next »

[mavrothal]

Disclaimer: I’m not associated with OLPC, SugarLabs, any local or wider Linux group or any software or hardware company/user group. I’m not in the field of computing and my only association with computers is as an end-user for the last 20+ years. Mac, Windows a brush with VML and Unix and since last May an owner of a G1G1 XO-1.
Some terminal experience, some scripting in defined environments, some rudimentary web page building, some looking under the hood, and that’s it regarding computing.
Regarding learning, some tutoring experience.
I have no first-hand or behind the scene information and all I know is coming from web pages, blogs and mailing lists. All suppositions are either educated or wild guesses...
I liked the original OLPC idea and thought was a worthy goal. Thus, I got a bit more into it.

So, If you go on reading, you have been warned… The issue at hand is the XO-1/Sugar marriage and separation and where we go from here.

Few years back 2 (for the propose of this writing) groups of people saw the possibilities and promises that computers could offer in teaching and learning particularly for the developing world and started an ambitious noble and worthy cause: provide “one laptop per child” in the developing world. One focused more on the development of the proper hardware, while the other on the development of the proper teaching software environment.
However, this merging had a fundamental flaw. The hardware could not support the software and vice versa. Hardware was focusing on minimal initial and running cost and minimal energy consumption resulting in a sophisticated but very underpowered for today's standards machine, circa late 90s.
The software on the other hand went mainstream with today's standards. Enterprise level infrastructure (RedHat/Fedora) and trendy, powerful and recourse demanding interpreted programming language (Python) that could better and easier implement the programming goals. 
The initial compromise was increased specs and cutting corners in software. The result was a pre-alpha product (G1G1-2007) that a year later advanced to an early beta (G1G1-2008) and later "advanced" to the realization that the two can not continue together, simply because of the inherited incompatibilities.

Parenthesis: Someone may say “is the economy stupid” but is not supported by the data. The bad economic turnout may have contributed, but was not the problem. XO-1 had an initial monopoly and within a year, from the 14 million Netbooks sold at prices $300-450, G1G1 2008 sold 7000. A good overpriced product will survive (see Apple). An alpha product will not. That’s what XO-1/Sugar was. In this economic environment, no one would throw way $400 for an alpha or early beta product no matter how noble the cause may be. That’s the contribution of the economic downturn in the OLPC-Sugar separation. Exacerbation of the inherited conflicts.   

At this point OLPC decided “the heck with Sugar” is never going to run properly on this machine and, if THEY want to run it on it “they are very welcome to try”.  SugaLabs decided “the heck with the XO-1”, Sugar is never going to run properly on that machine and, if THEY want it to run on it “they are very welcome to try”. Indications about this? Very little about SoaS on http://lists.laptop.org/pipermail/devel/ very little about the XO-1 in http://lists.sugarlabs.org/listinfo/sugar-devel and there is only a https://www.redhat.com/archives/fedora-olpc-list/ that is struggling to bring the two ends together. Basically is trying to make fedora run properly on the XO and as a consequence Sugar 0.84+ may do the same…

What is the way out from this?
Obviously better hardware and better software. The current but separate goals of OLPC and SugarLabs.
What is the problem then?
- A shuttered credibility of both organizations as incapable to deliver on their promises.
- Delivering developing world to Microsoft by opening the door and leaving it open on their way OUT.
- Discrediting FOSS as “scratcher” instead of “achiever”. 
But MOST IMPORTANT a million kids with XO (and hopefully many more to come) with forgotten and/or dieing software.

Should they abandon their VERY VALID goals for better hardware and software respectively and instead keep cutting corners and compromising on the XO-1/Sugar combo, then?  Absolutely NO. But both the XO-2 and a “mature” Sugar may be 2  or more years way and the “next quarter profits” are not so pressing anyway… What they should really do is to DEVOTE a portion (25-50%) of whatever resources they have to the support and development of the platform in the field. Progress in other areas may be delayed by 25-50% respectively, but when it arrives, will arrive on a bed of “happy customers” and “success stories”. And I’m sure that any progress on XO-1/Sugar is not going to be wasted regarding XO-2 and Sugar 2.0 development. So the delay may not even be so big.
It is arguable however if this can be done within each existing organization. They simply may not be able to “carry two watermelons in one hand”. One will fall and break for sure, and most likely both.

So the way to go is to form a 3rd independent organization from THEIR limited resources to do exactly that! Support Sugar on the XO-1, re-instate Sugar road map  and work on update-9.1, as originally planed. 

“They are very smart people. They know better” you may say. Yes, and No. Every one knows of very smart people doing stupid things when not in their field of expertise or under emotional stress. And that what I think happened with OLPC and SugarLabs that led to their separation. Separation is a fact of life, but usually both parents take care of their common kids the best they can. It is good for them. It is good for their future-parenting attempt. It is good for their future kids.
That’s what OLPC and SugarLabs should do and  don't cut corners on "child support".

So please, look after the REAL best interest of your "kids" and yourselves.
And who knows… MY XO even benefits from this… 

     

[Eddie Owens]

I certainly concur in Mavrothal's view that both OLPC and Sugarlabs need to step up and accept their parenting responsibilities for the million or so XO-1 users out there, rather than abandoning their child to the snakes and tigers of the binary jungle!

Sugarlabs needs to insure that each upgrade in "Sticksugar" is reflected in a comparable advance in the XO-1 based Sugar. 

OLPC needs to ensure that parts and technical support remain available for XO-1 for 3-5 years.

And both Sugarlabs and OLPC both need to make available whatever hardware or software information is required to ensure the widest possible utilization of the XO-1 by other Linux distros.

I am thinking specifically of DSL and Puppy Linux.  Both of these load into and run from RAM, and are reportedly very fast on hardware platforms even less capable than the XO-1.  But thus far no one has been able to to make either DSL or Puppy boot on XO-1. 

Either of these running on an XO-1 would seem to be a match made in heaven!  Early in OLPC's history, the Puppy community apparently made overtures to OLPC, exploring the possibility of being used in some fashion on the XO-1, but their efforts were rebuffed (thanks Ivan!). http://www.murga-linux.com/puppy/viewtopic.php?search_id=105084931&t=7992.

Come on, Walter and Nicholas!!!  You have uttered verbage suggesting that you are committed to FOSS...may your actions confirm that.  It is in everyone's interest that XO-1 remain a viable entity, running a broad spectrum of software (including DSL and Puppy Linux).