Re: Meeting Minutes Published - November 11, 2010
Not that I'm particularly well informed here, but having e-mail chatted with the NVDA guys about porting to Linux, and the Orca guys about porting to windows, and after reading a bit of the code and e-mail on the dev lists... Porting Orca to Windows or NVDA to Linux just isn't going to happen. It's not bad design on either team's part, it's just way too hard to do. That said, there's some sense in borrowing good ideas from each other. It turns out that both Orca and NVDA have an intense amount of code devoted to working around crud that doesn't work right in the rest of the system. Accessibility is low priority for almost all the apps the screen reader has to work with, so making it work with gum and tape in the screen reader is often how it's done. Bill On Fri, Dec 17, 2010 at 5:29 PM, Eitan Isaacson ei...@monotonous.org wrote: I agree with Cesar. In Orca's upcoming refactor a certain level of abstraction should be provided to allow porting to different platforms. Maybe in some kind of future NVDA and Orca could actually share a codebase or at least some modules. But this is just wishful thinking for now. This is how LSR was designed, it was abstracted (even to a fault), and kept a future Windows (or Mac) port feasible. Users win from a multi-platform screenreader because: 1. No learning curve when switching platforms. Today, besides learning a new platform, users need to learn different screen readers with very different interaction modes, NVDA or JAWS on Windows, VoiceOver on Mac, and Orca on Linux. It would be cool if a blind user did not have to worry about that. This would make the screen reader users competitive when it comes to technology proficiency since they are not locked down to the one AT they are used to on one platform. 2. It would do what Firefox did to the web to accessible computing. Desktop Linux is marketable today not because it reached feature parity with commercial offerings, but because it offers exactly the same web experience with Firefox or Chrome. The Free Desktop would be more attractive to blind users who are already familiar with it's great screen reader from the windows world. The screen reader and browser would be identical, and the experience would be too. Again, just wishful thinking :) On Fri, Dec 17, 2010 at 1:57 PM, Cesar Mauri ce...@crea-si.com wrote: El 17/12/2010 12:21, Piñeiro escribió: So now the problems. Take into account that turn Orca cross-platform is not just be able to compile Orca on Windows. There are more pieces that it would be required to turn: * at-spi: Orca is a screen reader that gets all the information from at-spi. So at-spi should also be migrated. * new bridge: right now, the communication path between the apps and at-spi is the ATK bridge or the QT bridge. Windows apps doesn't use ATK, AFAIK, it uses IAccessible2. So a new bridge should be required. So, turn cross-platform Orca means turns two modules, and create a new one. This is a really big amount of work to do. And we enter in a vicious circle. You proposed that turn in order to get funds. But we would require a really big amount of funds to get that. Thanks for your explanation which helped me to understand better how Orca works. Agreed. Given this scenario it seems clear that the effort is greater than the return. However, IMHO, I think that this approach could be taken into account for some new AT projects, especially those less dependant on specific api's (for instance, I'm thinking of AAC software). Beyond probably increasing funding opportunities (according to previous comments in this thread), a larger user base could be reached. Is just my opinion. César ___ gnome-accessibility-list mailing list gnome-accessibility-l...@gnome.org http://mail.gnome.org/mailman/listinfo/gnome-accessibility-list ___ gnome-accessibility-list mailing list gnome-accessibility-l...@gnome.org http://mail.gnome.org/mailman/listinfo/gnome-accessibility-list ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
Re: Meeting Minutes Published - November 11, 2010
On Fri, Dec 17, 2010 at 00:35, Juanjo Marin juanjomari...@yahoo.es wrote: On Thu, 2010-12-16 at 09:02 +0100, Tomeu Vizoso wrote: On Thu, Dec 16, 2010 at 08:28, Ben Konrath b...@bagu.org wrote: Hi Dave, On Wed, Dec 8, 2010 at 1:21 PM, Dave Neary dne...@gnome.org wrote: Based on the Orca (or even a more general a11y) roadmap, it may be possible to get some funding from companies or associations interested in seeing Orca get better (although a lot of the associations seem to be focussing more on NVDA because it works on Windows). Just thought I'd chime in here. I spent a bit of time searching for funding to work on Caribou after the ATRC / IDRC cut funding for my position. The feedback I received was a similar story; the potential funders seemed only interested in an applications that would serve their users who primarily use Windows. Obviously this will be an issue when searching for funding for GNOME a11y projects - especially new projects that don't have an established group of users like Caribou. What about those companies deploying linux desktops in masse in public organizations? From time to time appear in the news some big European city or government department that has migrated their desktops to Linux. Those vendors aren't asked to provide an accessible solution? Regards, Tomeu What about international disabilities associations like: - International Council for Education of People with Visual Impairment [1] - ONCE International I think we can use OLPC/Sugar/GNOME deployments in schools like a good argument for asking for this. Usually, national associations are very Windows-centric, but they can help to children in poor areas improving the GNOME a11y technologies and its translation to Sugar. I think Sugar people would agree with this (Maybe Tomeu or someone from Sugarlabs can help with this idea if we think is feasible) SugarLabs hasn't been successful at all with raising so far, but I think they would be happy to assist (I'm not that active there these days). Regards, Tomeu cheers, -- Juanjo Marin [1] http://www.icevi.org/ [2] http://www.once.es/new/Onceinternacional/0_pruebaonceint/index_html [3] http://wiki.laptop.org/go/Accessibility Cheers, Ben ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
Re: Meeting Minutes Published - November 11, 2010
Hi, Based on the Orca (or even a more general a11y) roadmap, it may be possible to get some funding from companies or associations interested in seeing Orca get better (although a lot of the associations seem to be focussing more on NVDA because it works on Windows). Just thought I'd chime in here. I spent a bit of time searching for funding to work on Caribou after the ATRC / IDRC cut funding for my position. The feedback I received was a similar story; the potential funders seemed only interested in an applications that would serve their users who primarily use Windows. Obviously this will be an issue when searching for funding for GNOME a11y projects - especially new projects that don't have an established group of users like Caribou. And what about turning Orca into cross-platform? I would attract funding from different sources. César ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
Re: Meeting Minutes Published - November 11, 2010
El 17/12/2010 12:21, Piñeiro escribió: So now the problems. Take into account that turn Orca cross-platform is not just be able to compile Orca on Windows. There are more pieces that it would be required to turn: * at-spi: Orca is a screen reader that gets all the information from at-spi. So at-spi should also be migrated. * new bridge: right now, the communication path between the apps and at-spi is the ATK bridge or the QT bridge. Windows apps doesn't use ATK, AFAIK, it uses IAccessible2. So a new bridge should be required. So, turn cross-platform Orca means turns two modules, and create a new one. This is a really big amount of work to do. And we enter in a vicious circle. You proposed that turn in order to get funds. But we would require a really big amount of funds to get that. Thanks for your explanation which helped me to understand better how Orca works. Agreed. Given this scenario it seems clear that the effort is greater than the return. However, IMHO, I think that this approach could be taken into account for some new AT projects, especially those less dependant on specific api's (for instance, I'm thinking of AAC software). Beyond probably increasing funding opportunities (according to previous comments in this thread), a larger user base could be reached. Is just my opinion. César ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
Re: Meeting Minutes Published - November 11, 2010
I agree with Cesar. In Orca's upcoming refactor a certain level of abstraction should be provided to allow porting to different platforms. Maybe in some kind of future NVDA and Orca could actually share a codebase or at least some modules. But this is just wishful thinking for now. This is how LSR was designed, it was abstracted (even to a fault), and kept a future Windows (or Mac) port feasible. Users win from a multi-platform screenreader because: 1. No learning curve when switching platforms. Today, besides learning a new platform, users need to learn different screen readers with very different interaction modes, NVDA or JAWS on Windows, VoiceOver on Mac, and Orca on Linux. It would be cool if a blind user did not have to worry about that. This would make the screen reader users competitive when it comes to technology proficiency since they are not locked down to the one AT they are used to on one platform. 2. It would do what Firefox did to the web to accessible computing. Desktop Linux is marketable today not because it reached feature parity with commercial offerings, but because it offers exactly the same web experience with Firefox or Chrome. The Free Desktop would be more attractive to blind users who are already familiar with it's great screen reader from the windows world. The screen reader and browser would be identical, and the experience would be too. Again, just wishful thinking :) On Fri, Dec 17, 2010 at 1:57 PM, Cesar Mauri ce...@crea-si.com wrote: El 17/12/2010 12:21, Piñeiro escribió: So now the problems. Take into account that turn Orca cross-platform is not just be able to compile Orca on Windows. There are more pieces that it would be required to turn: * at-spi: Orca is a screen reader that gets all the information from at-spi. So at-spi should also be migrated. * new bridge: right now, the communication path between the apps and at-spi is the ATK bridge or the QT bridge. Windows apps doesn't use ATK, AFAIK, it uses IAccessible2. So a new bridge should be required. So, turn cross-platform Orca means turns two modules, and create a new one. This is a really big amount of work to do. And we enter in a vicious circle. You proposed that turn in order to get funds. But we would require a really big amount of funds to get that. Thanks for your explanation which helped me to understand better how Orca works. Agreed. Given this scenario it seems clear that the effort is greater than the return. However, IMHO, I think that this approach could be taken into account for some new AT projects, especially those less dependant on specific api's (for instance, I'm thinking of AAC software). Beyond probably increasing funding opportunities (according to previous comments in this thread), a larger user base could be reached. Is just my opinion. César ___ gnome-accessibility-list mailing list gnome-accessibility-l...@gnome.org http://mail.gnome.org/mailman/listinfo/gnome-accessibility-list ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
Re: Meeting Minutes Published - November 11, 2010
Cesar, El 17/12/2010 12:21, Piñeiro escribió: ... So, turn cross-platform Orca means turns two modules, and create a new one. This is a really big amount of work to do. And we enter in a vicious circle. You proposed that turn in order to get funds. But we would require a really big amount of funds to get that. Thanks for your explanation which helped me to understand better how Orca works. Agreed. Given this scenario it seems clear that the effort is greater than the return. However, IMHO, I think that this approach could be taken into account for some new AT projects, especially those less dependant on specific api's (for instance, I'm thinking of AAC software). Beyond probably increasing funding opportunities (according to previous comments in this thread), a larger user base could be reached. Is just my opinion. This is already being done. Dasher is already cross-platform. The plug-ins to Open Office to generate DAISY books or print to braille are cross-platform (well, the braille part will be on other platforms soon). Etc. Having been involved in an attempt at making a cross-platform screen reader (we wanted to make the new outSPOKEN both for Windows and Mac - and I had an experimental port to SunOS going back in the day), I'm of the opinion that the overhead cost in abstracting the different approaches hooks such for the different platforms is likely not worth the cost - vs. just developing separate efforts which share ideas. Peter -- Peter Korn | Accessibility Principal Phone: +1 650 5069522 500 Oracle Parkway | Redwood City, CA 94065 Oracle is committed to developing practices and products that help protect the environment ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
Re: Meeting Minutes Published - November 11, 2010
On 18 December 2010 06:46, Peter Korn peter.k...@oracle.com wrote: Cesar, However, IMHO, I think that this approach could be taken into account for some new AT projects, especially those less dependant on specific api's (for instance, I'm thinking of AAC software). Beyond probably increasing funding opportunities (according to previous comments in this thread), a larger user base could be reached. Is just my opinion. This is already being done. Dasher is already cross-platform. The plug-ins to Open Office to generate DAISY books or print to braille are cross-platform (well, the braille part will be on other platforms soon). Etc. Having been involved in an attempt at making a cross-platform screen reader (we wanted to make the new outSPOKEN both for Windows and Mac - and I had an experimental port to SunOS going back in the day), I'm of the opinion that the overhead cost in abstracting the different approaches hooks such for the different platforms is likely not worth the cost - vs. just developing separate efforts which share ideas. Peter I kinda agree with both of you. Have platform agnostic AT will [potentially] increase the diversity of the community and so [potentially] improve sustainability. However it also adds considerable development complexity and support burden, especially when you consider we now need to include mobile/tablet platforms now such as Android and iOS. There is also project management complexity as each platform may pull in different directions due to variations in platform conventions and usages. With Maavis [1] I was careful to consider portability with GNOME in mind, even though the original requirement was Win32 only. Accordingly I used Mozilla XUL (actually Firefox) and llimited the contacts points with the platform by using VLC for media playing, XPCOM components for process launch/control and TTS (SAPI), plus an Outfox Python server is used for (video calling (Skype for win32) and switch (joystick) input. I'm fully aware that even with this care in architecture that supporting several platforms will be plenty of work. 1: maavis.fullmeasure.co.uk -- Steve Lee Full Measure - open source accessibility - http://fullmeasure.co.uk ___ foundation-list mailing list foundation-list@gnome.org http://mail.gnome.org/mailman/listinfo/foundation-list
