Browser improvements: Difference between revisions

From OLPC
Jump to navigation Jump to search
No edit summary
No edit summary
 
(8 intermediate revisions by 2 users not shown)
Line 1: Line 1:
For 2017, global support volunteers sought the best browser enhancements (tweaks, hacks, alternatives) to keep 1000's of XO-1, XO-1.5, XO-1.75 and XO-4 laptops most useful in schools that do not have budgets for new hardware. Browser wars, like most arms races, are not fully solvable in the end -- still we've learned a lot since we came together in 2007, so material progress may now be possible. Specifically:
{{draft}}
For 2017, global support volunteers seek the best browser enhancements (tweaks, hacks, alternatives) to keep 1000's of XO laptops most useful in schools that do not have budgets for new hardware. Browser wars, like most arms races, are not fully solvable in the end -- still we've learned a lot since we came together in 2007, so material progress may now be possible. Specifically:


A decade later compatibility questions are as much about fixed offline content (Khan Academy, KA Lite, Kiwix, RACHEL, XSCE / Internet-in-a-Box etc) rather than the online treadmill of constant upgrades. Glitzy, overweight broadband/cloud sites and their adverts are generally NOT the educational priorities of developing world schools, who regardless (even if it was pedagogically appropriate) cannot afford Internet costs nevermind streaming costs inherent to YouTube/etc (tragically even simple news articles > 1 MB in 2016, due to overweight adverts, for a fun aside see [https://manypossibilities.net/2015/11/zero-rating-a-modest-proposal/ Steve Song's counter-proposal] to provide 9.6kbit/s to all worldwide!)
A decade later compatibility questions are as much about fixed offline content (Khan Academy, KA Lite, Kiwix, RACHEL, XSCE / Internet-in-a-Box etc) rather than the online treadmill of constant upgrades. Glitzy, overweight broadband/cloud sites and their adverts are generally NOT the educational priorities of developing world schools, who regardless (even if it was pedagogically appropriate) cannot afford Internet costs nevermind streaming costs inherent to YouTube/etc (tragically even simple news articles > 1 MB in 2016, due to overweight adverts, for a fun aside see [https://manypossibilities.net/2015/11/zero-rating-a-modest-proposal/ Steve Song's counter-proposal] to provide 9.6kbit/s to all worldwide!)


So HTML 5.1 + Javascript is our moving target, that remains a noble goal despite: so long as we stay grounded in the reality that most all remote/offline deployments do software maintenance on an Annual Basis At Best. Even when serious security threats arise: even the best often forget that IT budgets and staffing are sadly unrealistic within most all high-poverty communities :/
So HTML 5.1 + Javascript was our moving target, that remains a noble goal despite: so long as we stayed grounded in the reality that most all remote/offline deployments do software maintenance on an Annual Basis At Best. Even when serious security threats arise: even the best often forget that IT budgets and staffing are sadly unrealistic within most all high-poverty communities :/


No question XO-1 laptops lack the horsepower of XO-1.5, XO-1.75 and XO-4. So XO-1 will be a secondary focus, whatever the outcome of our research. Particularly if we determine that [[Release_notes/13.2.8|OLPC OS 13.2.8]]'s browsers ([http://activities.sugarlabs.org/en-US/sugar/addons/versions/4024#version-157.3 Browse 157.3] in Sugar and [http://ftp.gnome.org/mirror/gnome.org/sources/epiphany/3.6/ Epiphany 3.6.1] in Gnome) already have made sufficient progress (e.g. for XO-1's more limited uses) in recent years!?
No question XO-1 laptops lack the horsepower of XO-1.5, XO-1.75 and XO-4. So XO-1 was a secondary focus, whatever the outcome of our research. Particularly as we determined that [[Release_notes/13.2.8|OLPC OS 13.2.8]]'s browsers ([http://activities.sugarlabs.org/en-US/sugar/addons/versions/4024#version-157.3 Browse 157.3] in Sugar and [http://ftp.gnome.org/mirror/gnome.org/sources/epiphany/3.6/ Epiphany 3.6.1] in Gnome) already have made sufficient progress (e.g. for XO-1's more limited uses) in recent years!?


Finally, diverging from Fedora 18 (released Jan 15th 2013, underlying 13.2.8) is also not our goal, despite its old age (Happy 4th Birthday, thanks to all who've kept everything together along the way!) Still, radical transformative solutions will be considered IF proven, e.g. certain high schools explicitly do not want Sugar, and many ask us to look into Firefox, Chrome (etc) on Gnome if not other kernels / desktops / platforms / OS's as appropriate?
Finally, diverging from Fedora 18 (released Jan 15th 2013, underlying 13.2.8) was also not our goal, despite its old age (Happy 4th Birthday, thanks to all who've kept everything together along the way!) Still, radical transformative solutions were considered IF proven, e.g. certain high schools explicitly do not want Sugar, and many ask us to look into Firefox, Chrome (etc) on Gnome if not other kernels / desktops / platforms / OS's as appropriate?


''Questions: beyond HTML5/Javascript industry norms, does it help everyone if we draw up a testing matrix right here, showing (1) which most common codecs, (2) which lightweight encodings-rates, and (3) which common content packs... work well (or tolerably well) on which platforms? An example content pack being the very popular [https://twitter.com/PhETsims PhET science simulations], a great educational offering, that is wasted when its browser/Javascript compatibility is hit-or-miss.''
''Questions: beyond HTML5/Javascript industry norms, did it help everyone if we drew up a testing matrix right here, showing (1) which most common codecs, (2) which lightweight encodings-rates, and (3) which common content packs... work well (or tolerably well) on which platforms? An example content pack being the very popular [https://twitter.com/PhETsims PhET science simulations], a great educational offering, that is wasted when its browser/Javascript compatibility is hit-or-miss.''


''Perhaps Most Important: can we communicate concisely/explicitly best browser & video "what works" guidelines, so that hard-working free content aggregators/publishers (many of whom volunteers) understand their tight packaging requirements across old Androids, old XO's, etc? Hence greatly enhancing the XO's practical purposes during its remaining years~''
''Perhaps Most Important: could we communicate concisely/explicitly best browser & video "what works" guidelines, so that hard-working free content aggregators/publishers (many of whom are volunteers) understand their tight packaging requirements across old Androids, old XO's, etc? Hence greatly enhancing the XO's practical purposes during its remaining years~''


In Conclusion: here are some people who have developed great expertise with modern browser/codec appropriateness for XO laptops, that I (Adam Holt) hope to call to gather all the best ideas here on this page, helping us all organize our legacy for 2017 and beyond:
In Conclusion: here are some people who have developed great expertise with modern browser/codec appropriateness for XO laptops, that I (Adam Holt) hoped to call to gather all the best ideas here on this page, helping us all organize our legacy for 2017 and beyond:


* James Cameron - [[Talk:Browser_improvements|suggests]] a "single representative test video which is transcoded into different picture sizes, frame rates, codecs, bit rates, codecs, container formats, and delivery network protocols. A smart proxy may be imagined that will transcode on the fly"
* James Cameron - [[Talk:Browser_improvements|suggests]] a "single representative test video which is transcoded into different picture sizes, frame rates, codecs, bit rates, codecs, container formats, and delivery network protocols. A smart proxy may be imagined that will transcode on the fly"
* Gonzalo Odiard
* Gonzalo Odiard
* Tony Anderson - many experiences around the world; I hope to talk to him by/around early January. Recommends measuring params of XO-1 ogg video streams, easily playable on all other XO's. Recommends avoiding all HD video and similar. Recommends downloads NOT streaming. Recommends a patch allowing users to experience larger {ebooks/EPUB/PDF, MP3 radio shows, MP4, txt} offline, e.g. larger media download to Journal by default in Sugar, or to download folder in Gnome (whereas very short media play directly in browser?)
* Tony Anderson - many experiences around the world; I hoped to talk to him by/around early January. Recommends measuring params of XO-1 ogg video streams, easily playable on all other XO's. Recommends avoiding all HD video and broadband pollution. Recommends A/V downloads NOT streaming. Recommends a [https://groups.google.com/d/msg/unleashkids/1yXqw5h-BqU/Qak0FJMtEAAJ patch allowing users to experience larger {ebooks/EPUB/PDF, MP3 radio shows, MP4, txt} offline], e.g. larger media download to Journal by default in Sugar, or to a download folder in Gnome (whereas very short media play directly in browser?)
* Terry Gillett - many experiences in SE Asia; will talk by/around early January
* Terry Gillett - many experiences in SE Asia; hoped to talk by/around early January; asked if we/all might [http://lists.laptop.org/pipermail/devel/2017-January/thread.html#39129 consider Firefox ~50] instead of Epiphany/Browse?
* Nathan Riddle - refreshingly has gathered experiences in US schools as well; hope to talk by/around early January
* Nathan Riddle - refreshingly had gathered experiences in US schools as well; hoped to talk by/around early January
* Jerry Vonau
* Jerry Vonau
* Tim Moody
* Tim Moody
* George Hunt - suggests Firefox on Gnome on XO-1.5 (does this also work on XO-1.75 and XO-4? is a Sugar icon possible?) and warns us that [http://www.pcworld.com/article/3010404/browsers/googles-killing-chrome-support-for-32-bit-linux-ubuntu-1204-and-debian-7.html Chrome is no longer supported on 32-bit OS's]
* George Hunt - suggests Firefox on Gnome on XO-1.5 (does this also work on XO-1.75 and XO-4? is a Sugar icon possible?) and warns us that [http://www.pcworld.com/article/3010404/browsers/googles-killing-chrome-support-for-32-bit-linux-ubuntu-1204-and-debian-7.html Chrome was no longer supported on 32-bit OS's]
* Kevin Gordon
* Kevin Gordon
* Nick Doiron
* Nick Doiron
* Alex Perez - Chrome is a voracious consumer of RAM, even if it did work on XO-1.5+ (but is Firefox that much better?)
* Alex Perez - Chrome will be a voracious consumer of RAM, even if it did work on XO-1.5+ (but is Firefox that much better?)
* ?


Thanks! Discussions most welcome on the [https://groups.google.com/group/unleashkids unleashkids mailing list] and [http://lists.laptop.org/listinfo/devel devel@laptop.org]. Aside/Clarification regarding proprietary codecs and Adobe Flash: thankfully there is much less demand for Flash in 2017 as compared to 2007, but still such schools exist, and real-world workarounds are welcome within this grassroots exchange -- even if such mods can't all be legally mainlined/published, by definition!
Thanks! Discussions most welcome on the [https://groups.google.com/group/unleashkids unleashkids mailing list] and [http://lists.laptop.org/listinfo/devel devel@laptop.org]. Aside/Clarification regarding proprietary codecs and Adobe Flash: thankfully there is much less demand for Flash in 2017 as compared to 2007, but still such schools exist, and real-world workarounds are welcome within this grassroots exchange -- even if such mods can't all be legally mainlined/published, by definition!

Latest revision as of 20:13, 30 March 2017

For 2017, global support volunteers sought the best browser enhancements (tweaks, hacks, alternatives) to keep 1000's of XO-1, XO-1.5, XO-1.75 and XO-4 laptops most useful in schools that do not have budgets for new hardware. Browser wars, like most arms races, are not fully solvable in the end -- still we've learned a lot since we came together in 2007, so material progress may now be possible. Specifically:

A decade later compatibility questions are as much about fixed offline content (Khan Academy, KA Lite, Kiwix, RACHEL, XSCE / Internet-in-a-Box etc) rather than the online treadmill of constant upgrades. Glitzy, overweight broadband/cloud sites and their adverts are generally NOT the educational priorities of developing world schools, who regardless (even if it was pedagogically appropriate) cannot afford Internet costs nevermind streaming costs inherent to YouTube/etc (tragically even simple news articles > 1 MB in 2016, due to overweight adverts, for a fun aside see Steve Song's counter-proposal to provide 9.6kbit/s to all worldwide!)

So HTML 5.1 + Javascript was our moving target, that remains a noble goal despite: so long as we stayed grounded in the reality that most all remote/offline deployments do software maintenance on an Annual Basis At Best. Even when serious security threats arise: even the best often forget that IT budgets and staffing are sadly unrealistic within most all high-poverty communities :/

No question XO-1 laptops lack the horsepower of XO-1.5, XO-1.75 and XO-4. So XO-1 was a secondary focus, whatever the outcome of our research. Particularly as we determined that OLPC OS 13.2.8's browsers (Browse 157.3 in Sugar and Epiphany 3.6.1 in Gnome) already have made sufficient progress (e.g. for XO-1's more limited uses) in recent years!?

Finally, diverging from Fedora 18 (released Jan 15th 2013, underlying 13.2.8) was also not our goal, despite its old age (Happy 4th Birthday, thanks to all who've kept everything together along the way!) Still, radical transformative solutions were considered IF proven, e.g. certain high schools explicitly do not want Sugar, and many ask us to look into Firefox, Chrome (etc) on Gnome if not other kernels / desktops / platforms / OS's as appropriate?

Questions: beyond HTML5/Javascript industry norms, did it help everyone if we drew up a testing matrix right here, showing (1) which most common codecs, (2) which lightweight encodings-rates, and (3) which common content packs... work well (or tolerably well) on which platforms? An example content pack being the very popular PhET science simulations, a great educational offering, that is wasted when its browser/Javascript compatibility is hit-or-miss.

Perhaps Most Important: could we communicate concisely/explicitly best browser & video "what works" guidelines, so that hard-working free content aggregators/publishers (many of whom are volunteers) understand their tight packaging requirements across old Androids, old XO's, etc? Hence greatly enhancing the XO's practical purposes during its remaining years~

In Conclusion: here are some people who have developed great expertise with modern browser/codec appropriateness for XO laptops, that I (Adam Holt) hoped to call to gather all the best ideas here on this page, helping us all organize our legacy for 2017 and beyond:

  • James Cameron - suggests a "single representative test video which is transcoded into different picture sizes, frame rates, codecs, bit rates, codecs, container formats, and delivery network protocols. A smart proxy may be imagined that will transcode on the fly"
  • Gonzalo Odiard
  • Tony Anderson - many experiences around the world; I hoped to talk to him by/around early January. Recommends measuring params of XO-1 ogg video streams, easily playable on all other XO's. Recommends avoiding all HD video and broadband pollution. Recommends A/V downloads NOT streaming. Recommends a patch allowing users to experience larger {ebooks/EPUB/PDF, MP3 radio shows, MP4, txt} offline, e.g. larger media download to Journal by default in Sugar, or to a download folder in Gnome (whereas very short media play directly in browser?)
  • Terry Gillett - many experiences in SE Asia; hoped to talk by/around early January; asked if we/all might consider Firefox ~50 instead of Epiphany/Browse?
  • Nathan Riddle - refreshingly had gathered experiences in US schools as well; hoped to talk by/around early January
  • Jerry Vonau
  • Tim Moody
  • George Hunt - suggests Firefox on Gnome on XO-1.5 (does this also work on XO-1.75 and XO-4? is a Sugar icon possible?) and warns us that Chrome was no longer supported on 32-bit OS's
  • Kevin Gordon
  • Nick Doiron
  • Alex Perez - Chrome will be a voracious consumer of RAM, even if it did work on XO-1.5+ (but is Firefox that much better?)

Thanks! Discussions most welcome on the unleashkids mailing list and devel@laptop.org. Aside/Clarification regarding proprietary codecs and Adobe Flash: thankfully there is much less demand for Flash in 2017 as compared to 2007, but still such schools exist, and real-world workarounds are welcome within this grassroots exchange -- even if such mods can't all be legally mainlined/published, by definition!