[IAEP] Sugar on Android via HTML5

[John Watlington]

On Sep 10, 2013, at 5:04 PM, Sameer Verma wrote:

> On Thu, Sep 5, 2013 at 7:51 PM, Caryl Bigenho <cbigenho at hotmail.com> wrote:
> One of the things that makes Sugar the ideal learning platform for children (and youth) is the wonderful compatibility of so many of the Activities ... both from Activity to Activity and from student to student. This facilitates the sort of learning we are all hoping to see more of... creative problem solving, project based learning and cooperative learning. Without this ability to integrate parts of projects, it would just be another collection of apps.
> 
> 
> I did not want to muddy the picture by injecting my own viewpoint, but now that I've heard from others (on and off list) it is clear that the split is driven by the role they play in the ecosystem. 
> Most technologists have come up with reasons why they don't think a complete Sugar experience would work on Android. Therefore, activities must run like any other app on Android. On the other hand, as Caryl said, "Without this ability to integrate...it would just be a collection of apps". 
> 
> Somewhat knowing the limitations of what can be done with Sugar stuff on Android, but disregarding that for a minute, I would say that Sugar as a *platform* is an experience. It has a UI. It has a UX. Everything from the Zoom interface to the activities to the Journal is Sugar. We have taken the original "Sugar on the OLPC XO" experience and replicated that to the classmate PC, SoaS, and other spins and distros, but in none of these cases did we break the holistic Sugar experience. Now, along comes a popular OS, and because the tech parts don't fit, we are advocating breaking up the pieces and taking whatever flies. Memorize will become one of the few hundred thousand apps on Android.
> 
> I disagree. 
> 
> It's like saying we'll do the cat sprite from Scratch, but nothing else. It's like saying we'll do the birds and pigs from Angry Birds, but not the slingshot. Sugar, without all its pieces isn't worth the trouble.

Sameer,
   I disagree somewhat with your thesis (and am very glad you started this discussion.)

From a technological standpoint, it is actually probably easier to implement what you describe:
Sugar as a monolithic Android application, which takes over the entire user interface when
launched.   The reason I never considered it seriously was the larger ecosystem.

The reason to move to Android from Linux is two-fold:
- Chip vendors are dropping Linux support in favor of Android.   The cheap chinese ARM
 vendors only support Android.
- Android/iOS are where application development is happening.  There is a much larger
community of Android developers than Linux or Sugar developers.

The hope was to provide the infrastructure underlying Sugar (the Journal datastore and
collaboration) as Android services, encouraging their use in new Android applications.
In this model, the Journal is another Android application, accessing the Journal datastore service.
New Sugar activities written in HTML should be capable of running in Sugar on Linux
or as Android activities (although perhaps with different execution wrappers).
In this manner, perhaps we can enlarge the Sugar community with developers mainly
targeting Android.   If we pursue Sugar as a single Android application, with embedded
Python activities, we are isolating ourselves from the Android community.

The danger of this approach is the loss of an integrated UX.  This could be addressed
by customizing the home UI, in the same manner that the XO tablet has a custom home UI
implementing the Dreams interface, but that would require "rooting" the tablet in some manner.
But the native Android UI isn't that bad...

Cheers,
wad

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sugarlabs.org/archive/iaep/attachments/20130912/64aa73eb/attachment.html>