Last night I went up to KRCB public television to help out with their election night coverage, at the request of Jake Bayless. So in a busy studio filled with hot lights and incessant chatter, I mashed reload on various web pages 'til numbers changed, then copied those numbers into a database, pressed "submit", and a process was started whereby Targa file slides eventually ended up in a place where the director could choose to use them.
The problems with this?
I can't believe that there isn't a standardized way for counties and the
state to publish this stuff. If ever there were a specific "useful" need
for the "semantic web", this is it. I'll be asking a few questions around
Marin and Sonoma to see if we can get something in place there for November.
This is exactly the sort of "news" I loathe. There was 2 hours of "male
answer syndrome", newscasters and commentators making stuff up and
prognosticating on a very small amount of actual data. I did get a different
view of some of the candidates, and boy I don't like either contender for
Looking at SFGate.com this morning
shows how bad "news" is: I have to actively read headlines and decide what
to click through on to figure out how the various propositions faired.
The workflow for generating those slides involved us mashing reload on web
pages, typing that data into a Filemaker Pro web interface, which triggered
something which output that data into some sort of pseudo-XML which got fed
into Motion, which
got dumped to .TGA files, which... So at the end of the night I thought "we
should have been Twittering the results", and I asked where we could stick
that in the process, and... yeah... It may be more pain to build your system
on top of open source to start with, but it'll pay off when you're a non-profit
publishing in a dying medium trying to figure out how to stay relevant to the