RE: [humanmarkup-comment] Base Schema-community: SEMIOTIC COMMUNI TY

[cognite@zianet.com]

At 10:19 AM 07-08-2002 -0700, you wrote:
>Getting closer. The notion that we can use semiotic topic maps 
>appeals to me as a way to organize sign systems into data 
>dictionaries. It also happens to answer the question of how we are 
>going to get the relevant data resources connected up for HumanML 
>applications.

It may work; there's already structure to connect to.  The current
graph structures may need extending....  It's not trivial, but hopefully
can be done in steps.  Perhaps -- moving from the shallower use of
already-explicit
notations that can be applied, onward to our own terms/tags/processes and API
for apps, though.  

Actually there's so much structure already that seeing where to go....  8-o   ;)

<snip>

> a WASP cultural entity exists 
>simultaneously in 1776 and 1976, though the attributes of said 
>culture for any particular time period will be slightly different 
>both in the way that the community views itself and how it is viewed 
>from without or by other cultural communities.

very interesting that the processual definition of 
those-who-communicate-constitute-a-[changing]-community

makes communities of whoever continues the dialogue...so that 
literary critics join with the authors they discuss even posthumously 
in a community [of common interest]!  Nice, I reckon.

Makes it quite consonant with topic continuity, all right.
There is work on Self-Organizing-Maps; for a quick overview
this is cool:  http://www.oreillynet.com/lpt/wlg/1081.  It gathers
from Finance, Bio, and NLP (Natural Language).

>REX> memberships in communities would occur by assertion, I
>>assume, or by behavioral tracking. Communities would be defined by
>>shared sign systems. Topic Maps group associations by categories,
>>categories are organized by ontologies, found by search engines of
>>search engines, mostly all done by metadata in the headers of
>>documents, delivering the sign systems to which an individual belongs.

Headers contain readily retrievable metadata cues to such things as sets of
 conventions pertinent to various natural languages: 1, pre-established.

The documents themselves bear information too about sign systems they
 embody: 2, intrinsic.  
        Some of it may reside in tags:  2a, explicit;
        Some may some not: 2b, derivable.  

Cumulatively, the sign System -- along with the referents (signifie's) that
are necessarily part of semiosis -- is emergent.

How does that come into play?

Here's an example I just ran across of XML tags that gave very pertinent 
info for HUML (REF:  http://www.ontopia.net/~grove/software/xmlarch/ )

Result

  <?xml version="1.0" standalone="yes"?>
  <persons>
    <author>Geir Ove Grønmo</author>
    <mentioned>Eliot Kimber</mentioned>
    <mentioned>David Megginson</mentioned>
    <mentioned>Lars Marius Garshol</mentioned>
    </persons>

from:

<ul>
    <li><a
href="http://www.ornl.gov/sgml/wg8/docs/n1920/html/clause-A.3.html";>Architec
tural Form Definition Requirements [AFDR]</a></li>
    <li><a href="http://www.ornl.gov/sgml/wg8/document/1957.htm";>ISO/IEC
10744 Amendment 1</a></li>
    <li><a href="http://www.isogen.com/papers/archintro.html";><i
persons="mentioned">Eliot Kimber</i>'s "A Tutorial to SGML
Architectures"</a></li>
    <li><a href="http://www.sil.org/sgml/topics.html#archForms";>The SGML/XML
Web Page: Architectural Forms and SGML/XML Architectures</a></li>
    <li><a href="http://www.megginson.com/XAF/index.html";><i
persons="mentioned">David Megginson</i>'s XAF package for Java</a></li>
    <li><a href="http://www.megginson.com/SAX/index.html";>SAX: The Simple
API for XML</a></li>
    <li><a
href="http://www.ifi.uio.no/~larsga/download/python/xml/saxlib.html";><i
persons="mentioned">Lars Marius Garshol</i>'s SAX for Python</a></li>
</ul>

I don't know what the rule is that lets these inserts be inside of
<i>...</i>, or if you 
could use <SPAN>....</SPAN> where there wasn't a handy enclosure by layout 
markup tags.  Do you?

In the example file biblio1.out for this python ADFR (XML Architectures)
parser, a relation
between the author's name (and original site of www.infotek.no ?) and
Norwegian is also shown, but I don't see where it came from; loss of the
slash thru Gronmo is noted -- perhaps that was done assuming Norwegian
character set, but metainfo has this -- doesn't it indicate English as the
document language?

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "">.

-------------

?test where="before"
?test where="after"
(biblio
-
  
(firstname
Anationality norwegian
-Geir Ove
)firstname
(lastname
Aht #IMPLIED
Amodified yes
-Gronmo
)lastname
(note
Aht address
-You can reach me at 
-grove@infotek.no
)note
-
  
-

)biblio

-------------

SC