[DAS2] DAS code sprint proposal

Andrew Dalke dalke at dalkescientific.com
Fri Dec 16 12:28:29 UTC 2005


Hi all,

I've been thinking more about the comments from Gregg and others
during the previous phone conference.  He was concerned that there
are some major spec changes this late in the DAS/2 grant period.

It does seem rather late, but from what I've seen in spec development,
especially ones like this which combine people from, what, 6 or
more sites and will be used by many more people, it's not uncommon.

It's just not possible to emerge the spec full-blown like Athena
from the head of Zeus, any more than it's possible to write 10,000
lines of code and have it compile the first time, much less run.
Even with peer review.

I think we have a solid idea of what the spec should look like -
though I'll have a busy few weeks assembling them into words.
I don't think the object model has had that much in the way of changes;
mostly it's been a question of pinning down a few details.

I think it's time to schedule a DAS code sprint, where the different
server and client people get together to implement the spec and
use that to provide feedback for the spec development.

My feeling is that people are ready for this too.  I know I would
rather code specs then doc them!

Christmas is coming up and there's usually the week or so at the
start of the year where people are getting back up to speed.
That sound like the end of January / early February is a good time.
For me the week of 6 February is the best.  That's probably far
enough in the future too for people to clear out time.

My thought is to have two groups of people, one in Cambridge and
one in Emoryville, but I've not heard of a geographically split
sprint like that.


					Andrew
					dalke at dalkescientific.com




More information about the DAS2 mailing list