Piggy-bank continues to haunt my dreams . . .

From: Dean Allemang <dallemang_at_acm.org>
Date: Wed, 23 Nov 2005 18:10:10 -0500 (EST)

I sent this personally to Ryan, but he suggested that it would do better
here in the general Simile list. Some background for those of you who are
not Ryan; I have been following Piggybank for a while, but two weeks ago at
ISWC finally got it to do all the things that the demos say it can do, and
well, I think it's pretty cool.

So here's the message I sent to Ryan:
>>>>>>>>>>>>>>>>>>>
I am impressed by the way Piggy-bank can display a bunch of items on a
google map. Just like the fellow who gave the talk said, this lets me show
my friends what I am doing with this weird thing called The Semantic Web.

I have some questions about how to extend piggy-bank. This might be the
impetus it will take to get me back into coding :) Or maybe, I can just
talk you into doing it.

Suppose I wanted to do the same thing you do for google maps, but do it for
a calendar. That is, the "map" is actually the 12-month calendar (shown on
a single page). The "pins" point to days (maybe with a different
graphic). Otherwise, the sytem works the same as it does now, only keying
on a different property.

Application to make my friends go "wow": Show the birthdates of everyone on
my livejournal "friends" page.

The same thing with a single month, where you have more space per day, so
that each "pin" can actually be a clickable block, for keeping a calendar.

Application to make my friends go "wow": Show all my appointments this
month. Merge in those of my boss in a different color, so we can find a
time to meet together. Merge in my private appointments that I don't want
to tell you or my boss about in my own view that she doesn't see.

Suppose I have a custom diagram that is meaningful to my business (or my
client's business . . .) Something like an org chart, or a system
architecture diagram, or something like that. I have metadata on artifacts
in my KM system. Show the results of some query laid on on this "map".

Application to make my customers go "wow": Well, I think I just described
it.


Also, how hard would it be to throw in an RDFS+[0] reasoner that runs on
the current triple set, just before you show it? That would make it a lot
easier to manage the key information that drives these applications.


Dean

[0] RDFS+ is the name I give to the very low-demand set of inferencing that
includes the RDFS stuff (subClassOf, subPropertyOf, domain, range) along
with a few simple OWL things (inverse, transitive, functional/inverse
functional). It can be implemented in about 15 minutes, once you include a
RETE engine or some such thing into your infrastructure. Yeah, I know,
JESS has licensing issues.
Received on Wed Nov 23 2005 - 23:03:54 EST