[IAEP] IAEP Digest, Vol 84, Issue 2

[Tony Anderson]

Hi, Sean

I thought the strategy of OLPC was very clear - sell national 
governments on providing laptops to all of their children. However, 
OLPC, independent of this strategy,  made two critical decisions: G1G1 
and using open software that have made all the difference.

The open software decision led to the development of the our current 
community,  Nicholas Negroponte's famous 'software' mistake.

The G1G1 model defines our community problem today. The Give 1 part gave 
focus on the need to provide and support the laptops at community 
schools in the developing world where electricity is a problem, internet 
is unknown, and hands-on experience with computers is minimal. The Get 1 
model gave focus on the use of the XO by one's grandchildren in the home 
where internet access is ubiquitous and everyone has hands-on experience 
with computers.

This is analogous to the difference between teaching a Spanish speaking 
child English as a second language in Massachusetts and in rural Peru.

The XO-1 is still viable in the Give 1 world. Outside of hardware 
problems, it delivers the same educational experience it did in 2007.

Naturally, the XO-1 is not marketable in the Get 1 world. The developers 
and supporting volunteers live in the Get 1 world. As a consequence 
going to a software model which jettisons etoys in the interest of 
keeping up with Fedora seems a reasonable tradeoff. Someone in Give 1, 
who has never before had a computer in their hands, is unlikely to know 
or care what model of Fedora is installed.

This begs the question, what has changed between Sugar 0.82 and 0.104 
that significantly improves the value of the XO in primary school 
education in the Give 1 world?

We should consider the real model of deployments (aside of the national 
ones). Some individual or group in the Get 1 world decides to sponsor a 
set of laptops for a specific school or library in the Give 1 world, the 
deployment.

The role of the sponsor is to coordinate with the deployment, develop a 
plan to provide electrical power (e.g. agreeing to pay for utility bill 
or getting an agreement that the deployment will pay), acquire the 
laptops, arrange for the laptops to be delivered to the deployment 
(often in luggage), and arrange for someone with technical skills to go 
to the deployment to set up the system and show the staff how it works. 
Naturally, my personal interest is that the sponsor should supply a 
school server and one or more routers to provide the XOs with access to 
some of the information the Get 1 world routinely obtains from the internet.

Normally, installation of software is not an issue. The sponsor handles 
that. The ongoing problem is that the community assumes the deployment 
has a similar familiarity with computers as is common in the Get 1 
world. The only introduction to computers is typically a few day 
workshop at the deployment when the laptops are delivered and installed. 
Further, the clear pedagogical vision is not communicated leaving the 
deployment to figure out how the XOs are to used effectively. The 
laptops are not used to provide continuing education in their use. The 
consequence is the often observed drift at the deployment into limbo 
(i.e. the computers spend most of thier time in the packing boxes).

If we need a marketing program, it is to find sponsors to fund and 
support deployments in the Give 1 world. This program should be 
accompanied by an effort to find unused XOs and get them deployed for 
the simple reason that the initial $200 investment is paid and they are 
immediately usable. Where are the XOs given to Mongolia? The program 
should include particular attention to making the task of sponsorship as 
easy as possible and on giving the sponsor a clear understanding of the 
pedagogical goal of the program.

Should we encourage or recommend deployments of computers/tablets/smart 
phones other than the XO? So far as I am aware there are no Sugar-based 
deployments on laptops other than the XO (Classmate in Argentina?). The 
initial reaction to Raspberry Pi is that when you added the essential 
peripherals (monitor, keyboard, camera, microphone) and packaged them in 
a portable package - the cost would be comparable to that of an XO. I 
have seen nothing to change that judgement.

If we are willing to accept a computer lab model in which the XOs never 
leave the school - the 'thin-client' model may be useful. In this model, 
the computer may be a Raspberry Pi with a monitor, keyboard, and mouse. 
Since it never leaves the lab, the packaging is not important. The 'thin 
clients' could be connected to the school server by an ethernet switch. 
One obvious consequence is that the learners will never have a chance to 
read ebooks, listen to music, listen to native English speakers, 
complete KA Lite exercises, or explore what they can do in Scratch or 
Turtle Blocks. They will get access to the computers on a schedule set 
by the school and will be expected to stay 'on task' while in the lab.

HTML5, CSS3, and Javascript certainly need to be supported by Sugar at 
the level of Python. Does this require replacing Python? Is Python ready 
to join APL and Basic in computer museums? A simple localhost (software 
feature) enables javascript to launch Sugar activities or resume them 
from the Journal. Lionel Laske's Sugarizer works on an XO as far as I know.

I think the concern about making Sugar viable on any platform is 
reasonable if Sugar is viewed as one computer application among many. It 
is certainly what the Get 1 world expects. However, how does this help 
the mission of providing a laptop to every child in a community school 
in the developing world (the 60% without internet)?

This issue of the plan for the future of Sugar and the olpc initiative 
is not simple and it does not have any easy answers.

Tony

On 03/02/2015 06:36 PM, iaep-request at lists.sugarlabs.org wrote:
> Message: 1
> Date: Sun, 1 Mar 2015 22:57:57 +0100
> From: Sean DALY<sdaly.be at gmail.com>
> To: Samuel Greenfeld<samuel at greenfeld.org>
> Cc: IAEP SugarLabs<iaep at lists.sugarlabs.org>
> Subject: Re: [IAEP] Planning for the future
> Message-ID:
> 	<CANnY+GN+QAv1jQiUJH380RQ7gFzoeD0_xF4z_iFqMav9cSEciQ at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Hi Samuel,
>
> thanks for this
>
> I believe Sugar has had a clear pedagogical vision from day one, but has
> not had a strategy for some time.
>
> Outside the XO, Sugar's historical technical architecture has unfortunately
> kept it out of reach from all but the most determined and tech-savvy
> teachers (and journalists). Without a pancake button download and one-click
> installer, the installation barrier is too high. OLPC's historical focus on
> the hardware was never helpful either, and the main reason OLPC got mauled
> by incorrect memes was they didn't want to accompany journalists past the
> unfamiliarity barrier of the XO (hardware+software).
>
> In my view there are only a few ways to overcome this issue:
>
> * Develop 1-click installers for Windows / MacOS / GNU/Linux. I had
> suggested maintaining a matrix of preconfigured (i.e. languages/keyboards,
> prepopulated Journal, selection of Activities) VMs over Oracle VirtualBox,
> whose license allows free distribution for nonprofit and educational
> purposes. Upsides were immediate fullscreen Sugar experience without
> touching the configuration of the host computer. The downsides were huge VM
> images and the effort required to build and maintain the matrix. At the
> time I suggested we approach Oracle for corporate sponsorship, but some
> community members voiced objections.
>
> * Arrange for Sugar to be preinstalled on low-cost, reliable machines other
> than XOs. This is complex and would require a sales force (or working with
> a partner's) since no OEM will make that investment without a prospect of
> selling many thousands of units. As an alternative I had suggested we ride
> the wave of Raspberry Pi units (five million sold in three years) by
> developing an SD card for it based on Sugar on a Stick, but there was no
> interest in that effort. I still believe a Sugar-branded version (case +
> teacher starters kit -documentation) could have an impact.
>
> * Migrate to a web-based Sugar compatible with browsers on any platform.
> Lionel's Sugarizer is I think a fabulous solution.
>
>
> I've heard it suggested that marketing could do fund-raising, but donors
> large and small won't want to contribute unless there is a plan. I've been
> bewildered what the plan is for some time.
>
> Sean