[Server-devel] Working on a theme - 'deployability'

[Martin Langhoff]

Early this week, in my xs-0.6 planning sessions I grouped the possible
tasks under general "themes" because I felt that it was the right
approach to prioritise/triage tasks with the very limited resources I
have for the XS. It worked very well, and it allowed Greg and I to
quickly figure out what our priorities and opportunities were, across
a sea of granular tasks full of complexities and interdependencies. It
also made our decisions easy to explain.

(The resulting plan is at:
http://lists.laptop.org/pipermail/server-devel/2008-November/002504.html
 )

And in the sugarcamp '9.1 brainstorm / planning' session, Ed talked
clearly about a dominant theme: Deployability. The themes for the XS
are "manageability" (server management, lease management), "large
scale schools" (3K users) and bits of moodle needed to administer
those other two. Both dominant themes for the XS are related to
blockers to deployments (Peru and Rwanda).

In a sense, it's a natural progression -- as we mature into having a
real product, face competition and difficulties we have to focus on
*removing each and every barrier to adoption*, every single thing that
prevents large numbers of users from using our XOs. And once they have
the XOs, every barrier that prevents using them effectively. Every
crisis is an opportunity to ask "are we focusing tightly enough on
removing barriers?".

Every bit counts -- if you like Joel Spolki's writing, this is a good
writeup of why this damn unsexy strategy is the way to win
http://www.joelonsoftware.com/articles/fog0000000052.html -- 'remove
barriers to $desired_action' is the mantra. In this case, deployment.
And there's a tipping point, but you can't know where it is in
advance, so row row row.

Personally, I find this energizing. I can get out of any meeting or
programming task that isn't very clearly and unambiguously removing a
barrier :-)

More importantly, it means that I can picture every hour I spend
eroding those barriers as a extending our collective reach to more
kids. A good morning of uninterrupted focus on scaling in Rwanda:
making a difference to 3K kids times the number of large schools OLPC
ever deploys to.

I didn't join OLPC to write PHP, Python, Ruby-on-Rails, Bash-on-Beer
or Erlang-on-Extasis. If it takes Cobol for the XS to reach more kids,
that's exactly what I'll do.

 http://www.amazon.com/COBOL-21st-Century-Nancy-Stern/dp/0471722618/

cheers,



martin
-- 
 martin.langhoff at gmail.com
 martin at laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff