XO-tablet development? [Devel Digest, Vol 90, Issue 8]

[James Cameron]

On Tue, Aug 13, 2013 at 10:10:38PM -0700, Yioryos Asprobounitis wrote:
> > On Aug 13, 2013, at 1:09 PM, Yioryos Asprobounitis wrote:
> > 
> >>>>  Paul wrote:
> >>>>  If not, which way the tablet/laptop development team is heading?
> >>> 
> >>>  don't know, sorry. 
> >> 
> >>  OK, thanks. 
> >>  Hopefully someone that does will care to comment.
> > 
> > It is lack of accurate knowledge of the future, not caring,
> > that keeps me from commenting.
> > 
> >> The current laptop and tablet implementations differ significantly and 
> > would be important  to have an indication where the (limited) resources should 
> > be deployed .
> > 
> > How would you devote resources to the tablet ?   It is closed source.
> > 
> > There is an attempt within the Sugar (and OLPC) community to move
> > the educational environment to HTML, where it can be used on both
> > Android (tablet) and Linux (laptop).   That is where I would suggest
> > deploying resources.
> 
> 
> Hopefully I'm not stepping on too many toes, but I would think that
> this might not be the best course.

Not so much stepping on toes but late to the party; this was being
talked about on sugar-devel@ months ago, and we would not want to
waste the work done so far.

> If you are to re-write the activities or even Sugar anyway, why not
> go native Android for ARM machines?
This is the current offering
> from OLPC and they appear to be much more popular and readily
> available in developing (and even developed) countries. 
> It makes a lot more sense developing for the economical, low power
> and already available machines.

I'm curious as to what you mean by native Android.  You mean Java?

The big advantage of rewriting from Sugar in Python to Sugar in HTML
is that the execution platform widens to cover both Android and
personal computer operating systems, with a single source code base.
I don't see how a rewrite in Java using the Android Software
Development Kit would help with that.

I have seen adequate performance of Sugar in HTML activities so far,
and in some cases better performance than Sugar in Python activities.

> I can understand that the android kernel is not currently supporting
> the ARM-XOs but OLPC since the beginning maintained its own kernel
> and still does, so it should be feasible.

Your assumption about feasibility is easily challenged; we do not have
as many kernel maintainers now as we had before, so our history is not
a reliable indicator.

> I can also understand that the Android stuck is not GPL3, but
> Apache2 is still "open" and between a closed system as the XOtablet
> is and a free one as the laptop, may be a viable
> compromise. Besides, you can still release your code under GPL3. In
> essence is the same as releasing GPL3 HTML5 activities to be run on
> Apache2 or even fully proprietary OSs.
> Yes, the entire offering will not be GPL3 but software-politics
> aside, will allow you to do as much as you currently can on the
> laptop and way much more than what you can on the tablet. 

I don't see how licensing is relevant to the technical problem at
hand, but the Sugar in HTML activities seen so far have been open
source.

> BTW given that currently the laptops are distributed to the end
> users mostly locked, licensing is not even relevant for them.

No, laptops are much less likely to be distributed secured now.  It
depends on the deployment what they choose, and they have to be able
to sustain the necessary technical effort before they choose secured
laptops.

No, licensing is still relevant to end users, regardless of whether
the laptops are secured.  But OLPCs main customers are the deployment
teams, rather than the end users.

> It is also likely that OLPC will be criticized for releasing
> something under a mixed Apache2/GPL3 license but given that the
> XOtablet is already there. there is nothing (more) to  wary about
> ;) 

Yes, I see you are criticising.  ;-)

-- 
James Cameron
http://quozl.linux.org.au/