has: "
I have modified my original ideas about there being three types of "queries", in two ways.
(1) The third type of "query", called "distance from prototype/archetype" can and probably should be generalised to something more like "distance from centroid/centre-of-mass" in the same mathetical state/classification space, (i.e. a functional space where the domain is a many-dimensional state or list of features, and where the range is classificational values for many classes). But, if we model psychology well enough, the centre of mass should turn out to be very close to the conceptual "prototype" anyway.
(2) There is a fourth type of query. Just as the third query is a fuzzy version of the first type (the first type being the idealistic-inductive or extensional-definition-ba sed query) the fourth type is a fuzzy version of the second type (the second type being the idealistic-deductive or intensional-definition-ba sed query).
More about all this later.
My next suggestion is "semi-symbolic", (semi-symbolic/semi-geome tric, is how I think of this), or, in other words, graph theory. Mark, I strongly suggest we/you look into what, in graph theory, are called "isomorphisms" and "homeomorphisms". They are very similar to a part of this process of identification. Two reasons: (1) if there are algorithms for determining homeomorphisms in weighted, directed graphs, then our job is easier. (2) If not, then our job will involve more word, but it will be work that will advance the field of graph theory, and should be published.
The way I see it, the process of homeomorphism-finding should be specialised to apply to a graph whose edges (internodes) are not just weighted and directed, but are of differing sub-types, corresponding to different dimensions in space; and every relation or attribute in the database (such as your GEDCOM files) would be encoded as nodes and internodes.
.
logical statements embedded in the model (used w/ queries etc too):
they are like a "query info cache" to save the system from repeating the same long queries many times, just to find out that "all ___ are ___"
or basic mathematical principles or something
they have to be validated
(or is really 4 something below? see "they have to be validated")
whether by a background process scanning all statements not checked since a certain "latest update" date on the individual queryies or lit of
them, or by checking them every time you enter new info into the model.
entities referred to in them are regular entities within the model.
they are related to contexts within the model: expected to be valid in some, not others (or valid to a fuzzy-defined extent as per tom?)
they could be used for constraints on individual entities or classes of them.
Put in something that checks om db'ss own integrity
for all files it stores (by storing a checksum of it to compare) & mark which may be getting corruptedband need to get from backup (or the
checksum is wrong). Also somehow checks all other data like the db integrity itself? Can the code be stored in it, and/or other system
backups?, then back it up separately, and check the integrity of everything as you do? What best long-term plan give this fallen world but
terrestrial/celestial goals?
How do we keep data integrity in general, where you would have constraints in a relational DB or something to make sure things stay right?
make it use a bundled database, so they can install just w/ java app & not have to inst postgres or anything, so small, lightweight, quick? (see
http://www.quicklist.org/testimonials.html)
Might also need to schedule reminders to check for all formats and make sure can still be read/displayed, or if there is a need to convert them
to new formats. for old ones, save the code or something? What best?
used for logical analysis?
what language or format are they stored in?
handling verbs: see Mark's comments on "sense & tenses", 10/2003 on om-list.
simulations (related to how I'd handle verbs?):
ideas of mine from 12oct2003:
Participants in simul. are the ones selected by user and all that are proximallyimmediately related. any entity "affected" (state
changed) during th simul. pulls in to th simul. as well all its related entities. (Limit toal #? or try both & see if affects so
optimized either by some limit to total # of participants in simul. or by guessing which to omit because relahionship is distant or not
affected enough?)
Entities to participate in simul. need scripts that respond to events like changes to geogr. location of neighbors., time, or whatever
events) could affect them. The simulation management code (or related objects?) calls these events (or related obj's call them to pass
on th effects, while still respecting the timslicing of the simul.?).
thread management tool makes sure the simeslicing happens fairly, and granular "enough" (per "n" instructions have a timeslice?)
How make these easy to invoke & run & write event handler code on objects & prepare & get and/or save results, or auto-update model &
show consiquences or inconsistencies in the model based on relatiosn & natural consequences of events in the simulation?--for the user
or internal system who may initiate it?
How make this flow into or use tools of neural nets or expert systems, that's easy for a non-articulate or non-tech user to run &
enhance/improve the model & the simulation stuff (like C)?
read up on simulations:
incl. Tom's links on SOAR toolkit, posted to om-list in oct2003.
.
.
.
security:
indvidual users: who can see what? owner/author/viewer? groups, or...? **collect requirements.
jdbc to server layer?--or other issues related to making sure the postgresql host authentication is what we want?:
see jdbc part of postgresql documentation on authentication methods it accepts, & admin guide to configure them?
.
when adding new features, document the use case/story/scenario that drives it, so ppl know what it is for? How do this documentation?
How do system documetnation in general?
.
How we could do "context" and "time"? (model this then ask others if seems good--get design feedback from ppl how best? read up on simulations etc
etc or just do the design then publicize it?) Make sure design is gpl 1st.
maybe: context is done by what we are pointing to--it is a matter of the database changing so that the getter methods return the data that fits the
situation. And the situation is defined by the time/place you are in, so you'd have to tell them that.
maybe: simulating time/place changing is done by the objects themselves knowing when & where they and/or their attributes are, and when the
time/place changes, they are notified by some loop which sends them the change event or simply calls a time or place change method, and they respond
by switching to different pointers for the values of their properties according to the internal info about what goes w/ what time/place.
Maybe this is done by having an event loop that has an array of all affected objects (they tell their own children somehow) and as we go thru time
it notifies them all and they react accordingly (by updating their own display or whatever). To optimize, we could have this loop (or a separate
one for those having the information) that tells the next time interval on which the notified objects would care about a change, so that we only
notify them if the time threshhold in the array has passed. Design this into the system for review at this time or later?
see notes elsewhere here re 'context'
can it search via dif't/multiple names for the same thing, and where there are > 1 results, prompt w/ the other names to see which, or go by linked
proximity w/in the model, or by context in the sense that one name is more applicable to another because the other name really means something
else? Example: air conditioner (used to mean 'swamp cooler' to me), but in this household both terms have individual meaning, so I should be able
to find either one, w/ either term, esp if using other terms or actions that specifically apply to just one of them, or it is nearer in context, or
more recently used term in conversation/work, or is the user's (more context) "favorite" term for the situation, or according to the language they
speak, even if others use dif't terms for it. (And the system uses the others' terms when communicating w/ them: if not just GUIDs and then it
looks up the terms for them on the spot, based fully on their context at that time.
Could it handle things like "ask the other questions for this person at the same time" even if the todo or whatever came up on a task that one only
sees now: serendipitously combine things that are "near" in person, place, mental or physical requirements, ....
.
review questions on the yellow notepad ppr etc, plus Mark's email I saved about class def's & "is-a".
.
make a copy of the discussion summaries from mail archive, by month, for the list, & save it in my backups? (do I still have too much data saved at
pair?)
read:
and read re related work!:
?-- http://developers.slashdot.org/article.pl?sid=02/11/27/1720249&mode=thread&tid=156
http://developers.slashdot.org/article.pl?sid=02/11/27/1720249&mode=thread&tid=156
http://mnet.sourceforge.net/intro.php
http://www.neurogrid.net/Decentralized_Meta-Data_Strategies-neat.html
http://cubicmetercrystal.com/alpine/
http://slashdot.org/article.pl?sid=02/12/06/1527214&mode=thread&tid=95
stuff about logic, in encyc britannica
a bit about star logo? (logo in java)) which alos does interesting simulation stuff w/ distributed systems (like life?)
article referred by Michael Iantosca (miantosc@us.ibm.com): something by david sklar in about '98, electronic book technologies, a theory for
knowledge representation or such.
so maybe I need to learn about the forms of reasoning & their AI representations as shown in that paper & references, plus statistics? Then revisit
my design ideas? Also learn re simulations.
Beyond Software Architecture
http://books.slashdot.org/article.pl?sid=03/07/03/1423256&mode=thread&tid=126&tid=156
Beyond Software Architecture explains how to bridge the gap between the marketecture and tarchitecture- how to create a product that not only
performs well, but which also appeals to the market. It is part of the Addison-Wesley Professional Series line of books (also containing notable
titles like Design Patterns, Refactoring, and Patterns of Enterprise Architecture) but this latest installment in the series is (thankfully)
paperback, so it comes at a paperback price ($39.99 USD).
important
put on list for s.o. else to help? do this for licensing/copyright reg, fsf etc?
There is no legal requirement to register your copyright with anyone; simply distributing the program makes it copyrighted. However, it is a
very good idea to register the copyright with the US Registry of Copyrights, because that puts you in a stronger position against anyone who
violates the license in the US. Most other countries have no system of copyright registration.
If the program runs on some form of the GNU operating system, such as GNU/Linux or GNU/Hurd, we would probably like to make a link to the web
site for your program. So please tell webmasters@gnu.org the URL. Also, we would like to list your program in the Free Software Directory once
it is working solidly. Please contact directory@gnu.org.
It is also possible to make your program a GNU package, a part of the GNU Project. (That's if we like the program--we have to look at it first,
and decide.) If you might be interested in joining up with the GNU Project in this way, please tell gnu@gnu.org, and we'll explain more.
.
for single-key data entry in this cmdline app, see comments etc w/ suggestions at
http://developer.java.sun.com/developer/bugParade/bugs/4050435.html, including http://www.ethannicholas.com/enigma/
semantic web compliance
check out the postgresql "routine database maintenance tasks" page, & what I should be regularly doing?:
file:///usr/doc/postgresql-doc/html/maintenance.html
update the web page to show where the dev't action is (on the list)
have them discontinue the other list?
post use case list again.
consider whether to use uniqueidentifiers in pgsql instead of long values (so larger db? more unique across installs? smthg?)
use genguid (or other tool) to make a globally unique number?
see: http://www.taurussoft.org/files/
description:
Uniqueidentifier datatype is 128-bit (16 byte) wide datatype claims to be unique across Universe. It's a good choice for primary indexes and
making relations between tables. This inplementation depends on libuuid library from e2fsprogs package by Theodore Ts'o. Licence is LGPL.
.
could post a mention of the web site to the slashdot discussion on gelernter, or email them, and/or put it on freshmeat. See their guidelines first.
Could also post to directory pages. Revu brint?--post anything there? (MAKE SURE MAILING LIST IS WORKING, FIRST!! -- SWITCH LIST TYPES?)
helpful
web site to-do notes (maybe):
add note on web site or here smwhr re why not logical assertions? (no inheritance; they may represent obj's but it seems you need a taxonomy
first--this model is the taxonomy; names built-in, but names are not data, just a name. objects are the data, which may have many names. I think
they richer, more real representations when full obj's & relats, & therefore can be processed more effectively, realistically, & thus we can think
better & get better results.
why not lisp (could be, later, if modularized well?) or...., & fewer libs/progrms (incl me; time2learn)
post our archives under the document list, linked to the regular archive page. Also could link to the email messages I saved from before--our earlier
digests, too.
later on (put this w/ other onemodel todo notes & start consulting them) get a bugtracking system & post our source in a tarball somewhere. And get
us an IRC channel to talk on?
.
"Copyright 2015-2024 Luke A. Call | OM Home
Contact (won't put you on a mailing list): comments@onemodel.org (but for replies see the mailing lists)
(Content generated selectively from a OneModel instance.)