[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

:zz: How much semantics to build in? (Was: Re: Conjoinment (was: User semantics for dimensions / sug'd guidelines



Mark-Jason hath wrote:

>In order to do [whatever], ZZ must know that d.sex is symmetric and
>transitive.  Otherwise it should give you a warning first, because
>what you are asking for makes no sense, and if you try to alphabetize
>in a dimesion that is not symmetric or transitive, you will change the
>meaning of the structure:

Thanks for helping me clarify, i.e. make explicit various
 thoughts which I may not have said out loud yet.

ZigZag as we see it here is intended as a very general
 mechanism demonstrationg--
- default visualization (connected cells) for everything
- default interface (connected cells as menu) for everything
- highly generalized, principled interconnectivity
- high functionality without the usual stuff, esp. a user world
 without--
-- hierarchies
-- metaphors
-- icons
-- applications

In the legendary Director's Cut version, which I intend
 to flesh out Real Soon Now, there will be about 16
 predefined dimensions with intended structure.

The guidelines for my *suggested* extensions will include:
- putting each new function in a cell (ignored in '98 by the
 highly-pressured Implementors)
- constraining what can happen on particular dimensions
- constraining what particular cell types can do (we don't
 even *have* cell types yet)
and, in general, a kind of reverse-inheritance by constraining
 particular stuff, rather than creating new functions that
 some things get and others don't (regular inheritance).

There are many possible other words buildable on the
 same connective structures with more constraints
 and more ideas.  In principle these are all valid.
 The directions I'm taking this, toward the DC version,
 reflect only my own preoccupations and predilections.
 Serving suggestion.  Mileage may vary.  Void where prohabited.

ChrzT


At 10:31 AM 10/28/98 -0500, you wrote:
>
>> This "conjoinment" may be a good term.  I have
>>  been noticing that you generally want new dimensions
>>  in pairs in just this way,
>> d.inside / d.contents
>> d.mark / d.marklist (tells what cells are part of
>>  "the same mark")
>
>That's what I was thinking.  Any time someone is going to want to have
>a relationship that is one-to-many, they will need to do this.
>
>> "Transitive" means that ordering is important. 
>
>Not exactly.  Order can be important wihout being transitive.
>
>Consider:
>
>	---> d.ancestor
>
>	A--B--C
>
>Here you want --- links to express `ancestry'.
>A is the ancestor of B, and B is the ancestor of C.
>But it is transitive, so A is also the ancestor of C.
>
>	---> d.child
>
>	A--B--C
>
>Here you want --- links to express a parent-child relation.
>A is the parent of B, and B is the parent of C.
>But is is *not* transitive, so A is *not* the parent of C.
>
>The two pictures look exactly the same, so if ZZ is going to
>understand the difference between these two pictures, there will have
>to be an explanation somewhere else.
>
>> See my paper on Preflets
>
>Where?
>
>> I think 1 and 3 come from somewhere else
>>  (as you see, I'm no mathematician) and may not 
>>  have meaning or relevance.  Hey, I'm open--
>
>I wrote them down because I ran into examples where they were
>important. 
>
>Symmetric means that in A---B, A is related to B in the same way that
>B is related to A.  If you hop A over B and make it B---A, the meaning
>does not change.  Example: d.contents is symmetric.  d.inside is not.
>
>If ZZ is going to do things like selecting and gathering cells
>automatically, it will have to understand something about the meaning
>of the structures you build.  Symmetry and transitivity are both
>important here.
>
>In an earlier message, I suggested that there could be an operation
>for automatically alphabetizing cells:
>
>	---> d.sex
>	MALE   -- Homer -- Bart -- Ned -- Barney -- Moe -- Snake
>
>You'd put the cursor on MALE, press the button, and the structure
>would become
>
>	MALE   -- Barney -- Bart -- Homer -- Moe -- Ned -- Snake
>
>In order to do this, ZZ must know that d.sex is symmetric and
>transitive.  Otherwise it should give you a warning first, because
>what you are asking for makes no sense, and if you try to alphabetize
>in a dimesion that is not symmetric or transitive, you will change the
>meaning of the structure:
>
>	---> d.son
>	Grandpa -- Dad -- Ted
>
>	||   Alphabetize
>	\/
>
>	Dad -- Grandpa -- Ted        (ooops)
>
>> >There were a couple of others I thought of, but my notebook is
>> >downstairs.
>> 
>> Uh-oh.  You DEFINITELY are one of us.
>
>My notebook is now in the compuiter, so beware.
>
>
>
____________________________________________________
Theodor Holm Nelson, Visiting Professor of Environmental Information
 Keio University, Shonan Fujisawa Campus, Fujisawa, Japan
 Home Fax from USA: 011-81-466-46-7368  (If in Japan, 0466-46-7368)
Professorial home page http://www.sfc.keio.ac.jp/~ted/ 
_____________________________________________________
Permanent: Project Xanadu, 3020 Bridgeway #295, Sausalito CA 94965
 Tel. 415/ 331-4422, fax 415/332-0136  
http://www.xanadu.net
PERMANENT E-MAIL: ted@xxxxxxxxxx
_____________________________________________________
Quotation of the day, 98.11.01:
"Life and death are both hereditary."  TN59