next »

[sracer]

 I'm on the fence about this. Bumps in the road aren't the problem: the issue is whether or not systemic problems are causing OLPC to fail. The article today about software development is indicative of a systemic problem. OLPC, I have no doubt, is guaranteed to fail if it refuses to work with the community. There is no point to the software being open source if changes will break, contributions aren't included, etc. The more OLPC insulates itself from the world, the more it ensures its irrelevance and failure.

 OLPC has surely been successful in bringing low cost computing to the market place, but as on organization, it is doing alot of things that indicate failure, not success. That doesn't make them dead yet, but it is certainly justification for real concern.

The user community-at-large did OLPC no favors by blindly supporting and endorsing everything they said and did without being open to discerning analysis and criticism.

The few of us who raised concerns, were dismissed as shills for Microsoft, simply "didn't get it", or accused of getting involved for the wrong motives.  We were "encouraged" into giving our XO to people who "actually wanted one".

The net result is that virtually all who were able to discern the areas for improvement are gone.   How can improvements be made if people see things as being just fine?  By the time that those people come around to seeing the issues at hand, it is too late.

[Brian Basgen]

The few of us who raised concerns, were dismissed as shills for Microsoft, simply "didn't get it", or accused of getting involved for the wrong motives.  We were "encouraged" into giving our XO to people who "actually wanted one".

The net result is that virtually all who were able to discern the areas for improvement are gone.   How can improvements be made if people see things as being just fine?  By the time that those people come around to seeing the issues at hand, it is too late.

Sracer, I'm sorry to hear you've had bad experiences here. Anywhere I saw the subject of Linux vs. Microsoft brought up on these forums, it was handled very well, with reasoned discussion. I'm sorry you had different experiences. I agree that criticism needs to be encouraged, and that group-think should be avoided.

Regarding OLPC's problems, I think we agree that their narrow focus and inability/refusal to create meaningful and effective relationships with other businesses, in whatever form that may take, is hurting their relevance considerably.

[Wayan Vota]

The few of us who raised concerns, were dismissed as shills for Microsoft, simply "didn't get it", or accused of getting involved for the wrong motives. We were "encouraged" into giving our XO to people who "actually wanted one".

Ahem, don't forget the Intel shill accusations either. 

Which is why I think you may have missed sracer's point, Brian.  I think he's referring to the reaction of the greater OLPC community to criticism - not the OLPC News Forum.  I know I strive very hard to make sure this is an open conversation about OLPC - its good and bad points - and I think for the most part, we've achieved that.

But its also why OLPC News and Forums exist.  I definitely felt that the OLPC Wiki was very much fanboy-only until G1G1 and even now, a little too Pollyanna.  (I haven't been on the official forum enough to know its bent, but others tell me that its more realistic than the Wiki.) 

I remember being digitally booed when I first brought up simple questions like "Where will these countries get 100+ million?" or "What's the Total Cost of Ownership for an XO?" and my favorite "Got implementation plan?"  They're not as sexy as Python vs. C++ or cute as child + laptop, but just as important for the success of any large program.

Sadly, this line of questioning was lost to blind hope/belief that Technology Is God.  And while there may be few pragmatists left in the core OLPC organization, I am still here hoping OLPC can still change and succeed.

We all need our blind hope in something.





 

[sracer]

Yep, Wayan nailed it.

[Brian Basgen]

 I see, thanks for the clarifications. We all certainly agree, and it isn't just fan boy reaction to criticism, but it is a pervasive problem throughout the organization. Intel becomes the devil because they are competing with OLPC; sugar goes on what continues to look like a proprietary path as opposed to an open one, and so on.  It all seems to reflect a bad culture, and I think fans of OLPC and XO can be easily swept up into this cultural baggage. 

[Sarah Elkins]

Regarding OLPC's problems, I think we agree that their narrow focus and inability/refusal to create meaningful and effective relationships with other businesses, in whatever form that may take, is hurting their relevance considerably.

Brian, I think their relevance has been hurt worse by what seems like a dismissive attitude to the open source community.   They've produced great hardware with the help of some businesses and without others.  i like the software they've produced, too, including Sugar and the Journal.  But asking and encouraging software contributions (giving away laptops on that premise, even) only to throw away activities people have developed?  Not ensuring minimal backwards compatibility?  I hope the new CEO will fix (or find someone to fix) the structural issues, or whatever is behind the development issues.

[pgf]

sarah -- what specific issues are you referring to?  can you give examples of activities thrown away, or of not maintaining "minimal backward compatibility"?

i feel like i must be out of the loop or something.

paul

[Sarah Elkins]

Paul,

First, please note that the immediate context when I wrote that was in response to Brian's comment, which talked about OLPC's (problematic?) relationships with "other businesses" but did not mention OLPC's relationship with the open source community. 

Second, the larger context for me was what I've read about the recent PyCon:  OLPC is not doing (much?) regression testing (no automated testing, anyway), no current OLPC folks were at this big Python event, some Pythonistas are ready to see OLPC burned to the ground.  Even discounting for personalities (I wasn't at PyCon and don't claim to know the back stories), it seems like there's a lack of regard for the importance of maintaining good relations with the Python community.
* http://ivory.idyll.org/blog/mar-08/pycon-08-thoughts.html
* http://ivory.idyll.org/blog/mar-08/software-quality-death-spiral
* Mike Fletcher (mcfletch)'s "OLPC at PyCon Wrap-up" from March 22; I'm having trouble with the direct link, but Google cache has it: http://209.85.207.104/search?q=cache:xXS2MwnGlJQJ:blog.vrplumber.com/2081+olpc+pycon+wrap-up&hl=en&ct=clnk&cd=1&gl=us&client=opera  -- your cache link may differ, or you can read excerpts at http://www.olpcnews.com/software/applications/sugar_open_developer_software.html

Third, from what I have read (last link above, again), I had thought Update 1 was going to break the Sugar functionality of the existing Activities.  Perhaps, looking at the comments there, I misunderstood.  In the case of OLPC-developed Activities, this may just mean downloading the Activities again. In the case of externally-developed Activities, though, this may still mean developers trying once again to puzzle  through arcane issues (different arcane issues?) to make their Activities work again.  Davewa wrote a very cool Activity for the XO:  StarChart.  It's educational, it's pretty, and it's one of my favorite Activities. It would be a shame if it became unusable for Update 1.

I do understand that backwards compatibility is a factor to be weighed among others.  What I don't understand is the mindset that privileged sending all those XOs to LCA2008 before honoring outstanding G1G1 commitments.  What's the point of giving away laptops to near-random passers-by there to do something "amazing" with them --  when sustainable development is of questionable priority?   (APIs aren't documented, backwards compatibility is uncertain, minimal testing?  Don't document the GUI until AFTER it's done and there's  turnover of OLPC developers? (my inference from Jay G.'s April 3 comment, last link above))

I'm a fan of OLPC's mission, of the XO and Sugar and Journal.  I think OLPC has made a difference already, and I do appreciate all the effort that's been put into it.  I don't want to see that effort squandered.  As I said, I hope the new CEO will fix (or find someone to fix) the structural issues, or whatever is behind the development issues. 

[pgf]

thanks sarah --  i don't really have much in the way of response, but i appreciate your time in writing.

from asking on one of the XO development lists about activities breaking with update.1, the only thing that came up is that there may be some tightening of a security model (i forget which) which may cause a few things to need fixing.

and don't worry about starchart -- i hear it's being included in one of the major builds -- for peru, i think.  i got a first-hand demo of the activity from davewa at a recent user's meeting in cambridge -- it's very neat.

as for testing, and some of the other quality worries -- i think some of the worries are well-founded (after all, OLPC is a young organization, and is still learning), and others are simply a result of not enough (or, perhaps, "too much" -- take your pick :-) information available to the outside world.

time will tell...

paul

[Sarah Elkins]

That's good news, thanks for the reply, Paul! 

I'm sure it's a tricky balance for the OLPC folks to maintain, trying to encourage contributions from and provide information to the outside world, without losing  ownership (mission creep) or starting brushfires.  I still have hope that the boat will float, not sink.  :-)

[Torokun]

so it officially sunk...

[Directive0]

You have a funny understanding of what the word "official" means.

[Peter Ruhe]

It's perseverance that will determine success,
not flawless execution.

Thanks, Directive0, I collect words of wisdom, and I really like that line.

I am curious where the money comes from to keep a staff of 30 employed. I doubt they work for free. I hope OLPC has enough resources to enable the necessary long-term perseverance. I can't see part-time amateurs like me making much of a contribution, because the XO software is so complicated that full-time focus seems necessary.  I tried loading build 703 to help test it, but ouch what a painful reality check it has been. I hope other XO enthusiasts are more capable than me, and have more time too.