slight were it not for the news that Oracle have let-go of important Gnome a11y community members working for the Sun Accessibility Project Office. Sun have been the major contributor to Gnome a11y, and this is a worrying signal that Oracle do not intend to continue the current level of contribution.

Assuming that Oracle do not wish to involve themselves in Gnome a11y, my back-of-the-envelope calculations indicate that we may have lost greater than $200,000 in anual funding over the last three years.

Although huge amounts of Gnome development takes place un-funded, by hackers, volunteers, users and hobbyists you would probably be surprised how much is done by folks working a 9-5. I don’t expect the figures to be the same, but as an example, 75% of kernel developers are paid by corporations for their work. The loss of the Sun Accessibility Project Office and other sources of funding will be felt very heavily by the Gnome a11y community.

Accessibility is incredibly important to the Gnome project, and not only to its users. Gnome has a fantastic, credible, accessibility story. This, to me, marks Gnome out as a class ‘A’ open-source project. Were we to lose this, it would be a turning point. In my eyes Gnome would then be a project in decline.

What can we do?

Firstly we need to go on a cohesive search for funding. The Linux Foundation has an accessibility group that I have been involved in for a long time. This seems to me the best place to combine our efforts in the great funding drive. Funding channeled through the Linux Foundation would not be Gnome specific, but cross desktop a11y technology is what we have long been striving for.

Ideally enough funding would be found to hire someone to work full time on Linux Desktop accessibility.

Outside of the search for cash all Gnome developers need to spend more time on accessibility. Its not always easy to make ones application accessible, and I’m sure it can seem daunting. There are still a11y community members ready to help out though. All is not lost. :) I’m damn near certain that we are going to pull together. Gnome 3.0 will have the same great accessibility that has made me proud of past Gnome releases.

AT-SPI2 First release

Its been a long time coming, but about a fortnight ago we did the first release of AT-SPI2, the D-Bus port of our Linux accessibility framework. As mentioned in the notes the release had a warning label.

‘DANGER: Development release – may cause psychological harm’

Overall though I’ve been very pleased with the work we’ve achieved so-far. I have not been funded to work on the project for six months now, but I’ve found time to fix a few bugs, and most importantly help out those who want to get involved. I never really believed Linus “Release early, release often”, but it turns out that he  was right. Since the software has had an official release we’ve had a good number of people providing patches and generally poking at it.

Thanks go to Mario Lang, Stephen Shaw, Halim Sahin and others who have tried things out. Its been extremely useful. Mario especially finding a couple of serious bugs.

We still have a long way to go, but the end has been in sight ever since the hack-fest in Dublin. We need to get the cspi port up-to-date, help Mike Gorse work through the issues regarding reference count removal and fix bugs, lots of bugs. With EDS quickly shifting to D-Bus Gnome is well on the way to ORBit removal. A11y and GConf shouldn’t be far behind.

File Organization & Gnome 3.0

Thank-you to Felix Kaser for pointing out Mark Shuttleworth’s interview with derStandard.at. It was an incredibly interesting read. While at the GCDS this year I made many guesses about the Ubuntu opinion on Gnome Shell and the plans for 3.0. They, as the largest distributor of a consumer linux desktop, have the most to lose or gain from the awesome 3.0 plans.

The part of the interview that struck me directly was Mark’s opinion on what is missing from our proposals:

Well initially there was a lot of discussions about something that was much less visual which is how files are organized and I even blogged about it. I think actually that could be a bigger  improvement in the every-day user experience of the GNOME desktop”

I don’t know about Mark, but I was surrounded by people who are desperate to solve this type of problem at GCDS. Rob Taylor & Philip Van Hoof were both present and really pushing tracker as a a usable, fast data-store for desktop metadata. I attended Thursdays ontology BoF with the two of them. Mark is right, this is generally an unsexy problem. People who missed the ‘Nature of e-mail containers’ conversation on Thursday afternoon avoided the GCDS nadir of boredom.

The real problems of how we present this to our users is still to come, and its much more difficult than providing a fast data-store or getting consensus on what the meta-data should look like. Still, its the sexy, exciting part where we should be able to get everyone involved. Mark may be surprised by how many at Codethink and elsewhere in the community are working on replacing the awful file-system metaphor for data organization. Obviously we haven’t done enough to get his attention yet, but that could change soon. (Small secret) Codethink should be putting resources into a demo app over the coming months for the purpose of showing off some cool new technologies, including the new Tracker. Wait for a blog post from Rob Taylor for the full details.

Accessibility 3.0

Last week I attended a week-long meeting at the Sun offices in Dublin. The purpose of the trip was to hammer out issues with the D-Bus AT-SPI port, so that everyone can feel more confident about its readiness for Gnome 3.0.

What a great week we ended up having. I should first say thanks to all that attended, Willie Walker, Mike Gorse, Li Yuan, Ke Wang, Brad Taylor, and Rob Taylor. Sun, Novell & Codethink also need thanking. They paid for everyone to be there, some flying from China, with no foundation money involved. The Gnome community is extremely lucky to have Willie devoting his time to accessibility. I dread to think what Gnome a11y would be like without his organization and direction

Willie organized the event, I imagine more out of fear at what he found in the D-Bus AT-SPI code-base than anything else. All told I think everyone was pleasantly surprised. The project will take a-lot of work to get to the level of maturity required, but loads has already been done. What is required of us for Gnome 3.0 isn’t our of our reach.

We started the week with a very brief overview of the design for AT-SPI D-Bus. Only Mike and I had much experience with the project and we wanted to get everyone up to speed.  This very quickly turned into a Q&A session and we figured out what the main issues were on Monday afternoon. Tuesday was spent on the important task of checking the D-Bus protocol specification against both the code and the IDL spec. This bought up a number of bugs, issues and improvements. It was well worthwhile. Amazingly Wednesday & Thursday were spent hacking by everyone, which is a great feat considering that four of us had only peeked at the code previously. So much got done. Orca is now running with a fair degree of success and we can move on to performance issues and application-specific bugs. For a full run-down of the still-to-dos and what-got-dones take a look at the weeks wiki.


Outside of work was great fun also. Dublin really is a beautiful city. Lots of green spaces nice buildings and extremely friendly people. The business park is pretty swanky also, and luckily the Oracle buildings are across the road from Sun. We went out to eat and drink in Temple bar, which apparently isn’t a bar. Wednesday involved a trip to the very very small Novell offices in Dublin. They are a way out of town, Brad and I were lucky to get there at all given the distinct lack of signs and road-names that have changed since Google scraped them. We met up with Alan McGovern, a Moonlight hacker, and along with a meal and a few drinks geeked the night away.  Alberto Ruiz took us all out for a nice meal on Thursday night. I forget where it was. You’ll have to ask him for the recommendation. Later we attended the Sun Pub Quiz. Luckily for me, our arrival was just as the Quiz part was ending. I’m not so good with the questions.

Me for 3.0

What am I doing for 3.0 I hear you ask? I’m sure you didn’t but I’m going to blurt it out anyway. I’ll be finding time any-where I can to work on D-Bus accessibility. Codethink has already dedicated lots of my time to the project, and there may be a little more in the future. The accessibility plan looks a little sketchy in the road-map, but I believe that with some hard work it can become a big success for the big release.


After ranting about my work I’d love to hear what everyone else us up-to. There must me more than a few semi-secret 3.0 charges taking place.

‘Do-ifying’ Gtk (and other applications)

Many thanks to racarr for posting about this earlier today.  I had seen posts on the ‘Do-ifying’ of GTK but they had passed me by. ‘Do-ifying’ applications is a fantastic idea. We shouldn’t limit ourselves to GTK apps though.

The bit that caught my interest was that racarr was thinking of exposing a UI heirarchy over D-Bus to allow Gnome Do to find key-bindings and make them available. I suggested that AT-SPI would be a better way to do this, as all key-bindings are already exposed. I come to no conclusions over whether A11y is the best way. I’m biased. :) It did get me thinking about how A11y technologies could be used to provide general application assistance with Gnome Do.

Context sensitive ‘Do’ commands

Orca, the gnome screen reader, uses a11y mainly by responding to ‘focus’ events which are emitted when input focus moves to a new widget. Gnome Do could use ‘focus’ events to provide context-sensitive commands. By knowing which application, and which part of an application the user is currently focused on it might be possible to provide a more tailored selection of commands.

Perhaps the choice of ‘Do’ commands for each application would depend on finding out what D-Bus interfaces it supports.

Application generated ‘Do’ commands

Obviously for application generated commands we need methods of exposing these to Gnome Do. My preference would be to do this via D-Bus, which again a11y could help with. All ATs (Assistive technologies) need to know when new accessible applications appear on the desktop. Towards this end, D-Bus AT-SPI has a daemon that acts as an application registry and informs ATs when new applications are added. This registry could be made more generic so that Gnome Do would listen for new applications and register their ‘Do’ commands when they are started.

I can’t describe the registration interface here. I don’t know Gnome Do. I’m imagining a list of actions with a command name and descriptions, but it could well be much much more complicated.

Key Bindings

As racarr suggests, it should be possible to inspect the applications UI hierarchy to find all the key bindings and present them and their descriptions to Gnome Do. GOK, the gnome on screen keyboard, already does something VERY similar to this by inspecting an application and presenting all the actions of the currently focused window in a simple manner. There may be other things we can infer from the UI hierarchy, but I wouldn’t want to get too ambitious here.

I’ve often thought that a11y technology was under utilized. As applications on our desktop are already merging together a little by providing and accessing more D-Bus services, a11y seems like it could be well placed to enable some really innovate interfaces for the general user. If this happens it could really help experiences for accessibility users too. More bugs fixed, more descriptions added to widgets with key-bindings. :)

Gnome, KDE & Mono A11y

Congrats Mono A11y

Many congratulations to Mono Accessibility, team for getting their first release out into the open.  The Mono A11y team must be one of the largest open-source A11y groups out there and I’m really excited about the work they are doing. WinForms and Moonlight are not yet my thing, but if Silverlight takes off the UIA provider they have created will undoubtedly form an essential part of Linux accessibility.

I don’t believe that it will be in the first release, but I’m really keen to see work start on the UIA client library for Mono. C# and Mono sound like a great place for developing new ATs.

AT-SPI D-Bus on freedesktop.org

For people who don’t know about Gnome accessibility or AT-SPI D-Bus:

AT-SPI D-Bus is a project which aims to use D-Bus instead of ORBit/CORBA as the IPC mechanism for Linux accessibility. For anyone interested in finding out about the Gnome accessibility architecture the developers page has some good information. Oddly enough KDE has a very good Gnome A11y overview, and Sun has a good diagram. Long story short the AT-SPI D-Bus aims to write a new, D-Bus based adapter for ATK, a registry daemon, and client libraries that are API compatible with the existing cspi and pyatspi.

The project has a new home on the freedesktop.org servers.

The code-base exists at:  git://anongit.freedesktop.org/git/at-spi2/at-spi2-core.git.

We are keeping a page on the linux-foundation wiki updated with all our progress. Unfortunately I’d say that the code is not yet ready for a first release. For reasons soon evident the code isn’t currently getting the love it deserves. (Help MUCH appreciated)

The reason we chose freedesktop.org and the Linux Foundation instead of Gnome hosting is that we wanted to emphasize the cross-desktop possibilities of a D-Bus based accessibility architecture.

Gnome, KDE & Mono: How it all fits together

The Mono A11y architecture diagram is missing something important that the AT-SPI D-Bus project can add – QT accessibility.

The drive to D-Bus accessibility came from ORBit deprecation, the embedded community and an ideal of cross-desktop accessibility. Its the last motive that has me most excited right now. QT currently has a D-Bus framework based heavily off AT-SPI, but unfortunately it has never been taken far enough to be compatible with existing AT-SPI ATs. The reason that the ATK, cspi and pyatspi libraries are not getting my attention right now is that I really want to get started on bringing QT into the mix.

A QT adapter for AT-SPI D-Bus will certainly round-out the Accessibility infrastructure on Linux. Not being involved in the KDE community I don’t have much say on how they do A11y, but I hope to make it as easy as possible for them to choose AT-SPI D-Bus. Along with the Mono work this could mean that QT, GTK, ATK, Winforms & Swing apps are accessible, using the same ATs, in both KDE and Gnome. I think that would be a fantastic achievement. If we work hard enough accessibility could be one of the big success stories of a joint Akademy/Guadec next year.

D-Bus Accessibility – Its alive(ish)

This may not look like much more than another screen-shot of Accerciser. You’ll have to trust me that is Accerciser inspecting firefox-2 over D-Bus. Its the first visible step in a long project aiming to replace ORBit with D-Bus for the AT-SPI accessibility protocol.


Mike Gorse and I have been hacking on this for a little over four months now, and its been pretty slow going. I’m very happy to finally have something to show for it.

So much to do…

What we have however isn’t quite the finished article. One of the main goals of the project is to help bring together Gnome and KDE accessibility. I believe that QT already has an ATK implementation, but I think it would make more sense for the QT accessibility interfaces to meet directly with the AT-SPI D-Bus protocol, a project that has not-yet begun.

On the ATK-Bridge / Pyatspi side there is still a huge amount of testing and bug fixing to do, along with a fair few missing features.

So much to test…

Its testing thats really occupying my thoughts at the moment. A while back, on a linux-foundation conference call, I remeber a conversation about how to test the pyatspi and cspi interfaces.

The conclusions were:

  • Have lots of small applications, each exposing only a few accessible objects and concentrating on a single accessible interface, such as the Component or Image interface.
  • Have a unit test written for cspi or pyatspi that is partnered with one of the mini applications, and inspects it over the AT-SPI protocol.

I have made a poor attempt to unit test pyatspi in this way, but I found it was a-lot of work to get any coverage. I think this is down to my choice of implementing a dummy ATK framework to create the applications. Mini GTK applications would have done just as well.

The advantage of lots of these mini applications and unit tests is that when someone comes to implement the AT-SPI protocol with a new widget set, or with new client side bindings they have something simple to validate against.

Where to go

The repository for the D-Bus AT-SPI work is still on the Codethink servers. It won’t be there for long though, there are plans for the project to be hosted at freedesktop.org soon.

A week in the life of a braindead hacker

Help needed

There was some discussion this week on desktop-devel about turning ‘Assistive technologies’ on by default in Gnome. See Willie Walkers original e-mailRob Taylor and I had discussed this previously, and at the time I didn’t have much of an opinion either way. After reading Willie’s e-mail though I was firmly in favor.

However Willie stumbled across a possible performance problem when accessibility is turned on. Using gtk-demo to test memory usage ‘top’ shows that accessibility can increase the ‘DATA’ memory section by 15mb, or over double the memory consumption. This was pretty shocking to me, so I tried to use massif to find the cause.

Thing is, massif reports that the memory allocation is not greatly increased by turning accessibility on.  The maximum stack size is not increased either. The ‘DATA’ field of ‘top’ apparently reports the DRS, or data resident set. The number of pages used by both the stack and heap. Rob suggested that fragmentation could be the problem, but in general I’m pretty stumped as to what is causing the discrepancy.

Help and suggestions greatly appreciated :)

Conduit conversations

We have John Carr in the office now. (Infact he’s been there more than Rob and I this week.) This means lots of conduit discussions. I was treated to a crash course on Conduit internals that left my brain feeling like a squeezed sponge, wrung cloth, and juiced orange all at once. This syncing thing is terribly hard and John has some difficult use cases that he wants to make work. (Tranferring contacts from his laptop go to GMail, and from GMail to the N810 and from that to his laptop. Whew.) Good luck to him. I think its the ‘Brain the size of texas’ award if he manages it.

The rest of the week…

Rob probably can’t blog to tell you this because he is gibbering somewhere in the corner muttering “Reference, Reference, where art thou?”. He’s been debugging lifecycle issues for a fortnight now. Its left us vowing never to touch ‘C’ again whenever humanly possible. (Which is not very often).

I have been working hard on the D-Bus version of pyatspi. Its going much slower than hoped, but i’ll get there in the end. I have a holiday coming this week, a trip to Finland for a friends wedding. I’m really looking forward to it. Were going to have a week of saunas in Lapland.

Memories of GUADEC & Istanbul

I had the pleasure of attending my first GUADEC last week. I doubt I’ll have a more enjoyable, tiring, embarrassing, or geek-filled time for the rest of the year.

I started the week locked up in my hotel room. Rob Taylor, John Carr and I were desperately trying to finish a demo that did not make it. Our trips out until Tuesday were mostly visiting restaurants, which would have been enough for me. Lamb Kebab and rice is just about my perfect meal, and during my entire stay in Istanbul I ate like a very happy king.

Once I got myself set up at the conference things got even better, there were some great talks. ‘UXD? WTF!‘ by Leisa Reichelt, ‘Introduction to the desktop data model‘ by Owen Taylor and ‘Syncronization and the GNOME deskop‘ by John Stowers all had me thinking.  It was a real shame that more people didn’t turn up to ‘Accessibility happenings in Novell‘, although it was the very last talk of the conference. Rob and I did a talk, ‘Wizbit a distributed revisioning file system‘, on Thursday morning. It was well attended, and I had a good time speaking.

The whole thing was smaller, and far friendlier, than I had expected. I was fortunate enough to meet Stormy just before, and soon after she had been announced as the GNOME executive director. She was just as lovely to me both times, so I assume power didn’t corrupt too much. :) Thanks to everyone I met for creating a thousand interesting conversations.

I can’t properly express my thanks to the people in Istanbul. They were incredibly helpful, patient and friendly. Even the ‘flexible prices’ that everyone seemed to experience only got bad when it came to taxis and drinks.  Thanks to Ali Sukru Goksu for his organization, and for showing us a great night out on Thursday. The conduit team and I went to an electronica festival on Friday night, which we never would have made it to without being looked after. Thanks Tuçe, Thanks Elvin.

The party prize and the most ‘flexible’ price prize go to John Stowers, for his 5-in-a-row 6am parties and for a certain vodka red bull.

I’ve come back really enthused about GNOME, about the people and the technology. Hopefully I’ll be able to put that enthusiasm to good use over the next year. See you all in Gran Canaria.