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

Re: Clarif. re inside/contents



> "Inside" is transitive-- (could it be otherwise?)--

I'm not sure, but I agree that it would have to be a rather bizarre
counterexample. 

> Now, the problem with only having that one dimension
>  to represent containment is that in this model,
>  each thing can only contain *one other* thing,
>  like nesting Russian dolls.

That is a persistent problem with the ZigZag model.

I am compiling a list of examples where that property is inconvenient
or clumsy; this was one of them; the male-female example I mailed
earlier is another.  

When I first understood how inside-contents worked, I said, ``Oh, what
an awful hack.''  If this one hack is the only one, that is not too
bad.  But sometimes when you have a hack it bespeaks a very deep flaw
that will have to be patched over and over again with hack after hack.
That's my biggest worry about ZigZag.

Of course, I'm not suggestion that you change it; the property is
fundamental to ZigZag and I think it's important to find out whether
it works or not.

>   A design choice is whether to have some title, or just the first item,
>  at the top of the list;

Yes.  That's precisely why I was asking about B in this picture:
> >> >        A - B - C      +--> +d.inside
> >> >        |   |   |      |
> >> >        X   B1  C1     v +d.contents


> The expected structures is: 
> d.contents \/   d.inside =>  ("|" here means "no connection)
> A a
>    b
>    c
>  B d
>    e
>    f
> 
> where for some reason (probably visual convenience in
>  some context), A *might* be connected to B, abd c to d,
>  but those connections have no system-supported connection.

Maybe I misunderstand your illustration, but if c and d are connected,
won't the system interpret d, e, and f as part of the contents of A?

(I usually write `c-d' as an abbreviation for `c and d are connected'.)