After XS 0.6 - future-of-the-XS wishlist...
martin.langhoff at gmail.com
Wed Jan 14 16:19:51 EST 2009
Yesterday, as part of XOCamp, Caroline Meeks pressed me a bit to talk
about my vision of the XS going forward. It is a somewhat hazy space,
as it's unclear whether I'll be able to work fulltime on this after
Now, even when you feel really certain about your future plans, a bus
could roll you over. Or your convertible could leave the road and
crash on the rocks in the sea.
So what's the XS plan then?
With 0.5.1 much of the infrastructure work is done, including network
services. With 0.6 that work should be complete, even if it'll be
lacking some nice-to-haves. Pat-on-the-shoulder, job well done.
The next stage -- if *I* get a chance to get it done -- is all about
focus on server-based education tools. First stop, Moodle. Right now
there is a barely usable moodle that auto-installs. I've written 2
plans for it -- one focused on the technical and more "mechanical"
issues, parts of it are already done, but many remain to do:
the second plan is more focused on making Moodle useful with primary
schoolers. Currently, moodle is great for text-centric interactions,
forums, wikis, reading content. But children 6 to 9 years are not text
centric yet. Surprisingly, it's not so hard to address that - along
with a host of other things to lower barriers of entry for classroom
use - see
It's probably about 8 months of work to get through all that. So a
year out, my crystal ball gets hazy -- in the fog, I can make out the
outline of some further plans, such
- merging edublog patches, make the moodle-based blog able to publish
to internet-based blogging sites - such as blogger.com
- add a wikipedia-slice to the server - while the laptops do have a
"wikislice", the server can host a much thicker slice, and perhaps
have it editable too
- complementary tools: Mahara, Mediawiki
- of course, better tools to manage the deployment...
- rebase on Fedora 11 or 12 (skip F10 as it's missing bits we need)
- rebase Moodle work on Moodle 2.0 - which brings a nice API to deal
with repositories, unlocking a lot of content-repository options
- replace squid with rproxy and hashcache
- use btrfs as soon as usably stable, wrap yum/rpm actions in
there you go. Now I can go back to living life dangerously without
feeling guilty :-)
Note note note note: this is pretty much *my* plan, my perception of
the most reasonable tradeoffs between the competing/conflicting
feature requests and the physical and resource constraints. Anyone
prepared to spend the 1% inspiration + 99% perspiration can refocus
this plan on areas that I might be ignoring... by getting it done.
In other words, the above is all hot air, and hot air doesn't count.
Working code does ;-)
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
More information about the Devel