RE: [humanmarkup-comment] Base Schema - symbol

[Rex Brooks <rexb@starbourne.com>]

Attribute quicksand=Confusion and overlap of terms between elements 
and attributes, in various schemata. Makes it more difficult to write 
apps correctly. Attributes are really second class elements that in 
some way are derivatives or aspects of the element to which they 
belong and which then get a datatype to which you can assign a value. 
Too many steps. Too many opportunities for errors to be generated. I 
think most attributes could be elements on their own. In any event, 
we don't have the choice except to limit use of attributes, for which 
we are doing an acceptable job, in my opinion. My opinion only.

I just like to keep things as simple as I can, and in that sense, 
having sign as the only element in the primary is actually simpler, 
but just requires reworking everything else. I think it can be 
handled in the semiotic process.

I think we will have problems later with classes, regardless, and 
what we need to do is what we have already done in the Requirements 
document by saying that there MUST be a means or method to change and 
adapt and add to the Schemata of HumanML. It is the how that remains 
to be thought about.

When I suggested that we think down the road about the Secondary 
Base Schema I was thinking about defining the levels of abstraction 
that go from symbolizaton to figuration/tropes in language groups.

I really do think that the semiotic processor is where the best work 
could be done on making sign the most atomistic principle, along with 
the semiote in building the most fundamentally accurate model of 
communication.

Ciao,
Rex



At 9:50 AM -0500 10/10/02, Bullard, Claude L (Len) wrote:
>No, not a heart attack, but my initial drafts for
>the primary were based on the assumptions that
>experts would tear into it and offer suggestions
>for redesign.   We can do as you suggest, but I
>think we will have some problems later with
>the classes.
>
>Saussure assumed that all semantic assignments were
>arbitrary.  I agreed with that until I looked into
>icons.  A signifier whose meaning is derived from 
>resemblance to the thing it signifies is not
>arbitrary.
>
>If we want to tie to emotions, I think Pearce
>comes closer to the model we need.   A sign
>system with no connections to the physical
>systems is essentially inert.  It has no
>source of energy.   It is good perhaps for
>the categorization of texts, but not much
>good for generative modeling.
>
>Why are attributes quicksand?
>
>len
>
>-----Original Message-----
>From: Rex Brooks [mailto:rexb@starbourne.com]
>
>I don't agree, but not because I don't think this is logically
>consistent, but because XML Schema attributes is a pit of quicksand I
>am desperately trying to avoid. I can live with humlIdentifierAtts
>and I could even live with humlCommIdentifierAtts. However, in terms
>of processing, if I had my way, we would only ever have elements with
>values for a clear set of datatypes, but that is not possible.
>
>Also, I think we would open a big door into much more work than I
>have the capacity to handle if we do this. While I freely admit that
>I am emotionally aghast at the prospect, it isn't the emotion which
>is driving my opposition to this. It is the amount of retrofitting
>work this would take. I just don't have the capacity. It is taking
>all I can do to complete what we are doing. If we do this, we need to
>find someone who can do the work within OASIS because I don't have
>the capacity. I have basically given up my income-making work, not
>just due to this work, but because graphic design always takes a big
>hit in economic downturns, and I suspect that once history takes some
>of the margin of error out of the economic statistics for the last
>year and half we will find that we have had a stagnant economy at
>best and a recession, especially if you factor out non-discriminatory
>(necessities) consumer spending and a Real Estate market thriving on
>the lowest lending rates the fed can manage to sustain.
>
>So, is it possible to construct something of a First Principle? Could
>we just make sign the basis of the semiotic process, which doesn't
>need the rest of our Primary Base Schema except as elements. We don't
>have a web language that fits it and we certainly do not have the
>accompanying structure to make procedures work for it within the
>existing structure of the web, but we also can't lose our XML Schema
>mentor.
>
>I agree, in the beginning there is the sign and the perceiver, but I
>wouldn't put too much emphasis on Peirce. Saussure had a lot to say,
>too.
>
>Ciao,
>Rex
>
>P.S. Len, are you trying to give me a heart attack? ;)


-- 
Rex Brooks
Starbourne Communications Design
1361-A Addison, Berkeley, CA 94702 *510-849-2309
http://www.starbourne.com * rexb@starbourne.com