I had been thinking along the web way, but this paper suggests another
approach more suited to the semantic web. Interesting insight.
This is almost exactly what I do with planner. I try to keep commits
as small and self-contained as possible. Not only does it make it
easier to roll back (although I still have to figure out how to mark
certain changesets as do-not-use, aside from deleting them from the
revision library), but it also makes it possible for people to
This works to my advantage as well, as other people are encouraged to
make their patches nice and small. I still have to hand-tweak some
changes. For example, Gary Vaughan uses a different tree structure.
However, I can review the patch logs and merge the changes in
If they mailed you responses, they’d effectively kill the community.
C’mon, are you going to keep checking back there? I wouldn’t. Good
point. Very very true.
I think I should show this table to my class to give them an idea of
their career path and the growth required.
Link from TerryP’s blog
I think this is why I have so much fun working on PlannerMode. I have
a clearly-defined set of users in mind (mostly the mailing list) and I
love threshing out interfaces and features with them. It’s not a
large-scale application and it certainly doesn’t appeal to the masses,
but it can be quite powerful when you get the hang of it.
Hey, Emacs is like that too. =)
Open source supports situated software very well, as you can tweak
something existing in order to fit your needs.
By the way, the author taught a class called Social Software.
Must forward to Dr. Rodrigo. This is her kind of thing.