OLPC upgrades

[david at lang.hm]

On Wed, 11 Feb 2009, Martin Langhoff wrote:

> 2009/2/11 Wade Brainerd <wadetb at gmail.com>:
>> I have spent the last 10 years optimizing software for a living, and can
>> tell you without a doubt that any system can be optimized without starting
>> from scratch.
>> It's just much harder to understand the performance characteristics of a
>> large, complex, existing system than it is to imagine an ideal performance
>> world in your head.
>
> +1 on both. Experience teaches that performance comes from
> understading and optimising a complex system. If you build a
> replacement system, it's initial performance _will suck_ until you
> study it, understand it and optimise it.
>
> We can skip the "build a replacement" step, and head for the goal faster.

the entire sugar infrastructure is a 'build a replacement' step. people 
are questioning if it was the right thing to do as opposed to using 
existing tools and infrstructure. (with the other part of the question 
being is it faster to go ahead and try to optimise Sugar or switch to more 
mature solutions)

David Lang

> (I've only seen one exception to the rule above -- and was the
> building of git. But it's a rare case where you get a band of
> battle-hardened kernel devs building a small and focused cli tool.)
>
> cheers,
>
>
>
> m
>