The purpose of this note is to update you on my Stanford TAP KB efforts, =
and the tools that I am using.
1. I have upgraded my Knowledge Explorer (KE) program to read all three =
of the download formats=20
of the TAP KB: mcf, rdf, daml..
2. I have written a command, mkrGetData, which translates MKR questions =
into TAP KB queries
which are passed to Rob McCool's Java TAP client to retrieve answers =
from the TAP KB.
3. I am currently integrating TAP and tabula rasa contexts into a single =
a. I downloaded tap.daml from TAP KB.
b I translated it into MKR using my rdf2mkr command
c. I am using MKR concept integration & differentiation commands, isi =
plus manual editing of hierarchy outlines (KE output becomes KE input)
d. I am using MKR scripts to automatically rebuild the KB after edits.
For the details of 3, you can browse my web site directory
Specific files of interest are
tap.daml TAP KB download (without instances)
tap.mkr MKR KB (without instances)
taphi1.ho merged TAP and tabula rasa context (partial KE output)
tapisi.mkr MKR concept integration commands (first step)
taphi2.ho integrated TAP and tabula rasa context (first step) =
(partial KE output)
tapscript.ho MKR script to read KB and apply integration rules
I also have the full TAP KB in both rdf and mkr formats. I haven't =
files on my web site because they're about 15 MB each.
4. As a fun project, I wrote an ExtractData command for a friend.
GetLetter is a KornShell script which downloads the daily batch of =
Letters to the Editor
from the Fresno Bee web site, using either wget or rebol. =20
ExtractData is a Unicon procedure, integrated into KE, which parses each =
(unfortunately, no RSS at Fresno Bee) and records the information in =
knowledge :=3D man do identify od existent done
knowledge haspart proposition list