[om-list] Introducing myself again ...
MrM0j0r15n Rajeev the kanehbosm
electrickanehbosm at hotmail.com
Thu Aug 29 05:29:33 EDT 2002
Hello Mr. Packer,
I am truly sorry for the insufficient introduction i posted earlier.
I am Rajeev Joseph Sebastian. i am an undergrad student of Information Tech.
at
the Centre for Engineering Studies, Cochin Univ. of Science and Tech.,
Kochi,
Kerala, India. I am currently in Semester 7(of 8 sems), and i am researching
Knowledge
Management and Applied Epistemology as the basis for my Seminar and two
Projects for this and the next semester (approx 8-10 months out of which i
could
spend atleast 3 hours a day, cept during exams and shows). I am also
interested in these things as an amateur philosopher, and as a future
software
engineer and filmmaker (ie, purely for selfish reasons!).
I have about 4 years of experience developing on the Java platform on the
Linux
operating system. Some of the systems (in Java/Linux) i have developed on my
own are:
Bark - a fully compliant multithreaded IRC Server, AcidProx - (not again ;)
a fully compliant
multithreaded HTTP/1.1 Proxy Server, LibAssist - a prototype for a
distributed Library mgmt
system and some other smaller projects of a systems nature.
I am also designing an information system to be implemented. This system is
called All College Knowledge System (AXIS). It is a system to encapsulate
all knowledge within the human and documentary components of this system. It
achieves this by providing information technological tools to convert mental
and documentary models into Knowledge Objects which encapsulate them, and
allow queries such as one of you wanted, ie, queries based on "inheritance"
or isa qualities of knowledge items. This system was envisioned by me, since
I really felt the abscence of an "Encyclopedia" like resource (which the
Internet is nowhere near no matter how many Googles arrive), to perform
"instantaneous" research operations. However, i dont have the expertise or
even knowledge to "abstract away" many of these things such as operations
suggested in the list (natural language understanding support for search
systems), which are too far away for my personal skills (im only beginning
to become confused by Chomsky :). But, i think i have the good fortune of
finding a group as yourself interested in such developement. So, i humbly
request your help.
As far as the "knowledge" components of the software system go, these are
the following projects in mind:
<this is an excerpt from another file (a draft) ... >
Alchemy - a knowledge system as one advocated by your group (to be
implemented
currently by myself):
Alchemy is a tool to automate research based on documentary evidence for
knowledge (such as paper Assignments, Books, Articles, Films etc etc) stored
by various means and organized on the DocBook system as a data mgmt
component (www.oasis-open.org) Data Objects themselves are to be stored in
Object-Relational Databases, and filesystems. Alchemy is a server device
considered to be on the Internet. It runs a Servlet-based backend connected
to various databases in a distributed fashion such as in LibAssist above
(both local and remote). Alchemy is to be developed in 3 phases:
Phase 1 : Analysis of requirements for a basic knowledge system supporting
word searches complemented by techniques such as relative weight of word
within single documents, etc.
Also, a sophisticated classification and naming subsystem to identify
varieties and instances of knowledge items is to be "analysed". At first,
this subsystem must encapsulate
the requirements above, but it must be extensible. To do so, we may start
from the comprehensive documentation language DocBook. Additionally, models
(UML diagrams, Eletronic Circuitboard diagrams, etc) used in engineering and
applied sciences are also to be stored/manipulated in the system in a
generic way.
the Subjective Logic API, or some non-monotonic, or second-order logic
formalism could be used to perform
"automated research" on knowledge items. The goal of this phase is not to
develop techniques themselves, but
to develop the frameworks within which this can happen in the near (or far)
future.
Phase 2 : design of prototype/release candidate 1 developed in iteration 1
of phase 3
Phase 3 : implementation of AXIS-RC1 in iteration 1 of this infinitely long
phase.
As needs to include additional knowledge items arrive, we may
extend the DocBook system and evolve a new model. Also, humanware activities
like
regularly populating the knowledge bases by the community. One such project
is Galactica Initiative, which is a free (GPL or other)
encyclopedia-on-a-CD, for the poor public. Students, from 9th grade onwards
will be involved in an intensive multilingual (English and Malayalam)
content production exercise, which will ultimately create Galactica. There
are plans to co-ordinate activities among various instituitions of repute,
to digitize and/or produce "documentation" on things from stars to cooking
(the latter including my sister, of which she is a good one ;).
In RC1, we must develop the system to facilitate the basic usage scenarios
developed for students of my University. However, the models developed in
the earlier phases must consider this requirement only as a special case.
This is because of certain design considerations for a knowledge-intensive
educational system which is a feature of the human aspect of this software
system.
Client Interface System
The client interface system is now prototyped in HTML/Javscript, but there
are plans to create a thin client, to allow more complicated gestures than
possible on standard Web browser interface. This allows Alchemy to be an
Application Service Provider, that is, no longer do u need to seperately
download installation files etc for newer version, except for the initial
download. All user data is stored on server-side. Users may set personal
information as secret, and thus prevent other users to specifically use this
data. But, there seems to be procedures to perform statistical queries, and
still hide individuals secrets. Also, the Client must be able to store
knowledge items to disk, and update them automatically if necessary. What is
left on the client side, is basically the GUI. However, by adding specific
components, to the thin client, one is able to set up ones own knowledge
base (this is standard for desktop version) and thus facilitates the local
data storage scenario.
<nostradamus mode>
Alchemy is designed to be a system where "no word is undefined, no concept
is unexplained", ie, it must automate research. Alchemy must go
"intelligent" by 2050, and take over the planet ! j/k.
eg: design bridge systems to connect to various national patent agencies,
and bulk transfer pre-Alchemized "knowledge" to AXIS Knowledge Central,
located here in Kochi
eg: interoperability with conventional AI projects
eg: expert system shell with NL queries
eg: high bandwidth internet connections
eg: large scale conversions of major Internet sites that contain knowledge
that does not change (like back issues of journals, etc)
</nostradamus mode>
The design for the project is to be done in Unified Modelling Language
(UML).
Help!
well, now to the "emotional" part: i hope, the group members are interested
to let me join your unified effort to build a free knowledge tool for the
planet. I also hope, all of you
reply favorably. Thanx.
insanekane
if i had my way, all "words" would be in quotes
PS: btw, since im learning DocBook now, i could develop your site for you
_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com
More information about the om-list
mailing list