Walter leaving and shift to XP.

Walter Bender walter.bender at gmail.com
Tue Apr 22 19:48:47 EDT 2008


First of all, just to clear, Flash does run on the laptop: there is a
choice of both the Adobe Flash player and the FOSS Flash player,
Gnash. We opted to install the Gnash player by default. Many of the
problems people have with Flash are actually related to codecs rather
than the player itself. We don't load proprietary codecs onto the
machine by default, but they are available for download and some of
our deployments in fact do opt to load some proprietary codecs--after
of course obtaining the proper licenses. I see this approach as a
reasonable compromise given the goals of the project. Apparently
others see this as fundamentalism?

Second, regarding Microsoft, I agree that if it is to be an open
platform, it should be open to everyone, including Microsoft. That
said, it is somewhat revisionist to suggest that the SD card was added
on behalf of Microsoft: it was added at the same time as the camera
because we had the opportunity while adding an ASIC necessary to
improve NAND Flash performance. The fact that it facilitates the
running of Windows was not the consideration at the time. I am not
aware of any current effort to port Sugar to Windows; I don't know
enough about Windows to know how much effort that would entail or even
if it is possible.

Third, in regard to the performance, feature sets, etc., the OLPC
software stack is immature--quite naturally, as it is a relatively new
product and project. The software development roadmap for the project
had included a phased approach where we first get a core feature set
built; do some initial triage of bugs and bring some stability to the
deployments; and then work to fine-tune performance. While have heard
a lot of noise about performance in the media and from some members of
the development community, it has not, in my experience been a major
road-block in the school trials and deployments. There are lots of
bugs and lots of things that could be improved upon, and these should
certainly be addressed, but the characterizations being made in this
thread do not reflect the realities of the OLPC deployments--the
children and teachers are using the laptops and are learning.

Fourth and final point for the moment: it is important to make a
distinction between the system software--drivers, power management,
memory management, etc. and the Sugar user experience. It is not yet
easy to always draw a clear line between them, but many of the
performance problems* are not related to the choices we made regarding
the UI, although, since the UI is how one experiences the laptop, they
are felt there. I am not suggesting that there isn't room for
improvement, but the call for dropping Sugar is not going to make as
dramatic a difference in performance as is being suggested. And at
what cost? Is the goal is simply to get laptops into the hands of as
many children as possible? If that is the case, why have we been
bothering to develop any software at all? And if others are making low
cost laptops that run Windows, why don't those efforts fulfill that
goal?

-walter

* Ironically, the majority of the system-level problems we had
experienced are directly tied to the two proprietary code bases on the
laptop: the wireless firmware and the embedded controller firmware.
While there are efforts to replace these, OLPC itself has been
diligently working with both Marvell and Quanta to make the best of
the situation. To suggest that fundamentalism has impeded progress on
those two subsystems is not correct.


More information about the Devel mailing list